Fix the inside. Win the outside.
The pressure comes from two sides. Inside, teams burn close to two hours a day hunting for information they already own. Outside, buyers ask AI before they ask you, and it describes you whether you shape it or not. We get your own knowledge in order, and make AI show you accurately.
Three problems. We fix all three.
01
Buried in your own files
Years of documents, scattered across tools. Nobody can find anything at the moment they need it, and search does not help because the data underneath is a mess.
02
The AI demo flopped
The pilot impressed everyone. In production it started giving confident wrong answers, because the data it was reading was never cleaned up.
03
AI can’t see you
Someone asks an AI who to work with in your market. Your competitor gets named. You do not come up, because there is nothing structured for the model to read.
The inside.
Everything your company knows, files, contracts, numbers, email threads, made findable, safe, and useful. Here is what that means in practice, without the jargon.
Ask your files anything. Watch.
Sales
Finance
Legal
HR
The technology is called RAG. What you see: answers with receipts, from your own files.
How many of these are true at your company?
A private self-check. Nothing is sent anywhere.
Tick the ones that are true for you.
Fix the inside.
Six things we do, and what changes for you.
01
Ask your documents a question
Type 'What did we agree with this client in 2024?' and get the answer with the exact document behind it, instead of digging through folders. The technology behind it is called RAG.
> What did we agree with this client in 2024?▌
02
One home for your files
You ask for the latest client contract and get three versions from three people. We set up one place where the newest version is always on top, whether you work in SharePoint, Notion, or Google Drive.
03
Safe AI tools for the whole team
Your team already pastes company data into ChatGPT. We set up ChatGPT, Copilot, or Claude properly, so nothing confidential leaks and you can see who uses what.
04
Automations and AI agents
The same offer gets typed into Salesforce, then into SAP, then into an email. We connect your systems and build AI agents that take over that copy-paste work.
05
Clean, consistent data
When the report says one thing and the system another, nobody trusts either number. We find the duplicates and wrong entries, fix them, and fix the process that created them.
06
Team training
Hands-on training in plain language: what these tools can do, what they must never do, and what the EU AI Act requires from you (Article 4 included).
What you gain
95%
of AI pilots stall on messy foundations. We fix the foundation first, so you land in the other 5%.
MIT, 2025 ↗78%
of organisations already run AI. The winners run it governed, on clean data, with a trained team.
Stanford HAI, 2025 ↗What searching costs you, today.
Hours lost per year
9'900
Cost per year
CHF 742'500
Set the three sliders to your reality. Hours per year assumes 220 working days.
We plug into what you already use.
And when something is custom-built or twenty years old, we connect that too.
The outside.
Get seen and quoted where AI engines and buyers now decide who to trust. Skip it, and that answer describes a competitor instead of you.
What we do on the outside.
01
Amplify your content
We take what your Creative Work team makes and put it in front of more of the right people, through channels most agencies ignore.
02
AI-visibility baseline (GEO)
We check how ChatGPT, Perplexity, and Google AI describe you today, and fix what they get wrong.
03
Machine-readable web
We prepare your site so AI engines can actually read it: llms.txt, structured data, and listings like Product Hunt.
04
Presence on sources AI trusts
We build an accurate presence on the places AI engines rely on: Reddit, forums, reviews, and Wikipedia.
05
Smart advertising
We use paid placement only where it measurably raises how often you show up.
06
AI answer monitoring
Each month we check how ChatGPT, Perplexity, and Google describe you, and correct what has drifted before your buyers see it.
Be the answer, not a footnote.
When someone asks about your market, an AI gives them an answer. We make sure you are in it, and described correctly.
What's at stake.
42%
of adults already use AI chatbots to search for information. That answer is your new first impression.
Pew Research, 2026 ↗10+
channels a B2B buyer now uses to interact with suppliers, up from five in 2016. Most sit outside your website.
McKinsey ↗1st
impression now happens inside an answer box you don't control. Unless you shape what it reads.
Whether inside or outside, never a leap of faith.
Same steps for both sides. The pilot pays for itself. You walk away with a plan you can act on, with us or without.
01
Pilot
We assess where you stand, inside, outside, or both, and hand you a plan: a data map, a prioritised roadmap, a risk register, and a proposal. Useful even if you stop here.
Scoped
02
Build
We put the plan into action: clean-up, systems, and integrations for the inside; llms.txt, earned surfaces, and amplification for the outside.
Project
03
Run
We keep it working: monitoring that catches drift, governance, upkeep, and a monthly impact report. Cancel-friendly.
Monthly
Our guarantee
If the pilot does not deliver what we scoped, you do not pay for it.
Every engagement starts with a scoped pilot. We write down exactly what you get before you sign anything. If we do not deliver those things as described, you get your money back. We can offer that because we control the scope, and because we would rather lose a fee than lose a reference.
Questions companies ask us
How does an engagement start, and what does it cost?+
With a paid pilot: a full assessment, a scoped plan, and a sandboxed proof, at a scope we agree in writing before you sign. You walk away with a map and a plan you can act on, with us or without. Everything after that is priced on the work, in a written proposal.
What if it does not work?+
The pilot is fixed scope, written down before you sign. If we do not deliver those deliverables as described, you get your money back. That is the entire risk you carry.
You are new. Why take the risk on you?+
Because we price like it. Early engagements are fixed scope, money back if we miss, and we over-deliver on purpose, because your result is what we sell next. You are getting a team that needs the reference more than it needs the fee.
What if our data is a mess?+
It usually is. Twenty years of duplicates, half-scanned PDFs and three versions of the same contract is the normal starting point, not the exception. That is exactly why the assessment comes before any build quote: we find out what we are walking into before either of us commits to a number.
Is our data safe with you?+
Yes. An NDA and data terms come first, we work on a sandboxed copy so nothing live is touched, and everything aligns with the Swiss nLPD and the EU AI Act. We touch only what the engagement needs.
We tried an AI tool once and it disappointed. Why would this be different?+
Most AI fails on messy inputs, not weak models. We fix the foundation first, on a sandboxed copy, then build only what the assessment justifies. No clean data, no AI.
Will this bury our team under new AI tools and workflows?+
No. We implement AI only where it earns its place and removes work, never workflow sprawl for its own sake. If a tool does not pull its weight, it does not ship.
Do we have to do both the inside and the outside?+
No. Take the inside, the outside, or both. Each works on its own: the inside puts your knowledge in order, the outside makes AI and search describe you accurately.