“There’s a meaningful difference between giving enterprises access to an AI model and building that model into workflows where real decisions are made.”
John Grace – North52 CEO
We are deep in the vibe coding era – Copilot / OpenAI / Anthropic. A competent person with a weekend and a prompt can spin up something that looks like a real solution, connected to real data, doing real things. It’s genuinely exciting, and it has shifted how fast teams can prototype.
So here’s the uncomfortable question: if you can vibe code any integration, any automation, any AI-powered feature, do you still need a dedicated business rules engine like North52 sitting inside Dataverse?
Yes. And the reason has almost nothing to do with what happens when you build the thing. It has everything to do with what happens after.
The Recurring Pattern of “Just Build It”
Enterprise tech has a short memory. Every few years, a new capability arrives and the conversation follows the same arc: this changes everything, we don’t need the old tooling anymore, just build it directly.
- Early 2010s: “Just write plugins in C#.” Developers could do anything they wanted inside Dynamics CRM. Many of them did, producing tightly coupled logic that lived in compiled assemblies nobody else could read or safely modify.
- Mid 2010s: Power Automate (then Microsoft Flow) arrived and suddenly everyone was building orchestration logic in a visual designer. Flows multiplied. Some ran. Some didn’t. Many were owned by a single person’s connection credentials.
- Late 2010s: PCF controls and custom code gave front-end flexibility. Teams built elaborate UI-layer logic that quietly duplicated or contradicted server-side rules.
- Now: “Just vibe code it with AI.” Fastest iteration yet.
Each wave produced technical debt, orphaned solutions, and decision logic scattered across the environment in ways that made the next person’s job harder. This is the same argument people once made about spreadsheets replacing databases. The ability to do something is not the same as the ability to govern it, audit it, or hand it to someone else.
North52 has survived every one of these waves. Not because it’s trendy, but because the underlying need for structured, auditable decision logic running at the data layer never went away. Trends cycle. That need doesn’t.
The “AI Access” Illusion
There’s a version of the AI story that sounds compelling on a slide deck: we’ve given the team access to a powerful model, so now they can make better decisions. But access and integration are fundamentally different things, and the gap between them is where enterprise reality lives.
- The last mile problem. AI outputs float in isolation unless someone wires them into the CRM transaction that actually needs them. A sentiment score means nothing if it’s not influencing the case routing that happens next.
- The integration burden. Somebody still has to connect the AI response to the record update, the approval chain, the routing decision. Vibe coded solutions do this, once. Then business requirements shift and the wiring breaks or becomes stale.
- The governance gap. No structured framework ensures that AI-informed decisions follow compliance rules, exception handling protocols, or approval hierarchies. The model doesn’t know your policies.
- The context deficit. A generic AI model has no inherent understanding of your pricing tiers, your territory assignments, your SLA escalation matrix, or the seventeen edge cases your ops team has been managing manually for years.
- The trust problem. When a VP asks “why did the system do that?”, you need a traceable answer. Not a probability score from a black box.
Think of it this way: having a brilliant consultant on retainer is not the same as having that consultant embedded in your approval process and your exception-handling workflow. One gives you advice. The other changes outcomes.
Where North52 Sits, and Why That Position Matters More Than Ever
North52 is a formula-based, no-code/low-code rules engine that runs natively inside Dataverse. In practical terms, it’s the decision-making infrastructure for Dynamics 365 environments. Opportunity scoring, case escalation logic, pricing calculations, approval workflows, data validation, enrichment, automated task assignment.
This is the layer where enterprise decisions actually execute. Not in a chatbot window. Not in a one-off script. In the transactional heartbeat of the business.
When AI gets built into this layer, with AI output feeding in as an input variable to a North52 formula, something meaningful happens. You move from “here’s what the AI says” to “based on what the AI says, combined with these twelve business rules, here’s what the system will do, and here’s the audit trail that proves it.” That’s the difference between a demo and a deployment.
The Vibe Coding Counterargument, and Where It Falls Apart
Let’s be fair to the other side. Modern AI coding tools are genuinely powerful. You can build a plugin that calls Azure OpenAI, parses the response, updates a record, and triggers a flow. Fast. Cheap. Impressive in a demo.
Now put it through the stress test of enterprise reality.
The pattern is consistent: vibe coded solutions optimize for speed of creation. North52 optimizes for speed of operation over time. Those are different objectives, and in enterprise environments, the second one wins.
What This Means Going Forward
The real opportunity isn’t choosing between AI and a rules engine. It’s recognizing that AI becomes dramatically more useful when it operates inside a structured decision layer rather than alongside it.
North52 doesn’t compete with AI. It gives AI somewhere to land. A place where model outputs get combined with business context, validated against policy, and executed with a clear record of what happened and why.
The organizations that figure this out early will have something their competitors won’t: AI that doesn’t just suggest things, but AI whose suggestions are woven into enforceable, auditable business logic. The ones that don’t will keep rebuilding the same integrations every time the tooling shifts, which, given the pace we’re moving at, could be sooner than anyone expects.