Don’t Step Into The Platform Trap: What Microsoft Build 2026 Could Mean for Your Next AI Stack Decision
Microsoft Build 2026 produced two announcements that, read together, describe something more interesting than the usual conference launch cadence: a plausible scenario in which enterprise AI stack decisions made in the next 12 months could become significantly harder to reverse. The operative word is “could”. Several pieces of the announced architecture are not fully shipping yet. But the direction is clear. The News Microsoft delivered two related announcements at Build 2026. The first came from Jay Parikh, EVP of CoreAI: the model is not the differentiator; the system governing it is. Microsoft’s answer is a six-step loop. Agents are built in GitHub, contextualized with Microsoft IQ, which grounds them in enterprise data from Microsoft 365, core business systems, knowledge bases, and the web, run in Foundry, governed via Agent 365, and continuously improved through a hill-climbing optimization cycle. Agent 365, combined with Entra, Purview, and Defender, catalogues every agent in the estate, regardless of where it was built, and lets IT enforce policy across all of them. The second came from Mustafa Suleyman, CEO of Microsoft AI: seven new MAI models built from scratch, with no distillation from third-party models. MAI-Thinking-1, the flagship reasoning model at 35 billion active parameters, benchmarks at parity with Anthropic’s Claude Sonnet 4.6 on software engineering tasks at significantly lower per-token cost. MAI-Code-1-Flash is integrated natively into GitHub Copilot. MAI-Transcribe-1.5 claims leading accuracy across 43 languages at five times the speed of competing models. Image and voice models complete the family. Alongside the models, Microsoft introduced Frontier Tuning: enterprises can train MAI models on their own workflow data using reinforcement learning environments. The model...
Your Sales Funnel Is an Architectural Disaster, And How to Change This
Every single week, I sit through pitches from enterprise software vendors boasting about the next iteration of their ” AI-powered sales pipeline optimization platforms”. They promise to auto-magically turn cold leads into closed contracts while minimizing human intervention. That sounds great on a slide deck designed to pump the stock price before an earnings call. In reality, however, these systems are automating an architectural flaw that has plagued B2B organizations since, well, forever: the linear sales funnel. Let me be clear here. The classic sales funnel is not an asset; it is a structural failure. It assumes a predictable, straight line where marketing captures raw interest, tosses a lead over a wall to a sales development representative, who then passes it to an account executive to close the deal. Once the contract is signed, the customer disappears from the pipeline, and is handed off to an underfunded customer success department that operates like a glorified complaints department. This system assumes that buying journeys have a finite endpoint. The B2B buying journey does not end when a contract is signed. By treating marketing, sales, and service as isolated phases with independent processes and technology stacks, enterprise organizations create massive amounts of friction. Norbert Schuster, a veteran B2B strategist who joined us in the latest episode of the CRMKonvos podcast, summarized this beautifully when he described the classic setup as the “Currywurst-Pommes effect“. Individually, a sausage or a plate of chips is acceptable; combined, they become something functional. Yet, in most organizations, marketing automation platforms and CRM instances do not communicate well. They sit side by side as poorly connected line...
Zendesk’s Specialist Bet Is the Right One; and Here’s What Would Make It a Moat
If you only read the press releases, Zendesk Relate 2026 told a strong, clean story. The era of the chatbot is over. Welcome the Autonomous Service Workforce. Resolution replaces deflection. Outcome-based pricing is the new norm. Specialization beats generalist orchestration. That’s strong. Really strong. If you also watched the customer panel, listened to the day-two keynote, and had the chance of having analyst one-on-ones, you got a richer story. One in which the strategic bets are well-placed, the customers describe a more nuanced reality than the slogans, and three specific refinements over the next twelve months that would turn a strong position into a durable moat. I came home quite positive. Here is why, and where I think the next twelve months are important. What Zendesk announced and why it lands The headline product story was the Autonomous Service Workforce: a network of specialized AI agents working alongside humans, orchestrated through what Zendesk now calls the Resolution Platform and improved continuously by the Resolution Learning Loop. Agent Builder gives customers a no-code interface to build bespoke agents. The Copilot suite expanded to four personas: Agent, Admin, Knowledge, Analyst. Voice AI handles 60+ languages mid-conversation. Employee Service AI agents from the Unleash acquisition live inside Slack and Teams. Knowledge Graph spans SharePoint, Google Drive, Notion, Guru, Contentful and Document360. Model Context Protocol support is bidirectional. Quality Score evaluates every interaction. This is quite a handful. Two of these messages are more powerful than the others. The first is resolution over deflection. Zendesk charges only when a resolution is verified by a second AI evaluation model; outcome-based pricing as the natural...