David Maldow and I have been cooking up The AI Chefs for more than 90 episodes. Each one serves a practical, bite-sized way to get more real work out of AI. We’ve now organized the best of them into one menu: The AI Chefs Bistro at aichefsshow.com. The goal is simple: find what you’re hungry for, watch a short episode and use the technique the same day. No hype; just recipes that work across the tools people actually use, including Microsoft 365 Copilot, ChatGPT, Google Gemini, NotebookLM, Claude and Zoom AI Companion. The menu is organized by effort and appetite: Appetizers: five-minute tips. Examples are “Short Order Grill” for cutting AI chatter down to bullet points, and “Twice Baked Potatoes” for asking AI to check its own work. Soups & Salads: foundational concepts, such as why AI gives dramatically better answers when it can see your email, calendar and documents. Entrées: full workflows. These include a Teams Channel Agent that writes weekly status reports, and “Obstacle Omelette”: a pre-mortem prompt that tells you how your project failed before it starts. Chef’s Specials: signature techniques. “Doggy Bag” packs a messy conversation into a reusable system prompt, and “Deconstructed Dish” captures your writing voice. Sides: platform-specific tips, from the Microsoft 365 Admin Agent to keeping Copilot Researcher from citing your own drafts back to you. Desserts: fun experiments, because AI should be enjoyable too. Each dish is also tagged by business outcome, so you can filter by what you need. The tags are Prompt Craft, Writing & Content, Research & Analysis, Workflow Automation, Meetings & Comms, Data & Documents, Learning & Upskilling and IT & Admin. If your organization is rolling out AI and wondering why adoption is stalling, start here. Tools don’t drive adoption; habits do. These recipes are designed to build them. Pull up a chair: The AI Chefs Bistro.

2026 may be remembered as the year AI agents moved from impressive demos to real business interactions. The technology is advancing quickly; the business models, operating practices, and expectations around it are struggling to keep up. In a recent BCStrategies podcast, five of our analysts and consultants, Tom Brannen, Blair Pleasant, Kevin Kieller, Chuck Vondra, and Steve Leaden, compared notes on what stood out most this year across customer experience, unified communications, enterprise strategy, and hands-on consulting. The common thread: AI's potential is enormous, but realizing it takes far more than deploying another agent. Tom Brannen: When AI Agents Become the Customers Personal AI agents are starting to research, book, buy, and request support on behalf of consumers. This is no longer theoretical: a major hotel brand is already fielding contact center calls initiated by AI agents. That raises questions traditional CX models were never built to answer. How do you authenticate an agent acting for a customer? Does it get the same service priority as a human caller? How do upselling, loyalty, and sentiment measurement work when the immediate participant is software? The takeaway: businesses will need to design experiences for both human customers and the AI agents representing them. Blair Pleasant: CX Is About Outcomes, Not Interactions The focus is shifting from handling and routing interactions to the complete customer journey and relationship, and from answering questions to completing workflows and resolving problems. Blair pointed to Zoom's "conversation to completion" message and Cisco's emphasis on relationship management as examples. If AI agents are digital workers, they need to be supervised, governed, and measured like workers. And with personal agents on one side and business agents on the other, agent-to-agent interaction looks set to be one of the defining CX challenges of 2027. The takeaway: the differentiator is moving from efficiently managing interactions to delivering meaningful outcomes across the whole customer relationship. Kevin Kieller: AI Has Thrown Enterprise Pricing Into Disarray Subscription licensing was predictable; consumption pricing, AI credits, and usage-based models are not, and customers often have no reliable way to forecast usage before they deploy. Kevin described the market moving from "token maxing" to "token axing" as organizations shift from encouraging experimentation to controlling costs. If an agent is positioned as a digital employee, what happens when its monthly allowance runs out mid-month? Which department pays when one agent serves several? Outcome-based pricing sounds appealing, but defining a successful outcome (was the issue really resolved if the customer calls back?) is far from simple. The takeaway: AI pricing must become more predictable, transparent, and outcome-aligned before organizations can build critical workflows on it. Chuck Vondra: The Problem Isn't Hype; It's Unrealistic Expectations Many organizations are enthusiastic about AI but haven't defined what they want it to accomplish. Chuck described a credit union that bought an intelligent virtual assistant and expected it designed and live within two months, overlooking the substantial requirements, integration, and testing work involved. As Kevin put it during the discussion: "No one wants AI. They want the magic that they think AI brings." AI isn't necessarily overhyped; it is often misapplied. The takeaway: successful adoption starts with the business problem and measurable success criteria, not with selecting the technology. Steve Leaden: Test AI to Fail, Not Just to Succeed In a healthcare deployment covering prescription refills and appointment confirmations, planned across 120 groups, pilot testing exposed long silences before transfers and repeated "do you want to be transferred?" prompts. Minor in a demo; damaging in a real patient interaction. The team recommended more testing rather than launching, including scenarios outside the scripted workflows, such as a patient describing a medication informally. The economics can still be compelling: one model compared roughly $1.2 million in annual staffing cost against about $200,000 in annual AI usage, though real results depend on resolution rates, implementation costs, and workforce decisions. The takeaway: try to break your AI before your customers do. ROI depends on disciplined implementation and testing against real-world scenarios. Five Lessons Heading Into 2027 Prepare for agent-to-agent interactions. Engagement models, security, authentication, and service policies will need to evolve. Measure business outcomes, not AI activity. Completed workflows and resolved problems matter more than interaction counts. Understand the economics before scaling. Consumption pricing introduces real budgeting risk. Start with clearly defined, achievable use cases. Executive enthusiasm cannot substitute for requirements, realistic timelines, and readiness. Test aggressively before deployment. A successful demo is not proof of production readiness, especially in high-stakes environments. Looking ahead, the panel expects context and orchestration to get much more attention as agents multiply across applications and departments. The winners won't be the organizations deploying the most agents; they will be the ones that understand where AI creates value, manage its costs, and implement it in ways that reliably deliver better outcomes. This post summarizes the original BCStrategies article. Read the full piece and join the discussion in the comments on LinkedIn.

