Blog: Data and decisions
Before AI answers the question, agree what the answer means
Published by Forever Foundry
Ask three departments how many orders were delivered on time. Get three different answers.
Now give an AI assistant access to all three reports.
You have made the disagreement easier to retrieve. You have not resolved it.
Trusted answers need agreed definitions
On 13 October 2026, AidenAI announced an expansion of its Data Services, including approved metric definitions, data ownership and answers traceable to their sources.
This is a supplier announcement, not independent evidence of customer results. But it highlights a practical question: what should an AI system treat as the business’s agreed answer?
Read the announcement.
The FOFO™ view
Some apparent data problems are disagreements about how the business works.
Imagine a distributor reviewing delivery performance.
The warehouse counts orders dispatched on schedule. Logistics counts arrivals within the delivery window. Customer service counts complete orders received by the date originally promised.
Each measure answers a different question. Combining the reports without understanding those differences could make performance look better while customers experience no improvement.
The first investment may not be another dashboard or AI assistant. It may be agreement on the customer promise, the evidence needed to measure it and who owns exceptions.
That requires listening beyond IT.
Fix the disagreement that changes a decision
Do not begin with a programme to clean every dataset. Choose one important decision and ask:
- Which figures do teams currently dispute?
- Do they use different definitions, dates or sources?
- Which measure reflects the customer’s experience?
- Who can approve the definition and maintain it?
- What decision becomes possible once the answer is trusted?
Record the time spent reconciling reports and the decisions delayed. Test whether agreed definitions reduce that effort and improve the underlying outcome, such as complete orders arriving when promised.
Disputed figures often sit where one team hands work to another. Read why your next AI opportunity might be hiding between two teams.
AI Compass™ listens across the organisation to uncover friction points, including conflicting information and expectations. Slid3rs™ helps people rank the resulting opportunities by business value and delivery complexity. FlowRoom™ supports leadership agreement on what happens first.
AI can help people find an answer. Leadership still needs to agree which question matters, and what counts as a valid answer.
Find the friction behind conflicting answers. Explore AI Compass™.
Related reading
Next