Join Convergence, a movement among startup technical founders & operators who are done with scattered tactics & ready to install the growth systems, decisions & leadership that move revenue.
The AI readiness gap your ERP vendor isn't talking about
Published about 11 hours ago • 4 min read
Reader, The other day Andy Park told me about a friend who runs a mid-market manufacturing company.
The guy has Epicor running from a server closet at his plant. Products ship on time. His team knows the system. Everything works.
Then the vendor starts pushing him to migrate to the cloud version. Three times his current annual spend. No direct migration path. A full reimplementation.
His question is completely reasonable: Why would I spend half a million dollars and risk my entire operation when what I have works fine?
But here's what I kept coming back to after my conversation with Andy and Marc Johnson, co-founder of TeamCentral. The friend's question is framed around the wrong risk. The real question isn't whether to migrate. It's whether he's building the data foundation that makes AI possible before his competitors do it without him.
Andy and Marc spent almost 20 years together at a global consulting firm, seeing the same integration failures across hundreds of enterprise implementations. That pattern recognition is exactly what led them to build TeamCentral, a middleware no-code platform built for mid-market manufacturers, distributors, and logistics companies that don't live in a single-vendor world.
As a licensed P.E. turned growth strategist, infrastructure-first thinking is exactly the lens I bring to GTM. When Marc and Andy started talking about data architecture as the prerequisite to AI readiness, I recognized the same pattern I see in every company that tries to scale before their systems are actually ready for it.
Why most ERP migrations are moving on the vendor's timeline rather than the customer's, and what that gap costs companies who haven't thought through data architecture first
The 94% supply chain visibility problem: why most companies are making operational decisions without complete data, and what that looks like in practice at the team level
How band-aid integrations compound into spaghetti architecture that makes AI implementation impossible, and why the cheap fix always costs more later
The four pillars of AI readiness (connected, quality, accessible, secure) and what it actually takes to assess where your organization stands on each one
TeamCentral's incremental governance playbook: why massive data definition projects fail, and how to start with the smallest possible scope and build from there
How MCP (Model Context Protocol) could solve the multi-vendor AI agent problem that most enterprise companies are about to run into hard
Oracle, SAP, Microsoft Dynamics 365, Salesforce, Epicor, GP, NAV, Sage, JD Edwards, SAP ECC, AS/400
TeamCentral's Corbi agent (Cortex of Your Business), including enterprise search, task automation, and Pulse (role-specific data feed): https://www.teamcentral.ai/products/
The framing that stuck with me most is this… The ERP migration conversation your vendor is having with you is actually an AI readiness conversation, and most vendors are not framing it that way. Companies that move to the cloud on the vendor's schedule without thinking through data architecture will miss the window to position for AI. That is not a recoverable situation in two years.
The 94% supply chain visibility problem is a data connectivity gap. You cannot solve it by selecting a better AI model. Andy described what it looks like in practice: a salesperson sees 20 units available, a customer needs 15, so they hold all 20 because they have no visibility into what's already promised or incoming. Multiply that across a sales team and you have inventory paralysis from otherwise rational people doing exactly their jobs. No model fixes that. The foundation has to come first.
Point-to-point integrations make sense at the moment you build them. The CRM needs to talk to the ERP, so IT finds the fastest path. Then another system. Then another. Each individual decision was rational. But architectural debt compounds exactly like financial debt, and the interest comes due the moment you try to build anything sophisticated on top of it.
Before any AI investment, this four-pillar check is the actual starting point: connected, quality, accessible, secure. Whichever pillar is weakest is where you focus first. Not model selection, not prompt engineering. The infrastructure work that nobody wants to talk about because it is not exciting.
Sequence matters more than ambition here. Model your data holistically before you think about which systems to connect. Define what your data should look like across every system, then figure out how to move it. The projects that try to align the entire organization on data definitions upfront fail because business requirements change before implementation finishes. Start with one workflow. Make the rules as simple as possible. Ship that. Then iterate.
MCP (by Anthropic) is the architectural layer that could make cross-vendor AI agent queries real in multi-system enterprise environments. Andy was direct that we are at chapter two of ten on this. It is not ready to bet your roadmap on yet, but it is worth understanding now because the companies building toward it are not waiting.
On jobs, Andy's framing here was the most grounded I heard in this conversation. AI eliminates the low-level repetitive work, the data entry, the integration monitoring, the spreadsheet reconciliation. It creates space for strategic work. But you will always need a person in the middle making judgment calls. AI should not be making critical decisions without human review. The companies that understand that distinction will actually capture the efficiency gains. The ones chasing full automation will create new categories of failure.
The companies I see moving fastest on AI are the ones who treated data infrastructure as a core business capability and did the boring foundational work everyone else skipped because it was not exciting.
If you are anywhere in the manufacturing, supply chain, logistics, or distribution space, this conversation is worth your time.
Practical GTM Engineering, AI Workflows & Growth Strategy For Tech Startups
Join Convergence, a movement among startup technical founders & operators who are done with scattered tactics & ready to install the growth systems, decisions & leadership that move revenue.