For years, the meeting room industry has measured interoperability with one question: how easily can a room join a meeting hosted on another platform? If your organization runs Microsoft Teams Rooms (MTR), joining Teams meetings is easy. Ideally, users should also be able to walk in, tap a button, and join a Zoom, Webex, or Google Meet meeting. In my latest article, I argue we may be asking the wrong question. Rather than "How easily can I join the meeting?" we should be asking, "How easily can I accomplish everything I came to the meeting to do?" As meeting platforms evolve well beyond audio, video, and screen sharing, that distinction matters more and more. Direct Guest Join: Simple, But Limited Microsoft's Direct Guest Join (DGJ) delivers a genuinely simple experience: find the meeting on the room controller, tap Join, and you're connected. For organizations standardized on MTR, joining other platforms without changing equipment or requiring a laptop is genuinely useful. For attendees who mainly need to listen, speak, and view a presentation, DGJ may be all they need. The limitations surface when users need to do more, and they can become critical when you are hosting. Microsoft's current DGJ implementation does not support sharing content through the room's HDMI input, interactive whiteboarding, or controlling the meeting lobby. It also carries restrictions on video resolution, display configurations, and access to certain collaboration features. Enhanced SIP connectivity closes some of the gap, supporting 1080p video, two displays, and HDMI content sharing. However, it is a paid interoperability service that requires a SIP dialing plan plus additional Exchange and tenant configuration. Neither DGJ nor SIP interoperability offers the complete functionality of the meeting platform's native application. An Interesting Alternative From Barco This is why I believe organizations evaluating Teams Rooms should take a serious look at Barco's ClickShare Hub. Barco's approach doesn't require abandoning an investment in Microsoft Teams Rooms. The Hub Core and Hub Pro provide a native, certified MTR experience when deployed with supported peripherals, so Teams meetings work with the familiar room controller and management capabilities. When someone needs to run a Zoom, Webex, or Google Meet meeting, they plug a ClickShare Button into their laptop, connect wirelessly to the room's camera, microphones, speakers, and display, and use their normal conferencing application. The important nuance: Zoom or Webex isn't running natively on the MTR. The meeting runs on the user's laptop, using the room's professional AV equipment. Employees use the platform exactly the way they already know how. The Experience Beyond Joining Consider an employee hosting a customer meeting on Zoom. With DGJ, they can join from the Teams Room, but may still need their laptop for host controls, content sharing, chat, or other features DGJ doesn't expose. Many end up joining the same meeting twice: once from the room for audio and video, and once from the laptop to collaborate and manage the meeting. That works, provided laptop audio is handled correctly, but it adds complexity. With ClickShare, the host uses the familiar Zoom application, with whatever host controls and collaboration capabilities their account and meeting settings permit, while the room handles the AV. There is no juggling two separate conferencing endpoints. The benefit extends beyond the meeting itself. Using a personal, authenticated client keeps recordings, transcripts, summaries, and follow-up activities connected across scheduling, conducting, and following up. DGJ doesn't prevent access to those materials afterward, but the room's guest experience sits apart from the user's own meeting account, which adds friction. The key differentiator isn't that joining is easier. It's that employees use the platform the way they already do on their laptops, while benefiting from the room's professional AV. That consistency has value. Can Direct Guest Join Ever Keep Up? Meeting platforms are evolving rapidly as AI becomes embedded in the experience: meeting assistants, transcription, translation, summaries, intelligent content, and automated actions. Each platform has its own implementation, identity requirements, licensing, and approach to exposing those capabilities to other applications. For DGJ to keep pace, that functionality must be available through the interoperability mechanism and supported by the room system. Because DGJ relies on the host platform's WebRTC experience, some improvements can arrive through the meeting provider's own updates. Still, there are architectural limits; features that work beautifully in a native client may not reach an interoperable room experience. And even native room systems don't support every capability of their desktop counterparts. My view: DGJ will likely always be playing some degree of catch-up. Not because Microsoft isn't investing, but because interoperability and native application functionality are fundamentally different propositions. As conferencing platforms become intelligent collaboration environments, that difference may become increasingly significant. ClickShare Isn't Without Compromises None of this makes ClickShare automatically the better choice. DGJ is excellent for straightforward meetings, especially when users simply want to walk in and join without opening a laptop. ClickShare requires a compatible laptop, adds another device to the workflow, and depends on the wireless connection and room peripherals working correctly. In practice, though, most employees already run meetings from their laptops when they aren't in a meeting room, so it is a familiar workflow rather than a new one. For organizations that primarily use Teams and only occasionally join meetings hosted elsewhere, DGJ may be perfectly adequate. Other options are also worth evaluating, including enhanced SIP interoperability services, wired laptop connectivity, and room systems that support more than one native meeting platform. The point isn't that ClickShare is universally superior; it's that it represents a different, viable approach to the interoperability problem. It's Time to Rethink How We Measure Interoperability Today we tend to judge room interoperability by which platforms a room can join, whether one-touch join is supported, and whether audio, video, and content sharing work. Those measures are important but increasingly insufficient. Organizations should evaluate the complete meeting workflow: scheduling, joining, presenting, collaborating, hosting, and following up. Can employees accomplish what they need without workarounds? Can they use the platform's features? Can they run the meeting with tools they already know? A room that joins every platform but restricts what people can accomplish isn't necessarily better than one that requires an extra step to connect. Successful meeting room interoperability shouldn't be measured by whether users can quickly join a meeting. It should be measured by whether they can accomplish everything they came to the meeting to do. As AI continues to reshape meetings, I suspect that distinction will only grow in importance. Update: Microsoft Cross-Platform Meeting Connect On the day I planned to publish the original article, Microsoft announced Cross-Platform Meeting Connect, which in the future intends to address some of these interoperability issues. Key points for now: it is not yet available; Zoom currently says it will arrive before the second half of 2027; and Microsoft hopes other vendors will announce participation soon. I think it's a great initiative, and I'll provide updates as features become generally available. My analysis focuses on what is available today. And the question I asked about DGJ, whether it can ever keep up, applies equally to Cross-Platform Meeting Connect. This post summarizes my original LinkedIn article. Read the full piece and join the discussion in the comments on LinkedIn.

For Episode 8 of Calibrating Copilot, Dino Caputo and I tackled a big question: can Copilot predict the future? We put Microsoft 365 Copilot, the Researcher and Analyst agents, and Cowork at Light, High, and Max through a forecasting exercise, then compared their predictions with actual results. Why flight data? We originally planned to use financial data, but we ran into something researchers call temporal data leakage. Even when you tell a model to use only historical data, it was trained on much of the internet and already 'knows' what happened later, which makes a fair test very difficult. Flight delay data is far less likely to be baked into a model's training. Canadian airlines report this data but it is not made public, and EU data is less detailed, so we used US flight records: a compressed folder of just under 70 MB. Several files exceeded Excel's limit of 1,048,576 rows, so you could not even open them in Excel to do the analysis yourself. (For the curious, Excel supports up to 16,384 columns.) The prompt Using the attached US flight records, forecast August 2026 percentages for arrival delays of 15 minutes or more and for cancellations on every listed route. Clean and analyze the raw data, identify seasonal, airline, and airport patterns, and briefly explain the forecasts and confidence levels. Deliver a forecast table, an executive-ready route risk map, and a comparative heat map, using only these files with no external research. Cowork: beautiful visuals Cowork Max produced a gorgeous risk map overlaid on the United States, with eight high-risk routes in red. Atlanta to Orlando, for example, carried a 33% chance of arriving 15 or more minutes late. It also produced delay and cancellation heat maps and an Excel workbook, taking 32 minutes and about $10 in credits. Cowork High was similar and only slightly less polished. Cowork Light took two minutes and $0.67, but its map lacked the US outline, its boxes overlapped, and its heat map was harder to read. Microsoft 365 Copilot, Researcher, and Analyst Microsoft 365 Copilot could not handle the ZIP file, and it also choked on the 11 or 12 extracted files. Dino's clever workaround was to put the files in OneDrive and give Copilot the share link, which worked perfectly. In about 16 minutes, it produced a single Excel workbook. Its 'map' was really a scatter chart plotting cancellation and delay percentages rather than geography, but its heat map was a sortable Excel table, which is more useful than a static image. Researcher asked clarifying questions about optimizing for predictive accuracy, how to express confidence, and report length, then produced the only Word report of the group, with more narrative. The Analyst agent, finally on a genuinely analytical task, created separate image files for the risk map and heat map; its map also was not geographic. The moment of truth We gave the actual August 2026 data to ChatGPT and Gemini and asked them to score each forecast. Microsoft 365 Copilot did best overall, ranking first for cancellations and second for delays, with a balanced normalized error of 36%. Cowork High and Cowork Max, with the most impressive visuals, finished at the bottom. All six Copilot options underestimated delays on every route. Their cancellation forecasts were also too optimistic: where they predicted roughly one cancellation per 100 flights, some routes saw closer to two. Most strikingly, a traditional forecast that simply used the average delays and cancellations from 2024 and 2025 was more accurate than any of the AI forecasts. One more lesson: when we first compared the results, ChatGPT flagged that our actuals file showed zero cancellations on routes with many flights and called it suspicious. It was right; our source extract had missed the cancellations. Read the thinking and output, not just the final answer. Interesting facts from the historical data More than half (51.4%) of Seattle to San Francisco flights historically arrived 15 or more minutes late, despite being a short hop. Direction matters: some routes had significant delays in one direction and not the other. LaGuardia to O'Hare and Boston to Washington Reagan accounted for 53% of all cancellations in the data set, despite representing only 19% of flights. The verdict Can Copilot predict the future? Not reliably, at least in our experiment. It produces plausible forecasts, and Cowork can produce visuals so polished that they draw you in, but as Dino put it, Cowork delivered more sizzle than steak. Sophisticated AI does not automatically beat simple statistical methods. If you need a forecast, start with traditional techniques, and if you use AI, slow down and verify the results against the data. Measure before you trust The most important lesson from this episode applies to Copilot investments generally: impressive output is not the same as measurable value. Organizations should base decisions about who gets which Copilot license on actual usage data, not on how good the demo looks. That is exactly what our Copilot Calibration service is built to solve. Using your existing Microsoft 365 usage data, we deliver person-by-person recommendations: who should get a Microsoft 365 Copilot license, which licenses should be reassigned, where Cowork makes sense, and where targeted training would pay off. It is fixed-price and requires minimal effort from your team.

