The Business Central Roadmap for the Next Decade

Every few months a new release wave lands, a new Copilot capability gets a splashy demo, and someone asks me the same question: is Business Central still the right horse to back for the next decade, or is Microsoft about to change the game underneath everyone who committed to it? I have spent enough years living inside ERP and integration projects to be wary of two opposite mistakes. The first is treating a vendor roadmap as prophecy, planning your business around features that exist only on a slide. The second is dismissing the roadmap entirely as marketing, and being blindsided when a direction Microsoft signalled for three years quietly becomes the default everyone depends on. The truth sits between those poles, and this article is my attempt to map it honestly.

The message up front: the biggest bet you make when you choose Business Central is not on a feature list, it is on a platform and the company steering it. Microsoft has spent a decade turning Business Central into the ERP core of a much larger ecosystem, and the next ten years are about that ecosystem, AI, low-code, and a unified data layer, growing around the ERP rather than the ERP growing new modules in isolation. Judge the direction, not the demo.

1. Where Business Central sits today, and how it became Microsoft's flagship SME ERP

To read the roadmap you first have to understand how Business Central got where it is, because the path explains the direction. The product traces its lineage back to Navision, a Danish accounting system that Microsoft acquired in the early 2000s and sold for years as Dynamics NAV, an on-premises ERP with a devoted partner channel and a reputation for being deeply customisable and slightly idiosyncratic. NAV was capable, but it was a classic on-premises product: you bought a licence, a partner installed it on your servers, and you lived with whatever version you implemented until a painful upgrade project forced you forward.

Business Central is what happened when Microsoft rebuilt that heritage for a cloud-first world. It kept the functional depth that made NAV trusted for real finance, inventory, manufacturing and project accounting, and it re-platformed the whole thing as a continuously updated software-as-a-service product running on Azure, with a browser and mobile client, an extension model that keeps customisations separate from the core, and a cadence of predictable twice-yearly upgrades that happen whether you plan a project or not. If you want the fuller story of that transition, I have written it up in the cloud ERP explained pillar and the introduction to Business Central.

Why did this become Microsoft's flagship ERP for small and mid-sized organisations rather than one of several? Partly deliberate strategy. Microsoft had a larger enterprise ERP in Dynamics 365 Finance and Operations aimed at big, complex organisations, and it needed a clean, focused product for the enormous mid-market where most businesses actually live. Business Central filled that space. But the deeper reason is that Microsoft stopped selling Business Central as a standalone ERP and started selling it as the finance and operations core of a single connected stack: the same Microsoft 365 that runs the company's email and documents, the same Azure identity that governs access, the same Power Platform that builds the workflows, and increasingly the same AI layer that Microsoft is threading through everything. An SME that already runs on Microsoft finds an ERP that speaks its language natively. That gravitational pull, more than any single feature, is what turned Business Central into the flagship. It is worth understanding that ecosystem story in detail, because everything in the roadmap flows from it.

2. How to actually read Microsoft's roadmap

Before we forecast anything, it is worth being disciplined about how Microsoft communicates the future, because reading it wrongly is how people get burned. Microsoft ships Business Central on a rhythm of two major release waves a year, one in spring and one in autumn, and each wave is preceded by published release plans that describe what is coming. Layered on top of that is a broader public roadmap portal where Microsoft signals features across its product families with statuses like planned, in development, rolling out, and launched. These are genuinely useful, and far more transparent than the old world of on-premises software where you found out what changed when the upgrade landed.

The discipline is in reading the status honestly. There is a large gap between a feature in public preview and a feature at general availability. Preview means Microsoft is committed enough to let you try it, not that it is finished, supported, or guaranteed to ship in its current form or its current timeline. Features slip. Features get reshaped between preview and GA. Occasionally features quietly disappear from the plan entirely. A roadmap entry is a statement of intent under current conditions, and conditions change.

