We used our own method.
This site is one result.

1. Start with what we had learned from people.

Our work with senior leaders and middle managers gave us a starting point: the questions people ask, the assumptions they challenge and the difficulties they encounter when changing how work gets done. Where recording was agreed, transcripts helped us revisit those conversations. We used the learning to shape the offer and its explanation, while keeping private client material out of the public site.

2. Give the website a customer and a job.

A CEO, a finance leader and an L&D lead arrive with different pressures. We wanted each to recognise a business situation, find a useful starting point and understand what they would take away. That meant talking about capacity, decisions, customers and outcomes. We kept revising language that explained our tools more clearly than the value to the reader.

3. Build small, review and change direction.

We worked with a generative AI co-worker in Codex to turn the brief into content, diagrams and working pages. It wrote code, ran checks, prepared releases and carried out deployment steps. We supplied the context, reviewed the result and made the decisions. The feedback was specific: the functional wedges needed to stay, the method needed a clearer visual build, and FUNCTION needed to show future vision as well as capacity allocation.

Moving from Wix also meant looking after the less visible parts of the experience: existing addresses, preparation pages, email destinations and both TIP domains. The website was a service to maintain as well as something to launch.

4. Validate the story. Verify the build.

For validation, we asked whether a reader could see their problem, understand the proposed value and find the next step. We asked AI to review the site from different CxO and buyer perspectives, creating an outside-in challenge to our own language. Those simulated reviews surfaced questions to investigate; real customer conversations remain the next test.

For verification, we checked the build, links, mobile layouts, keyboard interactions and the site guide’s answers and limits. We checked that unlisted pages stayed out of public navigation and search, and compared live pages with the intended release. We also challenged the content: an illustrative capacity figure must not become a claimed client result simply because the graphic looks convincing.

5. Keep learning after release.

The first plausible answer was often only a starting point. AI could produce more content quickly; deciding what to remove, what to prove and what mattered to a buyer still took judgement. Visual polish did not settle the evidence question. The next improvement is stronger, client-confirmed stories about why people started, what changed and what they learned.

We reached a useful release, kept a route back to the previous version and retained a backlog. The next decisions should be informed by how people use the site, where they hesitate and the conversations it generates.

The capability is bigger than the website.

Mindset meant curiosity, honest challenge and a willingness to revise our own assumptions. Method brought together customer thinking, AI tools, small releases and validation and verification. The result is a working site and a repeatable way to improve it. In a larger organisation, the same approach connects knowledge workers, product owners and IT around architecture, decision rights, governance and a business outcome they own together.