Should lenders use ChatGPT or build owned AI?
ChatGPT is the right tool for internal productivity (drafting, summarising, ad-hoc analysis). It is the wrong tool for the regulated, borrower-facing, differentiating layer — decisioning, servicing, collections, portal experience. Those must be firm-owned systems with an architected compliance envelope, or Consumer Duty and CONC survivability disappear.
ChatGPT is the right tool for internal productivity (drafting, summarising, ad-hoc analysis). It is the wrong tool for the regulated, borrower-facing, differentiating layer — decisioning, servicing, collections, portal experience. Those must be firm-owned systems with an architected compliance envelope, or Consumer Duty and CONC survivability disappear.
The right split
The false choice is 'ChatGPT versus owned AI'. The correct answer is both, with a hard boundary. ChatGPT accelerates every internal workflow. Owned AI runs everything a regulator, an auditor or a competitor could look at.
Related questions
Can ChatGPT touch borrower data?
Only under enterprise agreement with strict data-residency and no-training terms — and even then, not for decisioning or Consumer-Duty-visible interactions.
Cost comparison?
ChatGPT is cheap per user. Owned AI is a capex investment that compounds. Different budgets, different roles.
Vendor lock-in?
Owned AI is model-agnostic — the compliance envelope, retrieval and evals stay stable across model providers.
SMF ownership?
Both are SMF-accountable, but the accountability is meaningful only for the owned layer.
Where does the Diagnostic fit?
The Diagnostic maps which workflows should be which for your firm and produces the costed build plan for the owned side.
The AI-first lenders are pulling away on unit economics, not on rate. Every quarter you defer the architecture is a quarter of compounding disadvantage.
The £15k AI Diagnostic maps your lender stack, prioritises the systems that pay back fastest and produces a costed sequenced build plan.