ai solutions
in weeks, for a fraction.
Software, websites, web apps. What classic agencies take 9 months and a six-figure budget to build, we ship in 2 months - for a fraction. No throwaway prototypes, but a production stack from day one, in your repo, with docs.
this is what ai builds usually look like.
Three numbers that frame the market - and a fourth where we run against that frame.
typical duration for a custom business app at classic agencies.
usual team size per project - with the matching overhead in coordination, status reports and handovers.
budget anchor for comparable custom apps in the DACH mid-market.
2–3 months. 1–2 seniors. A fraction of that.
is this for you?
book a discovery callone senior. not a team of five.
Classic agencies need a team for each project: project lead, two engineers, a designer, someone on QA. Three quarters of the time go into coordination, status reports, meetings, handovers.
We work differently. One experienced person, equipped with a modern AI-assisted development toolchain, builds what would have required a full team two years ago. That saves three things at once:
time
no internal interfaces, no waiting between roles.
money
no team overhead, no multi-headcount for a single project.
quality
one person holds the entire stack in their head, instead of five who keep explaining it to each other.
- teamclassic agency
typically 4–6 people
tokyn1–2 seniors
- mvp durationclassic agency
often 3–6 months
tokyn2–4 weeks
- custom app durationclassic agency
often 9–12 months
tokyn2–3 months
- budgetclassic agency
typically six figures+
tokyna fraction of that
- coordinationclassic agency
weekly status reports
tokynweekly live demos
This doesn't mean less rigour. It means less friction. Code gets reviewed, tested and documented the same way as for any production system. The difference shows up in the price and the timeline - not in the result.
concrete. not abstract.
mvps in 2–4 weeks
From concept to a prototype you can test, demo, decide on.
full web applications
On a production stack - no rebuild for scale after six months.
websites and marketing sites
With AI-powered features (personalisation, search, lead qualification).
integration into your systems
ERP, CRM, data warehouse - whatever is already there.
ai features in existing products
One concrete use case, cleanly integrated, no rewrite.
clean handover
Incl. docs, repo access, team enablement - you run it yourselves, or we stay on.
three examples. what it might look like.
Mid-sized custom-machine builder, manual quote calculation, long lead times.
gebaut: Internal tool for AI-assisted quote calculation.
Quote turnaround from days to hours.
Built in a few weeks. A classic agency quote typically lands at several months.
Growing SaaS company, support function not scaling with customer growth.
gebaut: AI-powered search and auto-tagging in the help centre.
Support tickets visibly down in the first quarter.
Built in a few weeks. Internal effort would have been several months.
Freight forwarder with heavy manual workload on delivery-note intake.
gebaut: Web app for AI-assisted delivery-note intake.
Processing time per delivery note clearly down.
Built in a few weeks. Classic offer: several months.
ready to get concrete?
book a discovery callpragmatic. not hype-driven.
We pick the stack by three criteria - in exactly this order:
- 01use case - what does the thing need to do, now and 18 months out.
- 02compliance & data - what may be processed where, what has to stay in the EU.
- 03your team - what can you sensibly operate yourselves after handover.
What does not feed into stack decisions: what's trending on LinkedIn, what got hyped at the latest conference talk, what our favourite toy happens to be.
- We build on established open-source frameworks so you're never trapped in a platform.
- AI models are picked per use case - state-of-the-art where it makes sense, EU-hosted where compliance demands it.
- On request we host entirely in the EU, including all AI inference.
- Production stack from day one - no throwaway no-code prototype that has to be replaced after three months.
Throwaway prototypes in no-code tools that have to be swapped for a "real system" the moment real users show up. What we build still runs in 18 months - and can be picked up by any other engineering team.
The stack proposal for your project lands in the concept document after discovery - including reasoning per component and the alternatives we ruled out.
no workshop theatre.
discovery
Use case, data situation, systems and target KPIs. We talk to the people who'll actually use it.
Discovery doc with assumptions, risks and first architecture sketches.
concept
Scope, architecture, effort and risks on a single page. A clear decision, not a slide deck.
Fixed-price offer, timeline, milestone plan.
build
Iterative, runnable from week one. Production stack from the start. Weekly demos, no status-report theatre.
Repo access, staging environment, weekly demo calls.
handover & enablement
Clean handover, docs, team enablement. You run it yourselves - or we operate it under retainer.
Code ownership, technical docs, runbook, onboarding for your team.
On request we'll happily run the whole thing for you - maintenance, updates, evolution under retainer. Separate contract, no minimum-term games.
fixed price. fraction of usual.
We work on fixed prices per project - no T&M gamble, no hidden cost. Range indication after discovery, final price before build.
What you get from us typically lands at a fraction of what classic agencies charge for comparable scope. How small that fraction is depends on the competing offer - usually we're talking orders of magnitude, not percentages.
mvp
2–4 weeksOne concrete use case, runnable, in your repo. Ideal for testing, showing, deciding.
custom app
2–3 monthsFull application, integrations, roles & permissions, production-ready. Multi-user, real data.
ai feature in existing product
scope-dependentOne AI feature cleanly integrated into a shipping product. No rewrite.
Ongoing operations, new features after handover, token costs of the AI providers. On request we cover that under a separate retainer - its own contract, its own terms.
more in the journal.
let's talk about ai solutions.
30 minutes, no pitch deck. We look at your concrete situation and tell you honestly whether - and how - it's worth doing. If not, we'll say so.