Give AI the right facts about your business, in a shape it can read.

When someone asks ChatGPT, Perplexity or Google's AI for a supplier, those tools assemble an answer from what they can find and trust. A context system makes sure what they find is true, current and yours.

Book a 20-minute callRun the Two Readers Test

A context system is the structured layer behind your website: schema markup, a consistent entity profile, an AI-readable content map and, if you want it, a knowledge base for your own assistant. It helps people and AI tools describe your business the same way.

What AI reads on your site

People see layout. AI tools see headings, structured data, links and the words around your name. Both should land on the same answer.

The parts of a context system

Six pieces. Most companies need the first four. The last two are for teams with their own AI tools or high stakes on accuracy.

Schema graph

Organization, people, services, places, products and FAQs in JSON-LD, linked into one graph with stable IDs.

Learn the term

Entity profile

One true description of who you are, kept identical on your site, Google Business Profile, LinkedIn and directories.

Learn the term

Content map

An llms.txt file and a sitemap that point AI crawlers at the pages that matter, with plain summaries.

Learn the term

Pages that answer

Each service page opens with a quotable answer, then the detail people need to decide.

Learn the term

Your own assistant

A retrieval setup so your support bot or sales assistant answers from approved facts only.

Learn the term

Answer monitoring

A monthly check of what the main AI tools say about you, with fixes when they drift.

Learn the term

How the work runs

  1. Audit

    Week one. What AI tools say about you today, what your site and profiles say, and where they disagree.

  2. Model

    Week two. Your entities and their relations on one page. Company, people, services, places, proof.

  3. Mark up

    Weeks three and four. JSON-LD on every page type, breadcrumbs, llms.txt and the content map.

  4. Align

    Week four. Profiles and directories brought in line with the same names, facts and links.

  5. Watch

    Monthly. A short report on AI answers and search results, with fixes.

Who it's for

Six kinds of company I build context systems for. The work changes by field. The method stays the same.

Ecommerce

Product facts, prices and availability that AI shopping answers can trust.

Learn the term

Enterprise B2B

Long buying cycles and many readers. One consistent story across every page and profile.

Learn the term

Small and mid-sized businesses

A clean entity profile and schema graph that make a small team look as clear as a big one.

Learn the term

Consumer services

Clinics, studios and local services that AI assistants recommend by name.

Learn the term

Manufacturing and engineering

Specifications, standards and products described precisely enough for technical buyers and their AI tools.

Learn the term

Something else

If AI tools need to describe you correctly, the method carries over.

Learn the term

Also built to order

For teams with their own AI tools. Scoped and priced per project.

  • Retrieval (RAG) setups that answer from your approved documents only
  • Memory systems for assistants that need to remember customers and cases
  • Knowledge graphs of your products, people and places
  • Predictive scoring models for leads and accounts

People decide. Ai fetches. Give both the same true story.

The idea behind every context system

Build it with AI, then run it yourself

Context systems are part of both packages. We set them up together with AI, and your team learns to keep the facts current as products, people and prices change.

  1. Build it togetherWe make the real thing with AI tools, side by side, in your accounts.
  2. Learn the workflowYou see each step, keep the prompts and practise on live changes.
  3. Run it with AIYou manage and improve it yourself. I stay on call when you want a second pair of eyes.

Small and mid-sized businesses

Build and learn

We build the one thing that matters most right now, a website, an online store, a simple app or a workflow, with AI at every step. You work alongside me and keep every prompt, file and playbook.

Usually four to eight weeks. Quoted after a 20-minute call.

  • Goal, audience and a one-page plan
  • Your site, store, app or workflow, built with AI tools
  • A context file that tells AI tools about your business
  • Prompts and playbooks for everyday changes
  • Hands-on sessions for you and your team
  • Thirty days of support while you take over
Start with a 20-minute call

Enterprise teams

Build, train and scale

For companies that want AI in how they build and run digital products. We choose a pilot, build it with your people, then turn what worked into standards, training and guardrails across teams.

A pilot in six to ten weeks, then rollout at your pace. Scoped per programme.

  • Opportunity map and pilot choice
  • A pilot site, store, app or process, built with your team
  • Design system rules AI tools can follow
  • Shared context for your brand, products and data
  • Workshops and role-based training
  • Guardrails for quality, privacy and review
  • Measures for speed, cost and quality
Book a scoping call

Every example, guide and glossary entry on this site is part of the training. Read the insights or browse the glossary before we talk.

Questions people ask

Can you get us recommended by ChatGPT?

Nobody controls that. I make sure that when those tools look, they find clear, consistent and verifiable facts. That is the part you can influence.

Is this just schema markup?

Schema is one part. Most wrong AI answers come from inconsistent facts across the web, so the profile and content work matters as much.

Do we need a new website first?

No. A context system works on most sites. If yours can't hold structured data, I'll say so in the audit.

Is it safe to connect our own assistant?

The retrieval setup only reads documents you approve, and every answer can point to its source.

How soon will we see a change?

Search engines pick up structured data in days to weeks. AI answers change more slowly and less predictably, so I report monthly.

How is this different from SEO?

It overlaps. SEO aims at rankings. A context system aims at being described correctly wherever an answer is assembled.

More on AI search

All insights

What AI can't read about your customers

Context, bias and taste are still human work.

Read the article

Web design statistics

Including how people use AI search today.

See the numbers

Run the Two Readers Test

See your page the way people and AI see it.

Try the tool

Where to next

How to get your site cited by ChatGPT

The five changes, explained.

Read the article

Industries

Context systems by field.

Find your field

Website design

A new site with the system built in.

See the service

See your own site through both readers

A free 20-minute demo on your pages. I'll show what AI search can and can't read, and the three fixes worth doing first.

Book a free demoRun the Two Readers Test
Book a 20-minute call