The Sector

Microsoft is rebuilding Dynamics around Copilot. Are you?

By Marcus Whitaker · 10 July 2026

A sizeable tier of UK membership bodies does not run a purpose-built association system at all. It runs Microsoft Dynamics 365, assembled into a membership solution by a specialist partner: the base CRM, a set of association modules, and a layer of custom build that encodes how that particular body grades members, prices renewals and runs its CPD. Three structural facts now bear on everyone in that tier. Microsoft’s product gravity has shifted decisively towards Copilot, agents and the Power Platform. The partner channel that built these solutions is thinner than it was. And the economics of that channel, which bills by the hour, shaped the advice that put many of these builds in place. None of these is a reason to panic. All three are reasons to look carefully at the contract you signed.

Why did so many membership bodies end up on Dynamics?

Familiarity, pricing and people. Dynamics looked safe to trustees because Microsoft looked safe. The old procurement saying that no one ever got fired for buying Microsoft survives because it describes a real boardroom incentive: a trustee can defend Microsoft to colleagues and auditors in a way they cannot defend a 40-person software house. Charity and nonprofit licensing sweetened it: as of August 2026 Microsoft’s nonprofit price list offers Dynamics 365 Sales Enterprise and Service Enterprise at $23.80 per user per month, Business Central at $32, and Power Apps free for the first ten users, with additional users at $2.50. For a finance director comparing against a specialist AMS quote, those numbers did real persuading.

The third factor was the partner. Sector commentary makes the structure plain: Dynamics-core association solutions were “typically assembled by specialist partners stacking association-specific functionality on top of the base platform”, an approach that “won ground among mid-sized bodies”, as MemberWise noted in June 2026. The partner knew membership; Microsoft supplied the engine; the body got a system shaped to its processes.

And the strengths should be recorded plainly, because they explain why the tier persists. A body on Dynamics sits inside the Microsoft stack its staff already live in: Outlook, Teams, SharePoint, Excel. The member record and the mailbox are neighbours. Microsoft markets the arrangement with named customers: the Royal Horticultural Society runs Dynamics 365 and Azure, and Toastmasters International uses Dynamics 365 for its member experience, per Microsoft’s published case studies. For organisations whose members expect a portal that talks to everything else, that integration is the product.

What do the partner economics reward?

Hours. An implementation partner earns by configuring, integrating, migrating and then servicing what it built, and platforms differ greatly in how many of those hours they generate. An assembled Dynamics solution (platform, association modules, custom build) carries more delivery work than a productised membership system, and the channel that recommends it is paid in proportion.

The same MemberWise commentary, written by iMIS partner iFINITY and worth reading with that interest in mind, calls the Dynamics-core approach “a credible approach, if an expensive and slow one”, argues that a solution welded together from platform, modules and partner-built extensions “tends to carry more integration and configuration weight than the sum of its parts implies”, and records that industry consultants have assumed its own, lower iMIS implementation figures must be under-quoted, “conditioned by the inflated estimates competitors hand them”. That is a vendor talking its own book. It is also a vendor putting a number-shaped claim on the record: productised platforms arrive with lower delivery hours, and the consultants who write shortlists have been trained by experience to expect the higher ones. None of this requires bad faith from any individual adviser. A channel that bills hours will, on average, produce advice with more hours in it, and the brand safety of Microsoft underwrites the quote at board level.

The counterweight is the independent-adviser model: firms paid by the buyer, for advice alone. Hart Square, which works exclusively with charities, membership bodies and education organisations, reports 550-plus projects and states that it sells no software and takes no commissions. Buyers in this tier, or considering joining it, should establish which model any adviser belongs to before the shortlist is written. We examine the advisory layer across the whole market in our companion analysis.

What does Microsoft’s Copilot pivot mean for your implementation?

Look at Microsoft’s own Dynamics 365 pages as of August 2026 and the repositioning is not subtle. The product is sold as “agentic business apps” that move “from a system of record to a system of action”. Prebuilt agents proliferate: a Sales Qualification Agent, a Case Management Agent, an Account Reconciliation Agent. Copilot Studio is the customisation surface, and Microsoft has announced Model Context Protocol servers for Dynamics 365 so external agents can work against the platform. The roadmap money is going into agents acting on standard data models.