Zoho is making a significant move in the evolution of business intelligence (BI), repositioning Zoho Analytics as an agentic analytics platform that connects business data, context, AI-driven insights, and actions. The goal is to move beyond reporting and dashboards to help businesses understand what their data means and use those insights to make decisions and trigger workflows. The announcement reflects a broader shift in enterprise AI. As organizations begin deploying AI agents, access to data alone isn't enough. Agents also need to understand how a business operates, how its metrics are defined, and how different pieces of information relate to one another. Without that context, even a technically correct answer can lead to the wrong business conclusion. Zoho is addressing this through what it calls a trusted intelligence layer, built into the Analytics platform rather than requiring a separate data platform or cloud data warehouse. Business context is the foundation Two of the most important elements of the update are the new semantic and ontology layers. The semantic layer provides a business glossary for AI, standardizing terminology, synonyms, instructions, and metric definitions. This matters because organizations often use the same terms differently across departments. Revenue, for example, may have different definitions depending on the business context. Giving AI a consistent understanding of those definitions should help improve the relevance and repeatability of its answers. The ontology layer goes a step further by modeling business entities and their relationships. This gives AI a structured understanding of how the business fits together, rather than treating every data point as an isolated fact. These capabilities are particularly important as businesses move from asking AI to retrieve information toward asking it to analyze problems, recommend next steps, and take action. The quality of those decisions depends on the quality of the underlying data and business context. Zoho says these capabilities are native to Analytics, can be configured without code, and support multiple semantic models for different business contexts. The platform also adds more than 30 data connectors to its existing base of more than 500, including integrations with Zoho applications, SAP, Twilio, Intercom, PayPal, and other enterprise applications and databases. The broader opportunity is to reduce the disconnect between business intelligence and the systems where work actually happens. From asking questions to taking action Zoho is also upgrading Ask Zia, its conversational BI agent, with a multi-agent architecture. A central agent coordinates six specialized agents covering data analysis, data modeling, visualization, data preparation, coding, and actions. The idea is to let users move between tasks without losing the context of the conversation. A user might begin by investigating a business performance issue, prepare additional data for analysis, and then build a workflow based on the findings. Organizations can also build custom agents grounded in their own data, semantic models, and knowledge bases. These agents can connect to external tools through the Model Context Protocol (MCP), web search, and other integrations. Zoho is also introducing agentic actions that allow agents to trigger workflows directly. Another important aspect is that Zoho is extending its analytics capabilities beyond its own interface. Through MCP, a command-line interface, and a Claude plug-in, developers and external agents can access Analytics capabilities. Organizations can also embed agents in their own applications and portals while retaining permissions and governance controls. This openness is strategically important. Businesses are unlikely to standardize all their AI activity on a single vendor's assistant. Making business intelligence available across agents, applications, and workflows gives Zoho a potential role in a broader AI ecosystem, rather than limiting Analytics to customers who work exclusively within Zoho's applications. Governance matters as much as intelligence Zoho is pairing these agent capabilities with controls for managing their use. A unified data catalog provides visibility into data, metrics, models, and analytics assets, including lineage and impact analysis. AI observability lets administrators see which agents ran, who used them, what data they accessed, and how they responded. Usage and cost tracking, sandbox environments, version management, and rollback capabilities provide additional oversight. These aren't secondary features. As agents gain the ability to access business information and initiate workflows, organizations need to understand what those agents are doing and establish boundaries around their actions. Governance will be a significant factor in determining whether enterprises are comfortable moving beyond AI experimentation. Zoho also says pricing for Analytics remains unchanged despite the extensive updates, with Zoho LLM credits now included in each plan. The Basic plan starts at $24 per month, billed annually, for two users, 500,000 rows, and 10,000 Zoho LLM credits. The next step in Zoho's AI strategy The announcement includes more than 50 additional capabilities, spanning visualizations, pixel-perfect reporting, scenario analysis, forecasting, and data science tools. But the more significant development is the direction of the platform. Zoho plans to expand its business context capabilities, introduce decision intelligence based on the outcomes of agent-triggered actions, and build toward a broader unified data platform encompassing data integration, preparation, analytics, data science, and agent building. The challenge will be execution. Semantic models and business ontologies require organizations to invest time in defining their data and processes. AI agents also need to demonstrate that they can deliver reliable outcomes, not just compelling demonstrations. And while unchanged pricing is attractive, businesses will still need to evaluate the cost of AI usage, integration, and ongoing management. Zoho's approach is notable because it treats business context and governance as foundational components of agentic AI, rather than adding agents on top of existing analytics tools. If the company can make that intelligence accessible across applications and workflows while maintaining accuracy and control, Zoho Analytics could become more than a BI platform. It could serve as a common intelligence layer for how businesses use AI to make decisions and get work done.

Zoom is making a bigger play for enterprise AI, with a strategy focused on turning conversations into context, and context into completed work. Its new Unified Context Layer sits at the center of that vision. Zoom’s message at Zoomtopia 2026 was clear: the company wants to be much more than a place where people meet, chat, and make phone calls. It wants to connect conversations across an organization with the context, AI agents, and workflows needed to get work done. The theme was “conversation to completion.” Conversations generate valuable information, but decisions and commitments often get lost across meetings, chats, customer calls, and business applications. Zoom believes it can connect those dots and use AI to help organizations act on what they already know. It’s an ambitious vision, but one that addresses a real problem. Enterprises have invested heavily in collaboration tools and AI assistants. The next challenge isn't simply getting AI to generate a good answer; it's giving AI the right information and the ability to do something useful with it. The Unified Context Layer Is the Big Story Some of the most valuable business information never makes it into a formal document. A customer raises a concern during a sales call, a team makes a decision in a meeting, or a support conversation reveals a recurring product issue. Those details can remain buried in transcripts, chat threads, and employees’ memories. Zoom’s new Unified Context Layer is designed to bring this information together. It combines general knowledge, documented information from files, and connected applications such as Salesforce, ServiceNow, Workday, and Jira, and dynamic context generated through ongoing conversations. Zoom connects these inputs into an evolving picture of people, topics, decisions, and relationships. The goal is to let people and AI agents work from synthesized, current context rather than having to search through raw transcripts and disconnected systems every time they need an answer. This goes beyond simply connecting a language model to a company’s documents. By incorporating information generated through live interactions and linking it to business data, Zoom aims to give AI a more complete understanding of what is happening across an organization. If it works as intended, that context could help agents do more than retrieve information: they could recommend or initiate the next step. But execution will depend on how accurately Zoom captures information, how