Two of the Microsoft Dynamics community's most significant gatherings happened within weeks of each other this spring. DynamicsCon 2026 brought over 3,000 attendees to Las Vegas in May with 300-plus sessions across Finance & Operations, Business Central, and the Power Platform. DynamicsMinds 2026 followed days later in Portoroz, Slovenia, drawing a more technically oriented crowd of 1,000-plus to dig into Dynamics 365 Finance and Operations roadmap specifics, AI architecture, and the future of ERP infrastructure.
Sitting across both events, a few themes surfaced with enough consistency to be meaningful. These are not speculative forecasts; they are the conversations that filled breakout rooms, generated follow-up questions at lunch, and framed Microsoft's own keynote messages. For finance teams and AP operations leaders running Dynamics 365 Finance, they provide a practical lens for where the platform is heading and what decisions are worth making sooner rather than later.
The clearest signal from both conferences was that the shift from generative AI to agentic AI in D365 Finance is already underway, not something scheduled for a future wave. At DynamicsMinds, Microsoft CVP Georg Glantschnig revealed that Dynamics 365 AI ERP now has 14,000 customers, up from 2,100 in 2018, and reported 27% year-over-year ERP growth in FY2026 Q3. The growth figure matters less than what it reveals about adoption momentum: organizations are moving from evaluation to implementation.
Microsoft's product team presented a vision of what they call "agentic ERP," a model where traditional SaaS interfaces give way to collections of agents operating headless business services through the Model Context Protocol (MCP). The ERP MCP Server reached general availability in February 2026, and Wave 1 2026 release notes confirm that improvements to MCP server infrastructure are now a standard part of platform releases. In practical terms, this means agents built in Copilot Studio can read and write to Dynamics 365 Finance data entities, including custom extensions, without requiring ERP-specific development work.
Marco Smits, Microsoft Global Black Belt for D365 AI ERP, outlined a roadmap of at least 25 agents in development across core Finance and Operations capability areas. The ones with the most immediate relevance to AP teams include accounting policy compliance, cash management, budget management, and financial transaction processing. Microsoft also gave a preview of ClawPilot, an agent interface built on the OpenClaw project, demonstrating how it can surface vendor delivery delay alerts and initiate procurement workflows autonomously.
The practical implication for finance operations leaders: the agent architecture is real, the underlying infrastructure is live, and the question is no longer whether to prepare for it but how to sequence adoption without disrupting existing processes.
Business Central's Payables Agent has been in the ecosystem since 2025 Release Wave 1, but DynamicsCon 2026 marked something of a turning point in how the community relates to it. During the "Ask the MVPs: Finance Panel," the audience shifted the conversation away from feature descriptions and toward implementation realities such as what actually happens when you turn it on, where it trips up, and how the first few weeks of live operation differ from demo conditions.
That shift in conversational tone matters. When practitioners stop asking what a feature does and start asking what goes wrong with it, it signals that enough organizations have run it in production to have honest feedback. The consensus from that panel was encouraging in its nuance: the Payables Agent works well for standard, in-policy transactions but requires thoughtful configuration around exceptions, new vendor scenarios, and amount threshold controls.
The broader lesson from DynamicsCon is that automation maturity in AP is now being evaluated against a much higher baseline. Teams that previously benchmarked success against "faster invoice approval" are starting to ask whether their automation layer can handle payment execution autonomously. The gap between those two expectations is where most implementation decisions will happen in the next 12 to 18 months.
One of the more substantive debates that ran across both events was the one between native D365 Finance AP functionality and purpose-built AP automation solutions. This is not a new argument, but the framing has sharpened. Microsoft's Wave 1 2026 investments have made the native tooling meaningfully better, particularly for purchase order-based invoice processing, which has a low-friction path for teams already embedded in D365 Finance.
The counterargument, made clearly in sessions and expo conversations, is that native functionality and best-of-breed solutions are not actually competing for the same jobs. D365 Finance handles what it was designed to handle well. Where it encounters friction is in the broader AP workflow: multi-format invoice receiving, high-volume exception handling, approval routing across large approver networks, and the kind of predictive coding that learns from historical transaction patterns without requiring manual rule configuration.
