About Session
Every finance and technology leader in the room is weighing the same build vs. buy decision on AI, and for most of them the build path looks deceptively simple: a data warehouse, a BI tool, Excel or Sheets, and Claude or another LLM doing the analytical work on top. It is a real stack, and it gets better every quarter. It is also where most build vs. buy mistakes happen, because without a governed context layer underneath it, AI hallucinates, confidently and often invisibly. Finance needs more than a fast answer. It needs an audit trail, traceability back to the source, and role-based access controls, and none of those show up automatically just because the model got the number right this time.Christina Ross has sat on both sides of this decision as a three-time CFO before founding Cube. She will walk through the failure patterns she sees repeat across finance and technology teams, and lay out a practical framework for knowing, before the fact, whether building in-house can actually hold up to an audit, or whether it just looks like it can until someone asks.AI has lowered the cost of building. A prototype that used to take a team of engineers now takes an afternoon, which has revived the build case for problems that used to be an easy buy decision.Hallucination is a context problem, not a model problem. Without a governed layer defining the chart of accounts, the formulas, and the hierarchies, AI fills the gaps with a confident guess, and nothing flags it as a guess.The real cost of build does not show up on day one. It shows up on day 400, as maintenance work nobody budgeted for, and again the moment the person who built it all leaves, taking the only working knowledge of how the numbers trace back with them.
