Every stage of the AI Value Worx process draws on the same underlying resource: a continuously maintained, actively researched body of knowledge about what AI has achieved and what governing it requires. Without it, the process is a methodology running on opinion. With it, every recommendation, every business case, every governance decision is grounded in verified, structured, cross-referenced evidence.
The Opportunity domain contains evidence about what AI has actually achieved: specific organisations, specific initiatives, verified outcomes. It also contains analytical knowledge — expert research about how and why AI value gets realised, what conditions it requires, and what separates the initiatives that scale from those that stall.
The Governance domain contains the regulatory and policy landscape: what is legally required for your sector and initiative type, what expert consensus recommends, and how governance obligations translate from board intent to technical implementation — at every level from the EU AI Act to deployment security guidance.
Both domains contain two categories of evidence. Hard cases are specific and verifiable: a named organisation, a named initiative, a measurable outcome with a traceable source. Analytical cases represent accumulated expert understanding — richer and broader than a single case study, drawing on research syntheses and pattern findings across the field.
Hard cases answer “what happened?” Analytical cases answer “why does it happen, and what should we do about it?”
During initial seeding, a one-way transfer of reference evidence moves from our systems to yours. During ongoing extraction, the Active Intelligence Builder reaches only public sources — published case studies, regulatory documents, research publications. At runtime, the Opportunity Analyser queries your local Intelligence Base directly — no external API calls, no cloud lookups, no data egress.
Your AI Value Intelligence is not a window into our data. It is your data — running inside your infrastructure, maintained without external dependency, compounding in value as your organisation uses it and as the field evolves.
Every claim stored in the AI Value Intelligence traces to a source quote, a captured text file, and an original URL. This is not a documentation standard — it is enforced by the extraction pipeline’s verification agent. A claim that fails source verification does not enter the corpus. There is no path in for an unverified claim.
When the Opportunity Analyser surfaces evidence, the provenance chain is visible. The question “where does this come from?” always has a traceable answer: the source, the quote, the classification rationale.
Every claim is classified against a shared ontology: a structured set of dimensions that describes what a claim is about, what kind of outcome it describes, and under what conditions it applies.
This is what makes the Intelligence Base queryable rather than merely searchable. The Opportunity Analyser does not search for matching text — it queries for claims that match a client’s context on the dimensions that matter: mechanism, outcome type, transformation depth, APQC work function domain, sector.
The ontology evolves as the corpus grows. Each evolution sharpens the system’s ability to make distinctions that matter. The Structured Ontology Document is a deliverable of the engagement: not just the ontology itself, but the full record of why it is shaped the way it is.
Every response traces to verified evidence in the corpus. No assertion is unsourced.
At the end of the engagement, your organisation takes ownership of: