Our approach
Not a service we sell. Part of how every project is done.
From the first research to long after launch, the tooling examines your market, your site and your data page by page, and ranks what it finds by expected effect and effort. Our specialists check every finding and decide what happens next.
The three pillars of AI in our work
- Deeper analysisEvery page and query, not a sample
- Clear prioritiesRanked by effect and effort, agreed with you
- People decideEvery finding checked by a specialist
Scope
Where the tooling helps.
- 01Search analysisSearch intent, competitor gaps, structured data, and whether AI search such as ChatGPT and Google's AI Overviews mentions you.
- 02Site auditsTechnical health, speed, accessibility and content quality, page by page.
- 03Conversion optimisationYour own data on where visitors hesitate and drop off, turned into improvements that bring in more customers.
- 04Data analysisAnalytics and Search Console explained in plain language, with patterns and anomalies flagged.
- 05Market & industry researchCompetitors, demand and trends in your market, from open data.
- 06Research synthesisInterviews, reviews and customer questions turned into themes for brand and design work.
- 07Quality checksChecks before launch, redirect plans for migrations, and a comparison with the baseline after.
Principles
How we use it.
The same rules on every project, whatever the tooling finds.
- People make the decisionsThe tooling suggests; a specialist reviews and decides.Guards againstChanges nobody can explain.
- Checked before it reaches youNothing goes into a plan or onto a live site without review.Guards againstUnchecked AI output in front of you or your customers.
- Clear about limitsAI can be wrong. Every finding is marked as confirmed or to verify, with where it comes from.Guards againstConfident conclusions on thin evidence.
- Traceable to the sourceEvery finding leads back to the data: analytics, Search Console, the site or open market data.Guards againstAdvice from a black box.
- Only the data you allowYour data is used only with your permission, for research, analysis and optimisation on your project, and never to train AI models. Everything else comes from open data.Guards againstClient data used where it shouldn't be.
AI in your own product
Building AI
into your site.
The same principles, applied to your own product. It's a different kind of project: scoped on its own, with its own requirements, and running costs that depend on how much it is used. We go through all three with you before anything is built.
AI assistant
- Best for
- Sites and products with many recurring questions and a solid base of content to answer from.
- Less suited for
- A small site with few questions; a clear FAQ serves better.
AI search
- Best for
- Large sites, catalogues and knowledge bases where people search in their own words.
- Less suited for
- A site of a few pages, where good navigation does the job.
Automations
- Best for
- Repetitive work across forms, email and your systems, with clear rules for when a person steps in.
- Less suited for
- Processes that change often or depend on judgement at every step.
Have an idea for AI in your own product? We scope it with you first.
Talk to us about itProcess
From the first research to after launch.
The tooling is part of every step, each leaving something you can see and use.
01Research
Your market, your competitors and what people search for, before anything is designed.
You receiveResearch findings
02Audit
The tooling goes through the current site, its content and its technical health, page by page.
You receiveAudit findings
03Decide
Our specialists review every finding, add what the tooling can't see, and set the priorities with you.
You receivePrioritised plan
04Build
We design and build what the plan sets out, and check it before launch.
You receiveLive changes
05Optimise
After launch we keep analysing how visitors behave, and improve what turns more of them into customers.
You receiveOngoing analysis and improvements
After launch
Launch is where
the real data starts.
Once the site is live, the tooling keeps reading how visitors behave: what they search for, where they hesitate, where they leave. We turn that into improvements, measure them and keep going, so more visitors become customers. The analyses are shared with you alongside each plan.
- Ongoing analysis of visitor behaviour
- Improvements ranked by effect and effort
- Every change checked before it goes live
- Analyses shared alongside each plan
Frequently asked questions.
No. It's part of how we work on every project, not an extra line on the proposal.
No. It analyses and suggests. Writing, design and every decision stay with our specialists.
Your own data, such as analytics and Search Console, only with your permission and only for your project. It is never used to train AI models. Market, industry and competitor research uses open data.
Yes. The analyses are shared alongside each plan, so you can see what every decision is based on.
Yes. That's why every finding is traced to its source and checked by a specialist before anything changes.
No. After launch we keep analysing how visitors use the site and improve it, so more of them become customers.
Yes, as a separate project: an assistant, AI search or automations, each scoped on its own.
Mainly usage: how often the feature is used and how much it has to process. We estimate this with you before anything is built.
You dream it, we build it
Walk in with an idea, walk out with a website that moves you and your business forward.