Our IT department wants to build a custom AI integration layer to connect our legacy ERP to modern LLMs, but our leadership team thinks we should wait for our software vendors to release native updates. How do we resolve this build versus wait dilemma without hurting our short-term Rocks?

Question

Tyler Smith · Accepted Answer

This is a classic operational trap that can stall your traction. Building custom integrations requires significant cash and executive focus, while waiting for legacy software vendors to release native AI updates can leave you trailing your competitors.

To resolve this, use the IDS® process during your next weekly Level 10 Meeting™ and apply a simple rule: only build custom integrations if the workflow in question directly impacts your core differentiator on the V/TO®.

If the integration merely automates routine back-office tasks like invoice processing or basic data entry, do not spend custom development dollars. These are standard operational functions. Wait for your legacy ERP vendor to release their native AI updates, or use simple, low-code middleware to connect them in the meantime. Do not let custom software development distract your Integrator or derail your quarterly Rocks.

However, if integrating your ERP data with an LLM allows you to deliver a unique, high-value experience to your target market that your competitors cannot replicate, then building a proprietary layer is justified. It builds real IP that increases your company valuation. If you choose this path, assign a clear owner to the integration seat on your Accountability Chart and manage the project milestones strictly through your weekly Scorecard. This keeps your team focused on execution without blowing up your budget.