well it connects systems, and whether the resulting context is reliable enough to support business decisions. Capturing Context Wherever Work Happens One of the most potentially impactful announcements was around My Notes. Zoom is extending My Notes beyond traditional online meetings to capture conversations across phone calls, in-person discussions, and third-party meeting platforms. Planned enhancements include a dedicated app for offline meetings and mobile calls, hardware partnerships for capturing in-person conversations, and Follow the Meeting, which generates a personal transcript and summary even when someone misses part of a meeting. Users will also be able to capture screenshots and associate them with specific moments in a transcript, preserving visual details that text alone might miss. These capabilities address a familiar workplace problem: employees often need to be in the right meeting at the right time to stay informed. Better capture makes it easier to catch up and follow through without relying on someone else to share notes or a recording. Zoom is also making its conversational intelligence available beyond its own interface. New Search and Ask APIs will let AI agents retrieve relevant information or generate answers grounded in Zoom data. Model Context Protocol (MCP) servers provide access to capabilities including meeting intelligence, scheduling, My Notes, and agentic search. Integrations with Claude, ChatGPT, Codex, and Microsoft 365 Copilot are part of the strategy, with Gemini support planned. This is a smart move. Zoom doesn't need every employee to do all their AI work inside Zoom Workplace. Making its intelligence available in tools people already use could extend the value of its platform. The trade-off is that Zoom will need to demonstrate why its context and services remain valuable when employees increasingly interact with AI through other vendors’ interfaces. Customer Experience: Moving Beyond Insights to Action The conversation-to-completion strategy is particularly tangible in Zoom’s CX portfolio. New customer directory and profile capabilities are designed to create an evolving view of customer interactions across virtual agents, live agents, and Zoom Phone. SourceHub connects agents to third-party data, starting with Salesforce Service Cloud, allowing them to view and update contacts, cases, and opportunities without constantly switching between systems. Zoom CX Insights is also moving beyond traditional contact center reporting. It analyzes voice, chat, and email interactions alongside operational data to help businesses understand why customers are contacting them, what drives longer calls, and where agents may need support. New capabilities include recommended actions, scheduled reports, and executive briefs. Customer Workflow Orchestrator takes the idea further by connecting the contact center to CRM, IT service management, HR, and other operational systems. AI-powered reasoning can determine the next step in a workflow and resume it when a customer response or approval arrives, even hours or days later. That distinction matters. Creating a support ticket isn't the same as resolving a customer’s problem. Orchestrating work across systems offers a path toward closing that gap. The potential value is less manual work, faster resolutions, and fewer handoffs. The test will be how consistently these tools deliver results in complex, real-world workflows. Zoom is also extending Virtual Agent to physical kiosks, where customers can ask questions, complete routine tasks, or get help in environments such as retail and healthcare. If necessary, the interaction can be handed off to a live agent through Zoom Phone or Zoom Contact Center. Pricing in the Agentic AI age has become challenging for all vendors trying to figure out the optimal pricing model (consumption and usage, outcome-based, etc.). Zoom announced a new spend-based pricing model, with one unified spend pool across the ZCX portfolio. Zoom notes that the one annual commit covers the entire ZCX portfolio, and the commitment draws down across any ZCX product. This makes it easier for customers to use what they need, when they need it, without worrying about purchasing the right number of licenses for each product. I spoke with Chris Morrissey, General Manager of Zoom CX, about the announcements, the role of CX in Zoom’s broader portfolio, and how the Unified Context Layer can make customer interactions more personalized. I also spoke with Ram Rajagopalan, Head of Product – AI for Zoom CX, about the new kiosk virtual agent, as well as AI Architect, which uses AI agents to help build and test contact center agents. AI Architect is used to generate production-ready voice and digital AI agents from a simple prompt. I was particularly interested in how Zoom helps users create virtual agents. In this video, Ram demonstrated how AI Architect can simplify the process of building virtual agents. For businesses, the potential value is less manual work, faster resolutions, and fewer handoffs between systems and employees. From Productivity Features to Business Execution The same strategy runs through Zoom’s workplace and revenue announcements. My Notes workflows will let users send meeting summaries and action items to applications such as Asana and Slack using templates or natural-language instructions. ZoomMate team agents will coordinate specialized agents working on different parts of a project and combine their work into a single deliverable. For revenue teams, Agentic Workflow Builder will provide templates and custom agents to help recover stalled deals, improve forecast accuracy, and reduce CRM administration. Signal-based orchestration is designed to trigger workflows when buying signals appear, helping sales teams respond faster. Zoom Events is also getting AI capabilities. Content Studio will turn event recordings into clips, posts, and images, while an event setup agent will build configurations from prompts, agendas, or spreadsheets. The common thread is reducing the work between knowing what needs to be done and actually doing it. For enterprises, the bigger opportunity isn't just saving a few minutes on individual tasks; it's improving how work moves across teams and systems. Trust Will Be Critical to Adoption As AI agents gain access to more information and the ability to take action, security and governance become essential. Zoom is expanding capabilities, including deepfake risk detection, identity verification integrations, sensitivity labels, AI guardrails, and centralized AI management. Administrators will have greater visibility into active agents, the information they access, and their activities, with the ability to enforce policies and shut down agents operating outside defined boundaries. Zoom Compliance Manager is also being extended to My Notes and Zoom Canvas to help regulated organizations manage retention, monitoring, and audit requirements. These are more than compliance features. If AI agents can access confidential information, update business records, and trigger workflows, organizations need to know what those agents can see and do. The challenge will be maintaining consistent controls across Zoom products, connected applications, and external AI platforms—and establishing accountability when automated actions go wrong. I spoke with Alex Shockley, President and COO of Journey.ai, about its partnership with Zoom and capabilities for secure transactions, user authentication, and handling regulated data. The Bottom Line Zoomtopia 2026 signals a broader ambition for Zoom: building on its strength in communications to become a platform (or even operating system) that connects conversations, enterprise knowledge, and business execution. The Unified Context Layer is the centerpiece. Better meeting capture, cross-platform notes, customer profiles, AI agents, and automated workflows become more valuable when they share a common understanding of what is happening across the business. Zoom has two advantages. It sits across a wide range of conversation surfaces -meetings, phone, team chat, contact center, events, and revenue intelligence - and it is making its context available to external AI tools through MCP servers and Search and Ask APIs. Zoom is also leaning into open source with Onyx, a platform for interacting with different large language models and searching more than 50 internal data sources while preserving enterprise permissions. There is still plenty to prove. Integration quality, the accuracy of the context layer, measurable ROI, and customer confidence in AI governance will determine how much of this vision translates into business value. Many companies have figured out how to capture conversations. What happens next is what determines whether AI delivers value. Zoom wants to help businesses act on those conversations, carry work across systems, and reach meaningful outcomes. That makes “conversation to completion” more than a product tagline; it's Zoom’s attempt to define its role in the next phase of enterprise AI.

