The next chapter starts with one process
Quick answer: At the opening keynote of Community Summit North America 2026 in Nashville, Microsoft and its partners made a clear case that AI must now produce business outcomes, not just pilot projects. The path is to start with one process, build trust through governance, define what success looks like, and manage the change that follows.
The keynote opened with a reminder of the event’s scale: more than 600 education sessions, about 500 speakers, 300 partners, and 5,500 attendees. John Siefert, CEO of Dynamic Communities and Cloud Wars, and Dewain Robinson, VP of AI Analysis and Training, set the theme. Last year’s four Cs were curiosity, creativity, critical thinking, and community. This year, the call was to double down on critical thinking, because headlines about failing AI projects miss the real opportunity: rethinking business processes with AI on an agile system of record.
The keynote also described the community as four groups: the creators at Microsoft, the partners who bring solutions to market, the buyers who invest in those platforms, and the users who put them to work every day.
The conversations, in order
Each conversation on stage added a piece to the same picture.
A realistic first step
Mark Hamblin, Senior Solution Advisor at InsightWorks, was recognized for the firm’s VIP program. InsightWorks has attended Community Summit for 15 years and brought more than 100 clients to this week’s event. Hamblin said the expo is full of AI tools that look futuristic, and that the strongest first step is something realistic: a process the team can achieve and measure. His firm now hears more client questions asking what is realistic.
From pilots to production
Jennifer Harris, CEO of TMC, and Shane Young, TMC Global’s Chief AI Officer, described the shift from doing anything with AI to doing it the right way, with purpose and governance. Young highlighted enterprise data protection and the confidence that a platform does not train its models on a customer’s data. Harris urged attendees to stay open-minded and to put governance in place before fear stalls the work.
Redesigning the process
Georg Glantschnig, Corporate Vice President for Microsoft Dynamics 365 Agentic ERP Applications, offered three priorities. Reimagine the whole process, not just the task, across finance, operations, suppliers, and customers. Build a foundation of trust with governance, business context, controls, and an environment where AI can act. Start with outcomes: pick one process, define success, bring in partners, and iterate and scale.
Change management in a 12-month transformation
Ryan Fritts, CIO of Everon, and Mike Gillis, Owner and Principal at RSM, described Everon’s carve-out from ADT in Q4 2023. Everon replaced most of its core systems within 12 months, including moving its ERP and CRM to Dynamics, and its team worked very long weeks without a single resignation that year. Gillis said client conversations have moved from which product to buy to how to manage change, and that AI cannot be applied to bad data. Fritts described a shift from volume to value, with AI opening workflows the business would never have asked to automate.
Customers and platforms moving faster
Olof Hedin, GTM at Lovable, joined Glantschnig to discuss Lovable’s rapid growth: over 25 million people on the platform and 17 million projects built in about four months. Glantschnig described a Microsoft session where prototypes that once took weeks could be built in hours or days, and where the prototype became the start of the product, not the end of the customer conversation. Hedin said Lovable and Microsoft have signed a $245 million agreement and are developing products and marketing together. The keynote also cited HCA Healthcare, where nurses are building their own interfaces connected to patient records.
Transforming a manufacturer
Bonnie Duddlesten, VP of Enterprise Transformation at Barry-Wehmiller, and Aynsley Keller, SVP North America at Truvio, joined Glantschnig for a session on transforming an industrial machinery company with more than 12,000 team members.
Humans and agents on one team
The closing Microsoft keynote featured Dona Sarkar, Chief Troublemaker for Microsoft Enterprise AI Advocacy, and Leon Welicki, VP and GM of Power Platform. Welicki outlined a four-part strategy: improve the products, make the platform the best place for humans and agents to work together, turn every app into an agentic surface, and expose organizational knowledge so agents can use it. He cited 56 million monthly active users and use by more than 97% of the Fortune 500. A live demo showed agents joining a team space to fix color contrast on a home screen and run a performance analysis, then leaving once they were done.
Microsoft’s own transformation also offered lessons. Its engineers reduced incident-handling time by 35%, with the first 25% coming from better data quality. Sarkar closed with six skills people still need as agents take over technical work: imagination, judgment, empathy, curiosity, the courage to retire what no longer works, and presence, showing up in person.
What this means for partners and mid-market companies
For Microsoft partners, the keynote invited a shift from technology to outcomes. Glantschnig named change management as the biggest opportunity for partners, more than installing software. Clients can buy software anywhere, but they need help deciding what to change and how.
For mid-market companies, typically those with roughly $20 million to $250 million in revenue, the lessons are practical:
-
- Start with one process your team understands well and can measure.
- Set governance before AI acts on company data, including whether a vendor trains on your data.
- Define the outcome in one sentence that finance and operations would both accept.
- Clean and reconcile core data before adding AI.
- Plan the change with the same care as the technology, and keep the mission visible.
CEM works in this space, building Microsoft Dynamics 365 solutions for construction and manufacturing companies. The keynote’s message about starting small, governing well, and measuring outcomes matches how CEM helps clients plan.
Next steps
The fastest way to act on the keynote is to pick one process this quarter. Map it from request to result, mark where work waits, and write the outcome in one sentence. Then decide what the AI tool can see, who reviews its output, and how you catch errors.
Your turn: Which process in your business keeps people waiting the longest, such as approvals, month-end close, or change orders? If you can’t yet describe its outcome in one sentence, or you aren’t sure the data behind it is clean, start there. Check where your systems and data stand before your next AI investment.
