The reason AI programs stall a year in is almost always the data. Not the model, not the vendor, not the skills gap. We name the problem before the check clears.
The pattern is remarkably consistent. Executive team approves an AI initiative. Vendor lands. Prototype builds against a hand-picked sample. Prototype demos well. Rollout begins. Six months in, the initiative is quiet in leadership updates. A year in, everyone knows the real problem is that the operational data is inconsistent, undocumented, siloed across four systems, or missing entirely for the segment the AI was supposed to serve. The vendor is not going to say this out loud, and the internal team does not want to be the one to.
We say it out loud. Before you sign, ideally. After you've signed, still.
What we produce
- Data-readiness assessment. A written read of the specific data your initiative depends on — where it lives, who owns it, how clean it actually is (measured, not asserted), what gates it has to clear before it can be usefully fed into an AI system. Written in a way both your CDO and your CFO can read.
- Remediation punch list. A prioritized list of the fixes that unlock the initiative, ranked by cost and by expected impact. Includes explicit calls on what to fix now, what to defer, and what to simply live with. Most AI programs try to fix everything and finish nothing.
- Platform read. Where the conversation has turned to a “data platform” buy, an honest read of whether what your operation needs is a platform, a pipeline, a warehouse, a lakehouse, a semantic layer, a governance tool, or none of those. The answer is not the same for every operator, no matter what the reference architecture on the last analyst report suggested.
- Governance-adjacent inventory. A quick pass on which data assets your AI initiative touches carry regulatory, contractual, or reputational exposure — PII, PHI, customer-confidential, board-confidential, non-transferable licensing. Handed to your general counsel and your CISO before it's a problem.
What we don't do
We don't run the remediation. We are not the ones writing the ETL, backfilling the missing columns, or standing up the semantic layer. We produce the assessment and the punch list. Your team, or a build partner you retain, does the work. We can review that partner's approach on request.
We also don't recommend a data platform when the honest answer is that you don't need one. Some operators need a platform. Many operators need a spreadsheet with a written owner and a rewritten SQL query. We say which case you're in.
Readiness assessments we've written have ranged from a single line-of-business at a mid-market operator to an enterprise-wide portfolio touching six ERPs, two CRMs, three data warehouses, and a decade of migration debt. The method scales; the deliverables do not. A five-hundred-page data inventory is a document nobody reads. We keep them short on purpose.
Engagement
Data-readiness work is typically project-shaped — a three-to-six-week assessment, delivered as a written report and a working session with your data leadership. Fee structures are fixed-fee. Follow-on review of remediation progress at ninety days is optional and priced separately.
Introductions happen by referral. If you were pointed here by someone we work with, mention their name when you write. If not, tell us plainly what your initiative is and where your data currently sits. We reply within two business days.