The claims are worth taking seriously as a sign of where the market is going. They're also worth checking carefully before anyone reorganizes a department around them.
What jaam says M. does
The pitch starts from a familiar enterprise problem. In jaam's words, business processes rarely run through a single system or team. Work, knowledge and decisions are spread across people, applications and different stages of a process, making it difficult for AI agents to understand what needs to happen next.
jaam's answer is what it calls a Context Engine. According to FinTech Global's report on the launch, it combines organisational knowledge, business processes, live activity and historical data, and uses that to work out how work moves through an organization and what the next step should be. jaam's release also says M. learns from decisions, outcomes and patterns to improve later work and automation.
Other claims in the release:
- Agent choice: M. can run jaam's own built-in agents alongside agents a customer has already built.
- Microsoft fit: It's designed to work with Azure AI, Copilot, Power Platform, other business systems and custom applications.
- Gradual handover: Organizations can decide where people stay involved and where AI takes on more work over time.
FinTech Global framed the launch around financial services. It reported that M. is aimed at how to give AI agents enough business context to handle complex, multi-step processes while maintaining oversight and control.
Section summary: M. is pitched as an orchestration and context layer that sits above the tools you already use. It isn't a replacement for them. So far, all of this comes from the company.
The five-stage model
jaam describes adoption in five stages:
- Analyse: understand requirements, information, knowledge and processes.
- Build: turn that understanding into the structures and work needed.
- Work: support live work with the right context and information.
- Intelligence: monitor progress, spot risks and flag where action is needed.
- Optimisation: find where AI, agents and automation can improve the process.
This is a product framework, not an implementation guide. It doesn't list prerequisites, connectors, configuration steps or success criteria. Seasoned admins will notice the sequence looks a lot like a standard process-improvement cycle: map, build, run, monitor, optimize. That isn't a criticism, because the pattern is familiar for good reason. The difference jaam claims is that the platform records context along the way, so AI agents have something to reason over.
Who is behind it
jaam isn't a newcomer to the Microsoft world. Its own site says it design AI agents and automation on Microsoft and Nintex, with governance built in. Its services page lists building and running agents on the Microsoft stack and Power Platform, including Microsoft Copilots and "bring your own" agents. That consulting background explains the Microsoft-first positioning. It doesn't independently confirm how M. connects to any particular Microsoft service.
The release also quotes Lucy Bourne, co-founder of Oaka Studio, described as a specialist Microsoft partner growth and strategy consultancy. She says M.'s alignment with the Microsoft ecosystem and ability to work with Microsoft Partners also provides a practical route for organisations already invested in Microsoft technologies. Read that as a supportive comment from someone in the partner ecosystem, not as an independent technical review.
Checking the numbers
Press releases like big statistics, and this one uses four. We checked them against the release's own footnotes and Microsoft's published methodology.
Claim in release Where the release attributes it What we could verify 88% of organizations use AI; 7% have fully scaled it McKinsey (footnoted inconsistently, see below) Not independently confirmed 80% say AI improved individual productivity; 37% report positive EBIT impact McKinsey Attributed to McKinsey's State of AI survey Active Microsoft 365 agents grew 15x in a year Microsoft 2026 Work Trend Index annual report Metric definition confirmed; exact 15x figure not seen in the material we reviewed Gartner flagged "Universal Orchestration" in Feb 2026 Gartner A Gartner document on the topic is dated February 4, 2026
The footnotes don't match. In the PR Newswire version, the 88% figure in the summary bullet points to footnote [1], McKinsey's State of AI. The same 88% figure in the body points to footnote [2], McKinsey's Top Trends in Tech. It may be a formatting slip, but it means readers shouldn't assume all the McKinsey figures come from one survey or one time period.
What "active agents" means. Microsoft's Work Trend Index methodology defines its Monthly Active Agents Growth metric narrowly. It counts unique active agents seen in telemetry across the Microsoft 365 Copilot Agents platform and SharePoint agents, in a rolling 28-day window. An agent counts as active if it has at least one day of user-initiated use or completes at least one autonomous run in that window. The data runs from March 2025 to March 2026 and covers agents built with Agent Builder and the Microsoft 365 Agents Toolkit.
That's a useful measure of how fast agents are spreading across Microsoft 365 tenants. It says nothing about whether those agents deliver business value, and a single use in a month is a low bar.
Microsoft's own research actually supports jaam's broader argument. The same Work Trend Index says organizational factors such as culture, manager support and talent practices account for about twice the reported AI impact of individual effort. In Microsoft's words, people are often ready, but the systems around them are not. Microsoft also notes that its analysis shows statistical association, not causation.
Why orchestration is the next fight
jaam isn't alone in thinking that coordinating agents is the next hard problem. The release points to Gartner, which in February 2026 identified "Universal Orchestration" as an emerging enterprise challenge as companies seek to coordinate AI agents, bots and people across multiple systems and vendors. Gartner's related research, Universal Orchestrator — Governing AI Agents, Bots, and Humans at Scale, is dated February 4, 2026. Gartner didn't endorse jaam, and "Agentic Work Management" is jaam's own term.
For Microsoft shops, the obvious question is where M. fits next to Microsoft's own tools. Microsoft describes Power Platform as a way to unify apps, agents and Copilot to coordinate tasks and automate work. It offers Copilot Studio for building and managing agents, Dataverse as a shared data layer, and over 1,500 connectors. Microsoft also pitches built-in governance: DLP policies, environment management, identity, monitoring and auditability.
So jaam is building in the space between Microsoft's platform and a customer's messy real-world processes, and that space is getting narrower. An independent orchestration layer can be valuable if it covers non-Microsoft systems well or captures process context better than a DIY Power Platform build. If it mostly repeats what Copilot Studio and Power Automate already do, the case gets harder to make.
Section summary: The problem is real, and Gartner and Microsoft's own data both point to it. What jaam has to prove is that M. adds something Microsoft's platform doesn't already provide.
Is it actually available?
The release says jaam "has today launched" M. When we checked search listings, though, jaam's M. product pages were still indexed with a "Coming soon" label next to a "Keep me updated" prompt. That could simply be a stale search cache on launch day. The release itself gives no pricing, trial details, regions or general-availability terms, so anyone interested should confirm what can actually be deployed today before planning around it.
Questions admins should ask before a pilot
The launch materials don't answer the questions IT teams care about most. These aren't known product flaws; they're gaps to close during evaluation:
The bottom line
M. by jaam is a new Microsoft-oriented platform built on a sound idea: AI agents fail when they don't understand how a business actually runs, and organizations need to hand work to agents gradually and in a controlled way. Microsoft's own Work Trend Index makes a similar point about organizational readiness.
What exists today is announcement-level material. There's no architecture documentation, no security details, no published customer results, no pricing, and the product pages still appear as "coming soon" in search. Microsoft-heavy organizations looking at agent orchestration should add M. to their evaluation list and hold it to the same standard as Microsoft's own tools: governance you can test, permissions you can audit, and results you can measure.
If you're weighing Copilot Studio against third-party orchestration layers, or you've already hit the "our agents don't know how anything works here" problem, the WindowsForum threads on Power Platform governance and Microsoft 365 Copilot deployments are a good place to compare notes.
References
- jaam automation Launches M. Platform to Scale Agentic AI and Bridge the Business Productivity Gap - FF News FF News · 2026-09-29T09:44:14.501000+00:00
- jaam automation targets AI scaling gap across financial services fintech.global
- Universal Orchestrator — Governing AI Agents, Bots, and Humans at Scale gartner.com