My practical rule for clients is this: you can build operational plans on what is generally available today, you can pilot and experiment with what is in preview, and you can factor the direction of the roadmap into strategy while committing zero budget to features that have not shipped. Separate the signal from the hype by asking one question of every announcement: is this shipped and supported, is it previewing, or is it a vision statement at a keynote? All three are legitimate and all three are useful, but they carry completely different levels of certainty, and treating a keynote vision as a delivered capability is the single most common way I see people misjudge a platform.

A caution on demos: the polished AI demo you see at a conference is engineered to work. It runs on curated data, a rehearsed scenario, and often a capability still in preview. That does not make it dishonest, but it makes it a poor basis for a purchasing decision. The gap between "this worked in the demo" and "this works reliably on our messy real data" is exactly where disappointment lives. When you evaluate any roadmap claim, discount it to what you can verify on your own data, and treat everything above that line as promising rather than proven.

3. Copilot and AI as the new center of gravity in ERP

The single clearest direction in Business Central, and the one Microsoft has committed to most visibly, is that AI is moving from a bolt-on feature to the center of gravity of the product. Microsoft has branded its assistive AI across the whole stack as Copilot, and Business Central has been getting Copilot capabilities steadily: assistance with drafting product descriptions and marketing text, help reconciling bank statements and matching transactions, natural-language help finding and explaining data, and analysis assistance that lets a user ask a question of their business data in plain language rather than building a report by hand. These are shipped or shipping capabilities, not speculation, and the trajectory is unmistakable.

What I am confident about, as extrapolation rather than announcement, is the shape of where this goes. AI in ERP is moving along a predictable arc. It started as content generation, the easy and visible wins like drafting text. It is moving into transaction assistance, where the AI proposes the posting, the match, the classification, and the human confirms. The clear next phase is decision support that reaches across the connected data, an assistant that can answer "why is this customer's margin down this quarter" by pulling from finance, sales and operations at once, because it sits on a unified data layer rather than a single module. That is the direction, and it follows logically from what Microsoft has already built.

The important practitioner's reframe is that Copilot is not one feature you evaluate, it is a design philosophy Microsoft is applying to the whole product. The question for the next decade is not "does Business Central have AI" but "how much of the routine finance and operations work becomes assisted or automated, and how do we keep control while that happens." That control question is not a footnote. It is the whole game, which is why I keep returning to AI governance for enterprise operators as required reading before you lean on any of this in a system of record.

The insight worth internalising: in a system of record, the value of AI is throttled by trust, not by capability. An assistant that drafts a journal entry is only useful if a human can verify it quickly and an auditor can trace how it was produced. The organisations that win with Copilot over the next decade will not be the ones with the most AI features switched on, they will be the ones that built the review, approval and audit discipline that lets them actually rely on the output. The technology will outrun the governance, and the governance is what determines whether you can safely use the technology.

4. From modules to agents: the shift toward autonomous business agents

The most significant strategic shift Microsoft has been signalling, and the one I watch most closely, is the move from AI as an assistant that helps a person to AI as an agent that carries out a defined task with a degree of autonomy. Microsoft has been explicit that it sees agents as a major direction, describing capabilities where an AI agent monitors for a condition, takes a multi-step action, and reports back, rather than simply answering a question when asked. This is genuinely different from Copilot-as-helper, and it is the direction that will reshape how ERP work gets done if it matures the way Microsoft intends.

Picture the difference concretely. A Copilot helps a person clear the sales-order exception queue faster. An agent is configured to watch that queue, apply your rules to each exception, resolve the ones that clearly fit a known pattern, escalate the genuinely ambiguous ones to a human, and log everything it did. The human moves from doing the work to supervising a process that mostly runs itself and surfaces only the cases that need judgement. Multiply that across accounts payable matching, collections follow-up, inventory replenishment suggestions, and month-end reconciliations, and you can see why Microsoft frames this as a structural change rather than a feature.