Formal training often means hours of content, but sometimes you just want to reinforce a policy quickly and memorably. For Episode 7 of Calibrating Copilot, Dino Caputo and I asked Copilot and Cowork to turn a policy document into a short, interactive training experience. The prompt and the policy Our prompt: using Contoso's attached AI policy, create a five-minute interactive training experience with branching scenarios, feedback, and a final score. Dino had Copilot find a publicly available AI governance policy, which he anonymized for Contoso (Microsoft's favorite made-up company) and supplemented. Importantly, the policy encourages appropriate AI use rather than simply banning it, and it names specific tools rather than speaking only about Microsoft AI in general. Microsoft 365 Copilot (Auto), take one: an agent nobody asked for On my first attempt with M365 Copilot set to Auto, it decided to use Agent Builder and created a 'Contoso AI Policy Coach' agent in about two minutes. That sounded promising until I started the training. It took 1 minute and 8 seconds to present the first question, the experience was plain multiple choice in a chat, and there were only three questions. It also claimed it could adapt to the learner's role, then admitted it had no evidence it actually did. A fail. Cowork Medium: fast and cheap Dino's Cowork Medium run took just 1 minute and 25 seconds and $0.72 in credits. It created an attractive dark-themed app with an introduction, a timer, eight scenarios, a running score, feedback that cited the relevant section of the policy, and the option to retake the training. We gave it five stars for accuracy and four for completeness, mainly because others added a bonus print feature. Microsoft 365 Copilot with Opus 5.5: excellent I went back to M365 Copilot and explicitly selected Anthropic's Opus 5.5 model. This time it planned a single, self-contained HTML file, reasoned for 2 minutes and 33 seconds, and delivered an app every bit as good as Cowork's, including a print results button that actually worked. Five out of five. Cowork Max: beautiful, but we could not share it Cowork Max built a full React app themed around 'a day in the life' of using AI at Contoso, with 11 questions, four branches, and the most descriptive results dashboard of the group. But it published the app as a Power App, and Dino could not share it with me. Troubleshooting the issue in Cowork burned roughly another 700 credits, almost as much as building the app. Creating it took 31 minutes and $7.90; by the end, the total was about 1,490 credits, around $15, plus an hour and a half of our combined time. Five stars for accuracy, three for completeness. Microsoft 365 Copilot (Auto), take two Using the exact same prompt and settings as my failed first attempt, M365 Copilot on Auto this time used Python to generate an HTML app in just 39 seconds. The result was nearly identical to the Opus version, which makes me suspect Auto picked that model. Five out of five. Dino also tried M365 Copilot with the GPT-6.1 Sol model. In about two minutes it produced a strong app with source notes and a download-your-results option (a CSV clearly labeled as not a compliance certificate), although it did not include a timer or running score. Five stars for accuracy, four for completeness. Key lessons If you want a standalone HTML app, say so. Leaving it open invites creativity, which may mean an agent or a Power App you did not want. Build in Cowork, iterate in Copilot. Dino took a Cowork-created app into M365 Copilot and made changes, including colors, without spending any more credits. If Copilot gives you a link that is not clickable, you are not losing your mind. Tell it, and it will usually provide a working one. AI is a meticulous reader. It flagged that the policy's effective date still read 'insert date here' and that two different governance contact emails appeared on different pages. AI is probabilistic. Because M365 Copilot is all-you-can-eat, trying again is free. Our recommendation: for interactive training like this, you do not need Cowork. Microsoft 365 Copilot produced excellent results at no additional cost, especially when you specify a standalone HTML app or select a specific model. That is the opposite of Episode 6, where only Cowork could create a usable explainer video. The right Copilot, used the right way After seven episodes, the pattern is clear: you need the right Copilot, used the right way, to get value. Spending $15 and an hour and a half on one app may seem minor, but multiply that across a thousand employees using Copilot the wrong way and the impact on budget and productivity is significant. Visit enableuc.com/calibration for a summary of every Calibrating Copilot episode and the lessons we have learned so far. That is exactly what our Copilot Calibration service is built to solve. Using your existing Microsoft 365 usage data, we deliver person-by-person recommendations: who should get a Microsoft 365 Copilot license, which licenses should be reassigned, where Cowork makes sense, and where targeted training would pay off. It is fixed-price and requires minimal effort from your team.

Authentic natural beauty and artificial intelligence came together at NiCE’s annual analyst summit, held on the beautiful island of Maui. Against a backdrop of palm trees and crashing waves, the discussions focused on AI, particularly agentic AI, and how it is changing the customer experience (CX) landscape. The theme was “Huliau,” Hawaiian for turning point. And our industry is certainly at one. In just a few short years, we’ve gone from the introduction of generative AI and large language models (LLMs) to agentic AI and AI agents, and more recently, the emergence of personal AI agents. At each stage, NiCE has embraced the change, transforming itself from a CCaaS provider focused on customer service into a broader CX AI platform provider. The company is looking beyond the contact center to other areas of the organization that touch customers, including sales and marketing. Cognigy Changes the Game NiCE’s acquisition of Cognigy is central to this transformation. It gives the company a native CX AI platform with conversational and agentic AI capabilities built in to the CXone platform, positioning NiCE to address a much broader portion of the customer journey. NiCE has a history of smart acquisitions, with inContact being one of the most notable. But Cognigy may prove to be its most consequential acquisition yet, helping propel NiCE ahead of other CCaaS and CX players in AI and agentic AI. As CEO Scott Russell noted, “With Cognigy, we made a bet on AI. Cognigy changed the nature of who we are as a company. The capabilities Cognigy brought into NiCE are at the core. We made a bold bet, and it was successful.” NiCE Cognigy acts as the intelligence layer of the NiCE platform, with agentic AI providing “reasoning across channels, orchestrating actions, and powering AI agents that can move from conversation to resolution.” While many vendors are bolting on AI, partnering with third parties, or providing an AI layer without the underlying infrastructure (yes, I mean you, Sierra), NiCE has assembled the pieces of a more complete platform. In my view, the platform is NiCE’s biggest competitive advantage, or “moat.” It brings together: Agentic Experience Automation: AI agents that resolve customer needs instantly and proactively, including agents for customer service, sales, and marketing, along with knowledge management and agentic analytics. Engagement Orchestration: Orchestration across channels, people, AI, and systems, including omnichannel routing, outbound engagement, voice and digital engagement, and a newly designed agent workspace. Workforce Empowerment: Bringing humans and AI together as one workforce, with workforce management, quality management, performance management, and recording and compliance. CX AI models and layer: The AI capabilities that support the broader platform. Customer Engagement Data: The data foundation for connected customer experiences and intelligence. NiCE Cognigy’s conversational and agentic AI layer underpinning the entire platform. AI isn't bolted on or treated as a separate, standalone layer. Instead, it is part of the overall platform, providing a connected operating model. (Cognigy can also be deployed separately by companies that want to integrate its AI capabilities with their existing CCaaS systems.) The advantage of the integrated approach is that the same AI capabilities operate across the platform, enabling more connected and compounding intelligence and a more seamless experience. NiCE Cognigy offers two deployment paths: a fully integrated experience within CXone for customers seeking one connected operating model, and a best-of-breed agentic AI option that integrates with existing CCaaS, CRM, and enterprise systems. There are clear benefits to having Cognigy natively integrated into the architecture. Customers get a common user experience and shared data platform, while AI agents can be built using customer data and trained directly on interaction data. Data, channels, and orchestration are available out of the box, reducing the need for customers to build connectors or MCP integrations themselves. The result should be faster time to value and lower implementation risk. Orchestrating Humans and AI A key focus area at the event was the changing workforce. With agentic AI, humans and AI will increasingly operate as one workforce, resolving customer needs across the enterprise. This goes well beyond handing an interaction from an AI agent to a human agent. It requires common rules, tools, and governance for both human and AI workers, along with orchestration across the entire customer journey. NiCE’s latest architecture added key capabilities to the platform, focusing on humans and AI working together, along with learning loops across the entire experience: NiCE AI Agents: serving as the execution layer, the AI agents understand intent, take action across enterprise systems, and resolve customer needs across voice and digital channels. Agentic Engagement Plane: the orchestration layer, designed to authenticate, mediate, route, and govern interactions as customers, employees, personal AI agents, and enterprise AI agents increasingly interact with one another. NiCE Guardian Agent: for providing governance, Guardian monitors AI and human actions in real time, applying compliance guardrails and detecting risk. Agentic Analytics: moving beyond reporting what