The Diagnostic starts with a business problem or desired improvement, not with a pre-selected technology. It combines adaptive AI-assisted interviewing with Oakmere-controlled calculations, domain-specific process packs, evidence provenance and explicit uncertainty.
Oakmere does not force every opportunity into an AI category. The assessment can point toward one or more of the following solution patterns:
After the solution pattern is established, Oakmere identifies the technology capabilities that are likely to be required, for example workflow orchestration, APIs, RPA, document intelligence, conversational interfaces, analytics, AI reasoning or human approval controls. This is capability-level guidance, not a vendor recommendation. Named vendor selection belongs in deeper qualification where architecture, security, licensing, integration and procurement constraints can be considered.
Oakmere uses process-specific domain packs where the assessment needs more depth than the Universal Framework alone. A pack is only described externally as an Oakmere Specialist Pack after it has passed defined build, evidence, testing and practitioner-review gates. Packs still being developed or validated are not presented as specialist authority.
Validated packs contain process-specific questioning, evidence requirements, value drivers, risks, controls, solution patterns, technology implications, contra-indicators, regression tests and versioned review history. Real assessment evidence is then used to improve live packs over time.
Oakmere separates current cost/impact, addressable value and realistically capturable value. Labour addressability is derived from process evidence such as repeatability, exception patterns, data quality, judgement and control requirements. A customer's own savings estimate is shown separately and is not simply accepted as the Oakmere result.
An AI model interprets free-text descriptions, classifies the opportunity, selects useful follow-up questions and drafts parts of the narrative. Oakmere code performs the financial calculations, applies value logic and controls payment state. The model is instructed to begin with the transformation need rather than assume AI is appropriate. In high-stakes domains, the system restricts itself to process/value/readiness analysis and requires specialist validation.
Relevant standards and recognised frameworks are used to ground governance and process-design considerations. They are not used as proof of Oakmere's ROI assumptions.