Here is where I separate what is announced from what I am extrapolating. Microsoft has announced the direction and shipped early forms of agent capability. What is genuinely uncertain, and where I will not pretend to certainty, is how far and how fast this goes in the messy reality of real businesses with imperfect data and non-standard processes. Autonomous action in a financial system of record raises hard questions: who is accountable when an agent posts something wrong, how do you audit a decision the agent made, how do you constrain it so a bad rule does not do damage at machine speed. These are not reasons to dismiss the direction. They are the reasons the direction will arrive more gradually and more carefully than the keynotes suggest, and rightly so. My honest forecast is that semi-autonomous agents handling well-bounded, high-volume, low-ambiguity tasks under human supervision become normal within the decade, and that full autonomy for consequential financial decisions stays deliberately gated for a lot longer, because the accountability problem is not primarily a technology problem.

5. Power Platform and the low-code convergence

Running parallel to the AI story, and just as important to the next decade, is the convergence between Business Central and Microsoft's low-code Power Platform. This is not speculative. Microsoft has spent years deliberately blurring the line between the ERP and the tools that extend it, so that Power Apps can build interfaces over Business Central data, Power Automate can orchestrate workflows that reach into and out of the ERP, and Power BI reads its data for analytics, all sharing the same identity and increasingly the same underlying data fabric.

The strategic idea Microsoft keeps pushing is the "fusion team", the notion that a professional developer and a business-side "citizen developer" collaborate on the same platform, with the professional building the components that need real engineering and governance, and the business expert assembling and configuring the parts that do not. For Business Central this matters enormously, because historically extending an ERP meant a specialist writing code in a specialist language, which made every small change slow and expensive. The low-code direction promises that a meaningful share of the customisation and extension work moves to people closer to the business, faster and cheaper.

My practitioner's view, which is more cautious than the marketing, is that this is real and valuable but easy to oversell. Low-code genuinely lowers the cost of the eighty percent of extensions that are simple: a custom approval flow, a small app for a warehouse task, a workflow that connects the ERP to another system. It does not, and I do not believe it will within the decade, remove the need for real engineering discipline on the twenty percent that touch financial integrity, complex logic, or high-volume performance. The risk I watch for is ungoverned sprawl, hundreds of citizen-built flows and apps that nobody owns, nobody documents, and nobody can safely change. The convergence is a genuine advantage, but it shifts the hard problem from "writing the code" to "governing who builds what and keeping the estate maintainable." Organisations that embrace the low-code direction without the governance to match will trade one kind of technical debt for another.

6. Data, analytics and Microsoft Fabric

Underneath both the AI story and the low-code story sits a less glamorous but arguably more consequential direction: the unification of data. Microsoft has been building toward a single analytics and data backbone, most visibly through Microsoft Fabric, its unified data platform that brings together data engineering, warehousing, and business intelligence, and through the broader idea of a shared data layer that multiple applications read from and write to. The direction is that Business Central stops being a data silo you have to export from, and becomes a participant in a unified data estate where its transactions are available for analytics, AI, and cross-application processes without brittle custom integration each time.

This matters more than it first appears, and here is the connection people miss. Every capability I have described so far, Copilot answering cross-functional questions, agents acting across processes, low-code apps combining data from multiple systems, depends on that unified data layer existing. AI is only as good as the data it can reach, and an assistant that can see finance but not operations gives shallow answers. The reason Microsoft is investing so heavily in the data fabric is that it is the foundation the whole AI and automation strategy stands on. Without a unified, queryable data estate, the impressive AI demos stay demos.

For someone making an ERP decision, the practical implication is that the value of choosing Business Central compounds when the rest of your data lives in, or connects cleanly to, the same Microsoft data estate. An organisation that keeps its ERP data isolated captures a fraction of what the platform can do. An organisation that lets Business Central feed a unified analytics backbone, and consumes that backbone through Power BI and increasingly through AI, is positioned to benefit from every capability Microsoft ships on top of it. The data strategy is the quiet decision that determines how much of the roadmap you actually get to use.

7. The partner and ISV economy evolution

Business Central has always been sold and implemented primarily through a partner channel, a large ecosystem of implementation partners and independent software vendors who build add-on solutions for specific industries and needs. That channel is a genuine strength: it means there is likely an existing solution for your niche, whether that is a specific manufacturing process, a regional tax regime, or a vertical like professional services or field operations. But the economics of that channel are being reshaped by the same forces reshaping the product, and buyers should understand how.

Two shifts are underway. First, the move to a continuously updated cloud product changed the partner's job. In the old on-premises world, a large share of partner revenue came from big, infrequent upgrade projects and heavy customisation. In the cloud world, the upgrades happen automatically and heavy core customisation is discouraged in favour of clean extensions, so the partner's value moves from "install and customise" toward "advise, integrate, and continuously optimise." The good partners have made that transition to an ongoing advisory relationship. The ones still selling one-off implementation projects and disappearing are being squeezed.

Second, AI and low-code change what an ISV add-on is worth. When a chunk of what used to require a custom add-on can be assembled with low-code, or handled by a Copilot capability Microsoft ships natively, the ISV value proposition has to move up the stack toward deeper domain expertise and genuine intellectual property rather than filling gaps Microsoft will eventually close itself. My honest read for buyers is this: choose partners and ISVs for domain depth and a demonstrable commitment to the direction of the platform, not for the size of the customisation they propose. The best partner in the next decade is the one who helps you use more of what the platform already gives you and customises less, because on a continuously updated cloud product, every unnecessary customisation is a future maintenance liability. For the buyer's-eye view of running that evaluation, the implementation journey pillar walks through it end to end.

8. What will NOT change: the fundamentals that outlast every wave

For all the change I have described, the most important thing I can tell anyone planning for the next decade is what will not change. It is easy to get swept up in each wave of AI, agents and low-code and forget that these are all built on top of fundamentals that have not moved in thirty years and will not move in the next ten. Three of them matter above all.

The first is clean data. Every capability in this article, from the simplest report to the most autonomous agent, is only as good as the data underneath it. An AI assistant trained to answer questions about your business will confidently give wrong answers if your data is wrong. An agent configured to act on conditions will act wrongly on bad data at machine speed. The organisations that get value from the next decade of ERP are the ones that treat data quality as a permanent discipline, not a one-time migration task. Nothing on the roadmap rescues bad data, and I would go further: the more powerful and autonomous the tooling becomes, the more expensive bad data becomes, because errors propagate faster and further.

The second is sound process. ERP encodes how a business actually operates, and no amount of AI fixes a broken process. If your procurement approval is a mess, automating it just makes the mess run faster. If your inventory process does not reflect reality, an agent managing replenishment against it will make confident bad decisions. The fundamentals of designing a clean, well-understood business process are exactly as important in the age of Copilot as they were in the age of paper, arguably more so, because you are increasingly delegating execution of that process to software that assumes the process is right.

The third is disciplined governance. Who is allowed to do what, how changes are controlled, how you audit what happened and why, how you keep the estate maintainable. Every new capability, low-code apps, AI agents, cross-system data flows, expands the surface that governance has to cover. The platform will hand you more power every year. Whether that power is an asset or a liability depends entirely on the governance discipline you wrap around it, and that discipline is a human and organisational capability, not something Microsoft ships. These three fundamentals are the boring, durable truths that outlast every release wave, and every serious ERP decision should be anchored to them rather than to the feature of the moment.

9. What this means for an SME making an ERP decision right now

Suppose you are not a technology strategist, you are the finance director or owner of a mid-sized business who has to choose an ERP this year and live with it for a decade. How should everything above shape that decision? The honest answer is that the roadmap should inform your choice without dominating it. You are buying a system to run your business today, and it has to do that well on day one regardless of what ships in three years.

Start with the fundamentals of fit. Does Business Central handle your actual finance, inventory, and operational needs well, today, in its shipped form? That is the question the is Business Central right for your organization pillar is built to answer, and it is the question that matters most. A brilliant roadmap on a product that does not fit your business today is worthless. If the fit is there, then the roadmap becomes a genuine tiebreaker and a source of confidence, because it tells you the platform you are joining is being invested in heavily and moving in a direction that compounds over time.

