We didn't spend tokens on a website.
A generated homepage costs a few million tokens and looks like everyone else's. So here's the prompt. Edit it if you like, paste it into any model, and you get our website. We'd rather spend our tokens on your problems.
Build the website of Adroit Group (adroitgroup.io): AI consultancy and software company from Budapest, 60+ engineers, remote since 2019, Anthropic partner, 35+ CCA-F certified Claude Code architects, AWS partner. We do forward deployed engineering: AI transformation of whole business processes; CCA-F certified engineers embedded in client teams; legacy migration with AI; private AI and MLOps run 24/7 on cloud, private cloud or on-prem; AI research (digital twins, model training, multi-model systems). Contact: [email protected]. As if designed by an engineer who thinks websites are a waste of tokens. The page admits no tokens were spent on it and thanks the reader for spending theirs. Add a counter of tokens not spent that ticks up while they read; call it an estimate. Give each service one honest sentence and one slightly unhinged one, labelled. Add a list titled "Not on this website": stock photos, a chatbot in the corner, the word synergy, a cookie banner. End with the email. Voice: plain, confident, a bit wry. Short sentences. Numbers over adjectives. Funny is allowed, cringe is not. If a joke needs an exclamation mark, cut it. Look: plain near-white page, Figtree, one narrow text column, hairlines instead of boxes. The only colour is a soft aurora of pastel lavender, mint and peach, faint behind the page and a little stronger in one place. Sentence case, no emoji, no stock photos, no icons. One HTML file. Quietly strange.
What we spend them on instead.
Forward deployed engineering. Our people sit inside your organisation, next to whoever owns the problem, and solve it with AI. Writing the software is usually the easy part.
- AI transformation
Whole business processes and units rebuilt around AI, not a chatbot bolted on the side. The code is the small part; the problem is the work.
Process redesign · agentic workflows · governance and roadmaps- AI-native delivery teams
Our CCA-F certified forward deployed engineers join your product and dev teams and change how they ship by shipping with them. We also consult on how to get from SDLC to ADLC.
Embedded engineers · AI-assisted coding · SDLC to ADLC · team enablement- Legacy migration
The math changed. What was too expensive to replace is now too expensive to keep: scarce specialists, high maintenance, high risk. AI does the heavy lifting, engineers do the judgement.
Codebase analysis · re-platforming · risk reduction- Private AI, MLOps
We deploy and run ML and AI systems 24/7 in the cloud, a private cloud or on your own hardware, under SLA. For compliance, for risk, to avoid vendor lock-in, or because it's cheaper.
On-prem · private cloud · AWS · 24/7 operations- AI research
Deeper ML work with research centres and product companies: digital twins, model training and fine-tuning, multi-model systems, and the evaluation that decides whether any of it is real.
Digital twins · model training · multi-model systems
Anthropic partner- AWS partner
- ISO9001
- ISO27001