AI Readiness Diagnostic

Is the core data the business runs on brought together in one place, rather than living only in the systems that create it?

Is there a central data store, such as a warehouse or lakehouse, that holds the core business data, even if it is not yet fully governed?

Does data move between your main systems automatically, rather than by someone exporting and re-keying it?

Is data movement between key systems automated through scheduled pipelines or APIs, with manual transfer the exception?

When a data feed stops or breaks, does your team find out before the people relying on the numbers do?

Are pipelines monitored with alerting and error handling, so failures are caught before downstream users see them?

If a number in a board report is questioned, can you show exactly which systems and steps produced it, without someone reconstructing it by hand?

Is data lineage captured automatically by the platform, from source to report, rather than documented by hand?

Do the core reports each function depends on arrive on a fixed, predictable schedule?

Are the core operational and management reports produced on a scheduled, repeatable basis rather than assembled on request?

Do people open dashboards to run their part of the business without being asked to, rather than waiting for a report to be sent?

Are dashboards tied to each function's KPIs in regular, unprompted use across teams?

When someone needs a new view or metric, can they get it themselves or within days, rather than joining a weeks-long queue?

Can business users self-serve new questions on governed data, without raising a ticket with the data team?

Do you know which reports are actually used and which decisions they inform, and do you retire the ones that aren't?

Are BI usage and business impact measured and reviewed, with unused assets retired and investment adjusted accordingly?

Do your key figures, such as revenue, active customers or headcount, have one agreed definition that every team uses?

Are the critical business metrics and datasets defined in a maintained data dictionary or metric layer?

When a number turns out to be wrong, is there someone whose job it is to fix it at the source, rather than each team correcting its own copy?

Are data quality issues logged, triaged and fixed at the source by a named owner, rather than patched downstream?

Are errors in the data usually caught before they reach a leadership decision, rather than discovered afterwards?

Is data quality monitored proactively, with rules, thresholds and alerts on the datasets that matter most?

Is the quality of your critical data measured and reported over time, like any other operating metric?

Are quality results tracked over time against targets or SLAs, with cleansing and standardization automated for critical data?

Do you produce forecasts, such as demand, revenue or cash, on a set cycle and compare them with what actually happened?

Are forecasts produced on a defined cycle, with accuracy tracked against actuals?

Does at least one forecast or recurring decision rely on a statistical or machine learning model, rather than judgment and last year's numbers?

Is at least one forecast or decision driven by a statistical or machine learning model rather than judgment alone?

Is a predictive model or AI running inside a real business process today, beyond a pilot?

Is at least one model in production informing a recurring operational decision, with monitoring and retraining in place?

When a new analytics or AI idea is approved, does it typically reach live use in weeks rather than becoming a project of many months?

Can a new analytics or AI use case go from concept to production in weeks, on established platform patterns?

Has leadership decided which parts of the business AI should be applied to first, rather than leaving teams to experiment on their own?

Has leadership set AI priorities that name the first processes or use cases, rather than a general intent to use AI?

Do employees have clear guidance on what they may and may not do with AI tools, covering privacy, confidentiality and what data they can share?

Are AI-use policies documented and communicated, covering privacy, confidentiality, IP and generative AI tools?

Is the value of your AI efforts measured against a baseline, such as time saved, revenue or quality?

Are AI initiatives measured against a defined baseline, with outcomes reviewed, rather than judged on activity?

Where AI influences decisions that matter, such as credit, pricing or hiring, is there a defined point where a person reviews before it takes effect?

Do consequential AI-influenced decisions have defined human-in-the-loop checkpoints, with overrides logged?

Has leadership named the specific business outcomes better data is meant to improve this year, such as margin, working capital or service levels?

Is there a documented data strategy that ties the data roadmap to named business objectives?

Is one executive clearly accountable for the company's data: its quality, who can access it, and how it is used?

Are data owner and steward roles formally assigned for the key data domains, and actively held?

Are there written rules on who may see sensitive data and what it may be used for, and are they followed in practice?

Is access to sensitive data granted through a request-and-approval workflow, with policies enforced in the platform?

Does a leadership forum review data and AI investments, risks and policies on a regular cadence, with the authority to decide?

Is there a data and AI governance body that meets on a cadence, owns policy and has budget authority over the roadmap?