For a membership body, the opportunity is real. Renewal chasing, member-query triage, event follow-up: these are standard-shaped tasks on standard-shaped data, exactly where prebuilt agents land first. A body running close to standard Dynamics workloads can expect Copilot value to arrive with its licences, cheaply and soon.

The catch is the phrase “standard-shaped”. The Copilot wave assumes your data lives where Microsoft’s agents look for it, in Dataverse, in the standard entities, with the standard semantics. The further your build is from that assumption, the less of the wave reaches you automatically.

What happens to heavy customisation?

It gets re-examined, piece by piece. Membership solutions built a decade ago encode membership logic in custom entities, plug-ins and partner-built modules: grade structures, subscription rules, CPD registers. Copilot does not know your custom grade table exists, and an agent asked to draft a renewal offer will reason from what it can see. Making the custom layer legible to agents means rework: remapping logic onto standard entities where possible, exposing the rest through governed connectors, and testing that an agent’s proposed action respects rules that used to live in a developer’s head.

This is not an argument against customisation; some of it carries organisational knowledge that no standard product holds. It is an argument for an inventory. Which customisations still earn their maintenance? Which replicate what the platform now does natively? Bodies that ran that exercise before the Copilot wave will absorb it cheaply. Bodies that skipped it will fund a larger project later, at partner day-rates, under time pressure: the hours problem again, arriving from a new direction.

There is also a pricing wrinkle to model. Agent capability increasingly arrives as consumption-priced Copilot Credits layered on top of licences: as of August 2026 Microsoft sells Copilot Studio in tenant-wide packs of 25,000 credits at $200 per pack per month, or on a pay-as-you-go meter settled through an Azure subscription, per its published pricing. The discounted nonprofit per-user licence is the floor of the bill; agent usage sets the ceiling, and finance directors should model the agent usage line before a keen IT lead turns everything on.

Who supports your build in five years?

The harder question, and the one boards ask least. The same MemberWise commentary records that “a good number of the smaller suppliers behind it have since disappeared”. That is the quiet arithmetic of a consolidating partner channel: the small consultancy that built your membership layer may have been acquired, merged, or simply wound up, and the knowledge of why your renewal pricing works the way it does left with it.

Microsoft is not the risk here; the platform is not going away. The risk sits in the layer between Microsoft and you: the partner build. A Dynamics membership solution is a three-party dependency (platform, partner modules, custom code) and its resilience is the resilience of the weakest party. Where the original partner has gone, a body faces a choice: pay a successor firm to reverse-engineer the build, migrate the custom layer onto a supported product, or carry an increasingly fragile system.

This paper does not name specific firms, because the market moves faster than print and naming survivors as safe harbours would be its own kind of error. The question matters more than the label on any firm: can your current supplier show you the documentation, the source, and the succession plan for your build? If the answer is silence, that is your answer.

What should a Dynamics-running body ask now?

Five questions, all answerable without a consultancy engagement. First, how far is our build from the standard Dynamics data model, and what would it cost to close the gap? Second, which of our customisations duplicate what the platform now ships? Third, who owns and documents our custom code, and where is the escrow if the partner fails? Fourth, what will Copilot Credit consumption add to our licence line at realistic usage? Fifth, if we had to move off this build in eighteen months, what would we move to?

None of these questions implies that leaving Dynamics is the right answer. For bodies deep in the Microsoft stack, with disciplined customisation and a stable partner, the Copilot pivot is an upgrade path paid for by someone else’s R&D budget. The point is that the tier’s founding bargain (safe platform, expert partner, shaped build) now has all three legs moving at once: the platform is being rebuilt around agents, the partner channel is consolidating, and the hourly economics that recommended the build deserve the same scrutiny as the build itself. Bodies that re-examine the bargain on their own timetable will renegotiate from strength. Those that wait for the partner’s letter, or the licence renewal that includes an agent line nobody budgeted, will not.