happened to identifying what should happen next, feeding signals back into agents and workflows for continuous improvement. Shared Cases: provide clear ownership and tracking from request to resolution Experience Memory provides persistent customer and interaction context based on understanding of history, preferences, outcomes, and intents. This memory persists across every interaction Together, these capabilities support a learning loop in which the platform learns from human experts, turns that knowledge into AI agents, and continually improves them. NiCE is evolving toward a fully integrated, self-improving process: identifying automation opportunities, creating agents, evaluating them against synthetic customers, deploying them across channels, and using the resulting insights to improve subsequent iterations. Every conversation can contribute to what the platform learns and how it handles future interactions. I spoke with Jeff Comstock, President of CX Product & Technology, about the platform and the learning loop, and what it means for businesses. Personal Agents: The Future of CX? One of the hottest topics at the summit was personal AI agents, such as Meta Muse, Instinct, OpenAI Dots and Perplexity Comet. These personal agents have emerged relatively recently, but adoption has been rapid, with Muse reaching the top of the U.S. App Store. Personal AI agents point toward a fundamental shift in how consumers interact with businesses. Today, people commonly use AI to find information, get recommendations, or help complete a task. Personal agents take this a step further: they can act on a person's behalf, interacting with websites, applications, and potentially businesses to accomplish a goal. They may use voice or text to make reservations, cancel subscriptions, and handle other tasks. For contact centers, the implications could be significant. The entity initiating an interaction may no longer be a person typing into a chat window or speaking to an agent - it may be another AI system acting on that person's behalf. That raises questions about customer identity and authentication, journey design, service automation, agent productivity, security, and even how contact center performance should be measured. Consider a personal agent tasked with resolving a delayed shipment, an unresolved support case, or a medical appointment. If the business doesn't deliver the desired outcome, the agent can keep trying until it gets the result it was instructed to pursue. Instead of a frustrated customer giving up after waiting in an IVR queue or navigating a website, businesses could face persistent, automated follow-up. That could affect IVR queues, call volumes, and other aspects of contact center operations. Businesses will need to prepare for a world in which contact centers serve two types of customers: humans and AI agents. Contact centers will need to accommodate machine-initiated interactions, authenticate personal agents appropriately, establish security and governance controls, and determine what these agents are authorized to do. We're still in the early days of personal agents, but businesses should already be thinking about how to prepare. During the customer panel at the summit, David Mayer, Associate VP, Data, Analytics, Learning and Quality Strategy at Hyatt, noted that online travel agencies are beginning to use personal agents. Mayer told the audience, “Personal agents are coming and we have to prepare. Security and authentication are essential, and we have to verify if it’s a real person and customer behind the agent and not a bad actor.” Rather than waiting for customers to start asking for support for these interactions, NiCE is taking steps to prepare for this next stage of CX AI. The company is participating in the development of the Personal Agent Protocol (PAP), an open standard being developed by Meta and Sierra, along with several industry partners, to define how personal agents interact with businesses. NiCE also offers the CXone AI Agent Gateway, which Neeraj Verma, Vice President, Product Management, describes as the front door for personal AI agents contacting a business. It detects when a caller is an agent, verifies who it represents, limits what it is allowed to do, and gives the business visibility into the conversation. These efforts demonstrate that NiCE is thinking beyond what customers need today to what they will need as personal agents become more common. Personal agents also introduce a new dimension to workforce management and engagement. Customers and workers can be human or AI, and interactions may involve multiple combinations: a customer interacting with an AI agent, an AI agent escalating to a human, a consumer's personal agent interacting with a human agent, or eventually, AI agents interacting directly with one another. As Russell pointed out, brands no longer own the customer journey. The AI agent will never give up and keep on trying. This presents a number of issues, including how enterprises will control interactions and the customer journey. Based on my discussions at the summit, NiCE has clearly been thinking about these scenarios and how to prepare for this emerging environment. NiCE CTO Kevin Lee and I discussed personal agents and some of the considerations businesses will need to address. (Note: AI was used to remove the background noise and music) Creating Our Own AI Agents The highlight of the analyst event was when the analysts were let loose to create our own AI agents for “NiCE Air.” We began by using NiCE Agentic Analytics to identify automation opportunities and determine where AI agents could be deployed. The system did all the work in analyzing real customer interaction data to identify what to automate. We then built our own AI agents both from scratch and from human-to-agent engagements to run on chat and voice. We were able to see first-hand how much intelligence is built into the system, and how easy it is to create the agents. We gave them names, personalities, skills, guardrails, rules, and more. I never thought I’d be able to build my own AI agent, but the tools NiCE Cognigy provides make it easy. The Changing Boundaries of CX In this video discussion, Andy Traba, Vice President, Product Marketing, and Jennifer Wilson, Director, Product Marketing, shared their key takeaways from the summit, including the rise of personal agents, the changing boundaries of CX, and NiCE's move beyond the contact center. A recurring theme was that value is increasingly accruing around platforms rather than individual point solutions. Traba also discussed the learning loops that enable each interaction to make the next one better. Wilson highlighted the demonstrations presented at the event, including persona-based experiences in which the workspace changes according to different users and their priorities. The demos illustrated how humans and AI agents can work together to achieve the desired customer outcome. Wilson also discussed the Agentic Engagement Plane as a unified orchestration layer for managing the customer journey as a whole. Traba noted that the different workers share a Case Manager and Experience Memory, reinforcing the idea that work and context need to carry across people, agents, and channels. Looking Ahead One challenge NiCE - and other CCaaS vendors - will face is expanding into other areas of CX, particularly sales and marketing. As discussed throughout the summit, AI expands NiCE's strategic role beyond customer service. The company aspires to move from being a software partner to providing a broader CX AI operating model. Russell described the bet NiCE is making as AI across the full customer experience, encompassing every engagement, workflow, worker, and outcome, from first contact to final resolution and from the front office to the back office. “Customer service is the cornerstone, but we do so much more,” he explained, while acknowledging that NiCE doesn't claim to have every piece. That makes an open architecture essential. Expanding into sales and marketing also changes the buying dynamic. NiCE is increasingly dealing with C-suite executives and technical decision-makers, in addition to traditional contact center buyers. As Russell explained, AI agents touch models, data, integrations, and security, so the Chief AI Officer, platform team, and AI engineer now make decisions alongside operations. The economic model is changing as well, with monetization expanding beyond seat licenses toward intelligence and automation outcomes. Working with partners such as Accenture will help NiCE engage the C-suite in these broader discussions. But the company still needs to demonstrate that it is more than a CCaaS provider. In fact, the term “CCaaS” barely came up in the presentations. More customer wins beyond customer service and the contact center will be important in establishing NiCE as a broader CX vendor. Competition is another consideration. NiCE must contend not only with established CCaaS and CX players, but also with newer AI-focused competitors such as Sierra, Decagon, and Fin, which was recently acquired by Salesforce. Sierra has been successful in part because of its founders' personal relationships, which can help secure meetings with senior executives and enterprise buyers. However, Sierra only provides the agentic AI layer rather than the kind of end-to-end platform NiCE is building. NiCE's integrated platform is a differentiator, while its option to sell Cognigy as an agentic AI layer that integrates with existing CCaaS and CRM systems gives customers another path - one similar to the approach taken by standalone AI providers. That flexibility should help NiCE compete with point solutions as well as full-platform vendors. Ultimately, NiCE is making a bet on AI across the entire customer experience: every engagement, workflow, worker, and outcome, from first contact to final resolution, front office to back office. As Russell acknowledged, NiCE doesn't claim to have every piece. Instead, it is “betting on one platform that serves most of the experience, and an open architecture that brings in the best capability from everywhere else.” In my view, NiCE's AI-first platform provides a strong foundation for a workforce made up of humans and AI agents. Its integrated approach, combined with its work on personal agents, also positions the company to address the next stage of CX. The challenge now is to turn that vision into demonstrable customer outcomes and establish NiCE's position beyond the contact center.