The specific advice I give SMEs weighing this decision is fourfold. Choose the platform for how well it fits your business now, and treat the roadmap as upside rather than as the reason to buy. Implement clean, resisting the temptation to heavily customise, so that you can actually receive the continuous updates and future capabilities as they arrive rather than being frozen by your own modifications. Invest early in data quality and process discipline, because those are the fundamentals that determine how much future value you can capture. And build your governance from the start, so that when the powerful new capabilities arrive you are ready to use them safely rather than scrambling. If you are also carrying an older system, the legacy migration pillar covers how to get onto the platform cleanly enough to benefit from all of this.

10. Betting on the platform: an honest take on risks and upside

Committing to Business Central is committing to Microsoft's platform strategy, and intellectual honesty requires naming the risks as clearly as the upside. Let me do both.

The upside is substantial and, I think, underappreciated. You are joining an ERP that is not a standalone product but the finance and operations core of the most heavily invested software ecosystem on the planet. Every dollar Microsoft spends on AI, on the data fabric, on low-code, and on identity and security flows toward the product you are running, because it is designed to consume all of them. The continuous update model means you receive that investment as a stream rather than as periodic painful upgrades. And the sheer size of the partner and ISV ecosystem means you are rarely the first to solve any given problem. For an SME, being carried along by that scale of investment is a real strategic advantage that a smaller vendor simply cannot match.

The risks are equally real. The first is lock-in. The deeper you integrate into the Microsoft ecosystem, the harder and more expensive it becomes to leave, and that dependence is a genuine strategic exposure you take on with eyes open. The second is pace of change. The continuous update model that delivers all this value also means you cannot stand still, twice-yearly upgrades happen whether you are ready or not, and organisations that cannot keep up with that cadence will struggle. The third is the AI hype risk: some of what is promised will arrive later, smaller, and messier than the keynotes suggest, and if you make commitments based on unshipped capability you will be disappointed. The fourth is the governance burden, which grows with every capability and which the vendor cannot carry for you.

The lock-in caution, stated plainly: I am broadly positive on Business Central, and I still tell every client that ecosystem lock-in is a real cost, not a paranoid fantasy. The convenience of everything speaking to everything is the same thing as dependence on one vendor's direction and pricing. That is not a reason to avoid the platform, plenty of the best decisions involve accepting a known dependence for a real benefit, but it is a reason to go in deliberately, to keep your data in forms you could extract if you ever had to, and to never pretend the exit door is as cheap as the entrance.

Weighing it all, my honest position is that for a mid-sized organisation already living in the Microsoft world, the upside clearly outweighs the risks, provided you go in with discipline. The platform bet is a good one. It is not a free one, and anyone who tells you it is, is selling something. The organisations that will look smart in ten years are the ones that took the upside while managing the lock-in, kept up with the pace instead of fighting it, discounted the hype to what shipped, and built the governance the vendor could not build for them.

Final thoughts

A roadmap is a direction of travel, not a contract. Everything I am most confident about in this article is the direction: AI moving to the center of the product, agents shifting work from doing to supervising, low-code lowering the cost of extension, a unified data fabric becoming the foundation the whole strategy stands on, and a partner economy reshaping itself around a continuously updated cloud product. Those directions are strongly signalled by what Microsoft has already shipped and invested in, and I would plan strategy around them. The specific features, timelines and version details are far less certain, and I have tried throughout to keep the line clear between what has been announced or delivered and what is my own informed extrapolation.

The part I am most certain of is the part that never appears on a keynote slide: the fundamentals do not change. Clean data, sound process, and disciplined governance were the determinants of ERP success before any of this AI arrived, and they will be the determinants long after the current wave has become ordinary. If you are choosing an ERP for the next decade, choose for fit today, implement clean, and pour your discipline into those fundamentals. Do that, and the roadmap becomes upside you are positioned to capture. Skip it, and no amount of Copilot, agents or fabric will save you, because the most powerful platform in the world still cannot rescue a business that has not done the unglamorous work underneath.

Related reading: Introduction to Dynamics 365 Business Central, Cloud ERP explained, Business Central in the Microsoft ecosystem, Is Business Central right for your organization, AI governance for enterprise operators.