Dooap's Agentic AP platform is built specifically around this distinction. Rather than replicating what D365 Finance already does, it addresses the operational gaps that persist even after native features are fully deployed, particularly around touchless processing rates, vendor-specific coding logic, and the mobile approval experience. At DynamicsCon, Argano's presence on the expo floor made a similar point, noting that its AP automation solution cuts approval cycles by up to two times while improving extraction accuracy for complex or inconsistent invoice formats.
The most useful framing from these conference conversations is not "native versus third-party" but "what is your current touchless rate, and what would it take to double it." That question tends to reveal where the real gaps are faster than any feature comparison does.
Several DynamicsCon sessions, including Logan Consulting's presentation on Business Central's built-in Power BI reports and Perficient’s sessions on connected data and intelligent automation, pointed to a theme that is quietly changing how finance teams think about their systems. Automation is no longer evaluated in isolation. The expectation is that it feeds real-time reporting, surfaces decision-relevant data to CFOs, and connects to operational systems in ways that eliminate the reconciliation work that used to live in spreadsheets.
The "continuous close" concept, referenced in multiple sessions at both conferences, captures this shift well. Traditional month-end close processes depend on delayed postings, manual reconciliations, and spreadsheet-based adjustments. As agentic ERP infrastructure matures, real-time transaction posting across subledgers becomes possible, and the month-end close starts to look less like a sprint and more like a reporting formality because the reconciliation work has already happened throughout the period.
For AP teams, this has a concrete implication: invoice data, approval timestamps, vendor spend patterns, and payment execution records that previously sat in an AP automation tool as operational data are increasingly expected to flow into analytics dashboards that CFOs and finance directors use for real-time decision-making. The question is whether the AP automation layer is architected to produce that kind of structured, reportable output or whether it treats invoice processing as an end in itself.
DynamicsMinds 2026 included sessions on e-invoicing mandates, structured around the "Big Bang" versus phased implementation question that finance teams in regulated markets are actively navigating. The regulatory pressure is real and geographically expanding: several EU member states have moved from optional to mandatory e-invoicing frameworks, and the compliance requirements vary enough by jurisdiction that organizations operating across multiple markets face meaningful coordination complexity.
What shifted at this year's events was the tone of the conversation. E-invoicing used to be treated as a compliance checkbox; it is increasingly framed as a structural decision that affects how AP automation is architected. If invoices must be received in standardized formats, validated against regulatory schemas, and reported to tax authorities in real time, then the invoice receiving layer of an AP automation solution needs to be built for that requirement, not retrofitted to it.
For D365 Finance users, this means evaluating whether the invoice receiving infrastructure they have in place can handle structured e-invoice formats, Peppol network connectivity, and the validation rules required by local mandates, without creating a parallel process that sits outside the main AP workflow. The organizations at DynamicsMinds that were furthest ahead on this had treated e-invoicing compliance as an AP architecture question from the start, not as a last-minute integration project.
Taken together, these five trends point in a consistent direction: the expectations placed on AP operations are rising faster than most teams' current infrastructure is designed to meet. Agentic ERP is not a future state that can be evaluated at leisure; the foundational infrastructure is already live, and the organizations building habits and configurations now will have a meaningful head start when autonomous processing becomes the operational norm.
The most grounded next step for most teams is not a platform evaluation or a technology investment; it is an honest assessment of where manual effort still persists, what the current touchless rate actually is, and which parts of the AP workflow create the most downstream friction for finance reporting. Those three data points tend to clarify both the priority order and the nature of the investment worth making.
If you are running AP on Dynamics 365 Finance and want to understand how Dooap's purpose-built automation layer addresses the gaps that native functionality leaves open, the comparison between D365 out-of-box and Dooap is a useful starting point, and the team is available for a conversation if the specifics of your environment make the comparison more complex.