For Episode 6 of Calibrating Copilot, Dino Caputo suggested a use case I had not considered: creating a short explainer video. We tested Copilot, Copilot in PowerPoint, and Cowork at Light, Medium, Extra High, and Max. The prompt We asked for a 90-second explainer video describing the different Copilot licenses, Copilot Chat, Microsoft 365 Copilot, and Cowork, including the capabilities and limitations of each. It should be fun, engaging, and animated, with a natural-sounding talk track and captions in case someone watches without audio. We kept it short and let each Copilot interpret the rest. Copilot: voiceovers are the problem Copilot (using Auto) could not complete the task. Its text-to-speech tools were blocked, and the only voice available was a robotic eSpeak voice that it rightly said did not meet our brief. Its suggestion: it would write the talk track, Dino would record it, and it would merge the two. Dino recorded the narration in Audacity, but because Copilot cannot accept audio uploads directly, he had to embed the audio in a PowerPoint file to get it back to Copilot. It then synced his voice to the visuals. It worked, but it took a lot of back and forth. Copilot's experimental Advanced Reasoning mode wrote Python (as it always seems to) and produced a video that was precisely 89.98 seconds long. Unfortunately, it used whatever narrator voice was available on the device, sounded like a Speak and Spell, and looked just as dated. Copilot in PowerPoint, with help from Cowork Copilot in PowerPoint asked several clarifying questions and then produced an auto-advancing, captioned slideshow. It could not generate audio, and it could not invoke Cowork. So I handed the deck to Cowork and asked it to add voiceovers using the narration in the speaker notes. From there, PowerPoint's File > Export to video created an MP4. The animations were basic, but the advantage is that you can easily edit the slides when the content changes. Total time was about 10 minutes; the Cowork portion took 6 minutes and 42 seconds and $0.82 in credits. Three stars for both accuracy and completeness. Cowork Light: you get what you pay for Cowork Light finished in 2 minutes and 34 seconds for just $0.47, but it created an HTML page with JavaScript rather than an MP4. It relied on the browser's built-in voice, so it sounded different on every device: male on one playback, female on another. Captions drifted out of sync, narration was cut off at slide transitions, and it mispronounced words, at one point wishing viewers 'happy copy loading.' On my phone, Cowork became 'cow-work.' Cowork Medium and above: a major jump in quality Moving to Cowork Medium changed everything. The voiceover sounded natural, the transitions were smooth, and the content was accurate, all from a single prompt and at a small fraction of the cost of Max. We gave it five stars for accuracy and four for completeness. Cowork Extra High was terrific. It used a memorable analogy: Copilot Chat hands you a recipe, Microsoft 365 Copilot cooks with what is in your fridge, and Cowork caters the whole party. It deliberately omitted pricing so the video would not go out of date and added a small disclaimer noting that licensing details were current as of October 2026. Frankly, it explained the differences far better than I managed when my wife asked me. Cowork Max took 53 minutes and $14.43 in credits and earned five stars across the board. It framed the three options as video game levels, complete with sound effects. Its thinking showed remarkable attention to detail: it calculated that a natural narration pace for a fun tone is roughly 220 to 240 words, and it adjusted speech speed on individual slides by 2% to 13% to hit the timing exactly. The only flaw we noticed was a slightly odd pause in 'Microsoft 365.' Dino also asked for a hand-sketched line art style in one run, a good illustration of how easily you can steer the visual look. Is Max worth it? Max was not five times better than Medium, and Extra High was not two or three times better either. Dino felt Extra High offered the best bang for the buck. My view: at under $15, even Max is a steal. Producing something comparable through an agency would easily cost more than $1,000, and since you can start the task and walk away, the extra time matters less. Our recommendation If you need an explainer video, use Cowork at Medium or above. Copilot cannot yet create voiceovers on its own, Copilot in PowerPoint is a decent option if you want something easily editable, and Cowork Light is not worth the savings. A product demo, onboarding overview, or internal announcement could all be produced this way for a few dollars. Some tasks need Cowork. Many do not. This was a use case where only Cowork delivered real value. In other episodes, the included Microsoft 365 Copilot was the clear winner. Knowing which tasks, and which people, actually need Cowork and AI credits is the key to getting a return on your Copilot investment. That is exactly what our Copilot Calibration service is built to solve. Using your existing Microsoft 365 usage data, we deliver person-by-person recommendations: who should get a Microsoft 365 Copilot license, which licenses should be reassigned, where Cowork makes sense, and where targeted training would pay off. It is fixed-price and requires minimal effort from your team.

Unified communications has traditionally been built around the needs of the desk worker. Give employees a phone, softphone, messaging, meetings, presence, and a collection of collaboration tools, and then try to make those capabilities work for everyone else. Mitel is taking a different approach with its new Mitel Workforce Experience (Mitel WX). Rather than starting with the communications client and trying to make it work across the organization, Mitel is starting with the employee’s role and the workflows associated with that role. That distinction could be particularly important as enterprises look beyond the traditional knowledge worker and try to bring better communications capabilities to the large - and often underserved -frontline workforce. UC Isn't One-Size-Fits-All During a recent analyst briefing, Mitel executives talked about what they see as a fundamental problem with UC - most UC platforms have historically been designed primarily for knowledge workers sitting at desks. That works reasonably well for employees who spend their days in desktop applications, but it doesn’t necessarily work for a nurse, field service technician, hotel employee, retail associate, or manufacturing work who may spend most the day away from a computer, who represent 80% of the workforce. While the communications requirements of these workers may be dramatically different from deskbound workers, many organizations have essentially given all of them the same communications -tools, which has led to implications like shadow IT. . Mitel’s answer is a role-based communications framework that can be tailored to the job an employee actually performs. Mitel WX takes a role-based approach instead. Administrators can create different roles and templates and determine which capabilities are presented to different groups of employees. The demo presented to the analysts illustrated how this works. Users can have different widgets and applications available depending on their role, while administrators control which capabilities are provisioned. Employees can then customize aspects of their own interface without having to navigate functionality they don't need. This is more than a user-interface change. If done well, it could make communications technology more relevant to employees who don't spend their days sitting in front of a PC. That includes what Mitel describes as the frontline or deskless workforce, which represents a significant portion of employees in industries such as healthcare, manufacturing, retail, hospitality and the public sector. The Hybrid Architecture May Be Just as Important For Mitel customers, however, I think the architecture may ultimately be as important as the user experience. Mitel has been an early leader in the hybrid architecture approach, which is now seeing a resurgence and is being embraced by companies of all sizes in the AI era, where privacy and security are paramount for both regulated and non-regulated businesses. Mitel has a large installed base, and many of those customers aren't necessarily looking to rip and replace their communications infrastructure. They may have significant investments in existing Mitel platforms, while also having good reasons for keeping some communications capabilities on-premises or in a private cloud. Mitel WX is designed with that reality in mind. The new customers can work with existing Mitel call-control platforms, allowing customers to introduce the Mitel WX experience without immediately replacing the underlying infrastructure. The architecture also supports combinations of on-premises, private cloud, and public cloud deployments, including multi-domain environments, providing customers some flexibility in how they modernize. The cloud migration conversation has often been framed as a binary choice: move everything to the cloud or stay on-premises. That's increasingly unrealistic for large enterprises, particularly in regulated or mission-critical environments. Mitel is instead betting that customers want to determine where different pieces of their communications environment should live. For existing Mitel customers, that could provide a more practical modernization path. Bringing Communications Into the Workflow Another interesting aspect of Mitel WX is its emphasis on workflow integration. Mitel is positioning the Mitel WX less as another standalone communications application and more as a platform that can incorporate the applications and information employees need to do their jobs. The company demonstrated the ability to integrate applications such as CRM and EHR systems and described a marketplace and SDK approach that can allow partners and customers to develop additional widgets and integrations. This concept isn't entirely new, as most UC vendors talk about integrations. The difference is that Mitel is tying those integrations to the employee's role. The objective isn't necessarily to put every possible application into everyone's communications client. It's to give a particular employee the applications and information that are relevant to the job. That is an important distinction as organizations try to reduce application overload rather than adding another application to the pile. Where AI Fits In AI is also part of Mitel's longer-term strategy. Mitel executives talked about voice as a natural interface for AI, particularly for frontline workers. That's an interesting area to watch because much of the current enterprise AI discussion still revolves around knowledge workers interacting with AI through a screen. For a worker in a hospital, factory, hotel or field environment, voice may be a much more practical interface. Mitel isn't positioning WX simply as an AI product, and I think that's the right approach. The more interesting question is how AI becomes part of the role-based workflows over time. For example, an employee shouldn't necessarily have to open a separate AI application to get assistance. If AI can be embedded into the communication and workflow tools the employee is already using, it becomes much more useful. What This Means for Mitel Customers The big question for Mitel customers will be whether Mitel WX can deliver on the promise of modernization without forcing them to abandon what already works. That is particularly relevant for organizations with complex communications environments and significant investments in Mitel infrastructure. If Mitel WX can provide a modern, role-based employee experience while allowing customers to maintain existing call-control systems and choose where workloads are deployed, it gives Mitel customers another option besides a wholesale platform replacement. There are still questions around the breadth of integrations, the depth of the workflow capabilities and how the platform will evolve as Mitel adds more AI functionality. Those will be important factors in determining how differentiated WX ultimately is. But I think the basic premise is sound. The future of enterprise communications isn't necessarily about finding one UC application that works equally well for everyone. It may be about delivering the right communications experience for the job an employee actually performs. That's a somewhat different way of thinking about UC, and for Mitel customers with large frontline workforces and complex hybrid environments, it could be a useful one.

