Expertise

Four connected questions shape my work.

A feature changes the user's experience, the system's responsibilities and the business's commitments at the same time. I work through them together, and I link each claim here to work you can inspect.

Human behaviour

Understand what the person is trying to get done and what makes the next action trustworthy. In the Scorecard, the user-facing consent promise became a requirement for every later email.

Product

Choose the problem and revise the commitment when evidence changes. Payment Truth's commercial review changed its role from a potential product to a reference build.

Technology

Make the system's responsibilities explicit. Payment Truth separates what the provider reports from whether an order can be released; the Scorecard enforces consent when a queued message is sent.

Business

Understand who adopts the product, who operates it and what makes it worth paying for. In the current Haylu definition, the first question is which recurring enquiry consumes enough effort to justify a reliable workflow.

AI workflows

AI needs a defined job and a verifiable result.

The model should interpret the request; the business systems should supply the facts. I define what an agent may do, which tools it can call, and how it behaves with incomplete information, conflicting records, failed tools and uncertain outcomes. Quality, response time and operating cost belong in the same product decision.

See the current AI workflow

APIs and developer products

The API contract is part of the product.

A developer needs to know what to send, what the response establishes, and what to do when the result is uncertain. The integration story includes what happens after the first successful request: verification, retries, tenant boundaries and recovery.

Inspect Payment Truth

Technical toolkit

What I've used, and where you can see it.