Insights & Ideas

A neutral executive guide to the three regulatory routes a UK retailer can take to launch a digital wallet on a banking-as-a-service stack: acting as an agent, partnering with an authorised e-money institution, or seeking its own authorisation. Includes who carries which obligations and what to confirm with the FCA's published guidance.

A neutral comparison for retailers and platforms deciding between a fintech specialist and a generalist agency to launch a wallet, card or BNPL product, with a decision checklist you can apply to any partner.

Reported pilots suggest accounts payable is one of the first places AI agents will release real payments. Here is a controls-first guide to what finance leaders should check before any agent is allowed to move money, and why a scoped proof of concept is the sensible way to start.

The agentic-commerce debate has moved from whether agents can pay to who authorised the payment and who is liable if it goes wrong. A UK explainer for retailers, platforms and risk teams: how authority is delegated and proven, where liability sits under existing frameworks, and what to implement first.

Buyers use different words for overlapping ideas, and the search data shows which ones UK buyers actually type. A short guide to agentic payments, agent payments, agentic commerce and delegated payments, and how each maps to the payments stack.

A buyer's guide to AI consulting for UK financial services and fintech-adopting businesses: how use cases are triaged, what a proof of concept should prove, where governance fits, and how to tell a decision-ready engagement from a slide deck.

On 19 October the FCA published good and poor practice for AI and asked AI providers to add disclaimers to AI financial advice, while applying existing frameworks rather than writing new rules for agents. Here is what that means for the teams building regulated products, and a control checklist to work from.

Accounts payable is where AI agents are beginning to release real payments, which makes it the natural place to run a proof of concept. Here is a framework for what to test, what to measure and the rule that tells you when to stop or scale.

What is graph engineering in AI, and when does it beat a prompt or a feedback loop? A practical guide to nodes, validation gates, parallel work and recovery paths.

When should an AI workflow be a fixed chain of prompts, and when does it need an agent? A practical guide to cost, debugging and hybrid architectures for production AI.