Five9 held its annual analyst summit in a new location this year: Seattle, WA. While some of us initially assumed the location was chosen because Seattle is such a cool city (which it is), the real reason was more practical: Five9’s new CEO, CMO, and CTO all reside there. Yes, there’s a new leadership team at Five9, along with new energy, positioning, branding, and more. While some things - and people - remain the same, there are definitely changes afoot. As CEO Amit Matradhas explained, the new leadership team and organizational changes are designed to remove internal friction and enable Five9 to build, think, and react faster. One of my favorite quotes from the event was: “We’re building for scale, for the next evolution. We’re using our past as an anchor but not letting it be an anchor.” AI and the Evolution of CX The growing role of AI in customer experience cannot be understated. Five9’s new tagline is “Smarter CX with every interaction,” highlighting how customer interactions feed into AI’s knowledge base in a continual learning loop for both human and AI agents. According to Jay Lee, CMGO (Chief Marketing and Growth Officer), smarter CX with every interaction means faster resolutions, better outcomes, stronger relationships, and deeper insights. As Matradhas explained, Five9 sees this as a whole new system where AI and humans come together to create new outcomes. As he put it, “Based on customer feedback, we’re on a direction to capitalize on this new wave.” When GenAI first came on the scene, many people - particularly on Wall Street - assumed that the role of contact centers and contact center agents would decrease, with AI taking over customer service interactions. That hasn’t happened. Matradhas noted, “AI was supposed to shrink contact centers, but it’s still growing, with new jobs and agents added.” The implication is important: AI isn’t replacing the value of the contact center; it’s expanding it, with Five9 putting the market expansion opportunity at $50 billion by 2029. Matradhas attributed this in part to a simple dynamic: when service gets easier and better, people engage more, not less. In today’s contact centers, AI and humans are increasingly working hand in hand, with each doing what they do best and the right resource being used for the right task. It’s not necessarily an either-or decision. It’s about finding the optimal combination of humans and AI. This is what Five9 refers to as “Humantic CX” - a combination of Human + Agentic. Humantic CX is Five9’s vision of a unified approach in which AI is a constant partner to human work, rather than a replacement for it. Humantic CX Humantic CX was clearly the star of the show, with demos - unfortunately under NDA - helping clarify what Five9 means by the term. Most CX vendors have been talking about “hybrid AI” or “hybrid workers,” generally referring to AI agents handing interactions, along with relevant context, off to human agents when necessary. Five9’s Humantic CX concept takes this a step further. Five9 describes Humantic CX as providing: Constant presence, where AI is always part of the interaction Elevated agent, enabling humans to focus where they matter most Compounding loop, where every interaction makes the next one smarter Optimal mix, with the right balance of human and AI for every situation The idea is to combine humans, AI agents, and channels in a single conversation, without losing context. Orchestration enables smooth handoffs during the interaction, while backend workflows can help resolve the underlying issue. It also brings together customer memory, domain knowledge, and real-time awareness. Five9’s ambition is to be “the customer outcome partner for the Humantic CX era, harnessing and orchestrating the best of humans and AI.” The company sees an open, unified CX platform as the foundation for achieving that goal - a platform that makes AI real, where voice and digital both play important roles, and where every interaction can contribute to a CX that compounds over time. Five9 intends to pursue this opportunity by focusing on several areas: Where humans will continue to exist and where humans and AI will work together The midmarket and above Key verticals and regulated industries, including financial services and healthcare, where complexity and compliance are essential Specific geographies, including North America and Europe, with a particular focus on markets such as the UK and Germany Differentiation With so many CX vendors now offering AI agents, AI studios, and related capabilities, differentiation is becoming increasingly difficult. Five9 believes its differentiation comes from its network, platform, and Humantic CX approach. Five9 is a global connected voice provider with a global telco network, providing global connectivity, voice analytics, and reliability. But the company is also placing significant emphasis on what it calls the interaction layer - the place where channels, orchestration, memory, intelligence, and customer interactions come together. As Matradhas noted, “If we can connect channels, provide orchestration, and pull memory into interactions, we’ll succeed.” That interaction layer could become an important battleground as AI agents become more capable and increasingly participate in customer interactions. Putting Humantic CX Into Context I spoke with Matt McGinnis about what Humantic CX is all about and how Five9 differentiates in an increasingly crowded market. In this video interview, McGinnis explains how Humantic CX brings humans and AI together to create synergies across the customer journey and customer experience. He also discusses how Five9 is differentiating through its single platform and by bringing together data, intelligence, analytics, and other capabilities. McGinnis also discusses how Five9 sees AI agents being deployed in the enterprise, including the importance of trust, security, and guardrails. He explains how Five9 is approaching these issues for customers in regulated industries and how the company sees trust as important for both contact center workers and consumers. I also spoke with Five9 President Andy Dignan about some of the key messages from the event, including: The speed required to operate in the AI era Five9’s focus on specific segments and geographies Executing on AI And, of course, Humantic CX In this video, Dignan also discusses customers’ readiness for AI and the growing importance of trust and governance, particularly in the regulated industries Five9 is targeting. He also explains Five9’s vertical approach and how it extends across marketing, product, go-to-market, sales motions, services, partner relationships, and more. Closing Thoughts Five9’s focus on Humantic CX definitely resonated with many of the analysts at the summit. The concept reflects where the market appears to be heading: toward a model in which humans and AI work together rather than AI simply replacing humans. Five9 presented a couple of interesting use cases, but the company will need to provide more examples to help businesses understand the practical value of Humantic CX and, more importantly, the measurable outcomes it can deliver. The concept is interesting; the next step is demonstrating what it means in practice. Five9 also recognizes that there are more and more players entering the space and that frontier AI companies could pose a challenge in the future. The company’s focus on the interaction layer, its open platform and ecosystem approach, customer and conversation insights, and its global network capabilities could help it stand out. Companies such as Sierra and Decagon are gaining both market share and mindshare, particularly among the C-suite. Five9’s opportunity is to position itself as a more complete solution rather than just one, albeit important, piece of the customer experience puzzle. Finally, Five9’s new executive team - Niranjan Vijayaragavan (CTO), Rob Hornish (CSO), Sevn Linsmaier (EVP of Transformation & Strategy), and Jay Lee (CMGO) - brings a fresh perspective to the company. While the team is relatively new to the CX market and has a significant learning curve ahead, its members have worked together at previous companies and appear to be moving quickly to reposition Five9 while focusing on growth and innovation. There’s clearly a lot changing at Five9. The next question is whether the company can translate its new Humantic CX vision into differentiated products, measurable customer outcomes, and sustained growth.

I recently had a discussion with Mitel's Paul Ginn, Verticals PLM, and Tom Boyle of the NHS and Mitel User Group (MUG) board member, about unified communications in healthcare. Tom's enthusiasm for communication technology in healthcare in general, and Mitel in particular, is impressive. In this video interview, we discuss: - What does a ‘communications failure’ actually look like on the ground for healthcare organizations — and what’s the human cost? - The top communication pain points in healthcare - How Mitel provides reliability and the right deployment model for healthcare organizations - Where AI actually adds value in a clinical communications context — and where is it not ready yet - What's next when it comes to providing patient care and much more