who we are

the company
building blankcollar.

the blank collar is a swiss AI transformation company. we build blankcollar — AI woven into how your company works — and we run the programmes that install it. founder-led. embedded with our customers, not floating above them on a slide deck.

why we exist

most companies are
running AI pilots.
almost none are compounding.

walk into any large company today and you'll find a long list of AI pilots: a chatbot here, a "copilot" there, a marketing copy tool, a code assistant. spend has gone up. impact, mostly, has not. across the industry, the overwhelming majority of generative-AI pilots never deliver measurable P&L — and this is happening on top of the most capable models humanity has ever built.

the easy explanation is that the technology isn't ready. it's the wrong explanation. the technology is fine. the bottleneck is whether your AI gets smarter the longer you use it — whether it remembers, codifies your way of working, and earns the right to act on your systems. tools that can't do that stay flat. companies that wire those things in compound.

we built blankcollar for the second kind of company. and we built it as a way of working — five layers (memory, skills, tools, agents, routines) installed inside your company, shaped by your DNA — because no off-the-shelf product can do the wiring for you. the wiring is the work.

what we believe

five beliefs that
shape everything we ship.

01
AI is an operating posture, not a tool.
the companies that win don't add AI to the org chart — they re-shape the org chart around it. every artifact a company produces (slack, email, decisions, contracts) becomes part of a queryable model of the company itself. that's the posture.
02
capabilities compound. they cannot be skipped.
memory before skills. skills before agents. trust before any of it. you can accelerate the build, but you can't leapfrog. the companies that try to skip layers stay in pilot purgatory.
03
trust is a precondition, not a feature.
no audit trail, no right to deploy. no eval harness, no right to deploy. no kill-switch, no right to deploy. we built the trust plane first because regulated industries are the early-adopter market — and because it's right.
04
the org chart is collapsing into three roles.
builders augmented by agents. owners directly responsible for outcomes. coaches developing humans. the middle of the pyramid is being absorbed by routines. we structure our own company this way — and we help our customers do the same.
05
prototypes beat decks.
we don't sell with slides. we walk into the first conversation with a working prototype on your sources. we run quarterly reviews against measured outcomes, not adoption metrics. speed is now a measurable property of a company, and we treat it like one.
talent: 70 / 70 / 70.
70% of the build in-house. 70% of engineers as doer-engineers, not architects on slides. 70% of the team at senior skill density. that's the talent shape that makes AI compound. we hire for it. we install it inside our customers.
how we run

four operating principles.

01
swiss based
hosted in switzerland by default. jurisdiction is a default, not an option.
02
european focused
built for the shape of regulation that actually governs how european companies operate.
03
sovereign by design
nothing leaves your perimeter unless you decide it should. not on day one, not on day one thousand.
04
multi-model independent
claude, gpt, gemini, open weights — route by task, change with a config.
how we work with you

embedded.
not in a slide deck.

the data is unambiguous: companies that partner with a specialist outperform companies that build AI alone — by roughly 2:1. our model is the partnership done right. we sit inside your business, not above it.

scope first

we map the leverage.

before any code, we map your highest-leverage 2–3 workflows — the ones where AI compounds, not where it looks shiny. you sign a written scope before we install anything.

install fast

first value in week one.

memory layer ingesting on day one. first codified skills by week two. first agent acting under a human owner by week four. measured outcomes from day one — not "we'll know in six months".

hand off

we leave when you can run it.

the goal is not a permanent retainer. by the end of the engagement your team owns the platform, the skills, the agents. portable by design — your DNA, your data, your right to leave.

founder & team

who's building this.

a team of operators, engineers, and domain specialists, distributed across switzerland and europe. hands-on depth in banking, insurance, law, and manufacturing.

kristian kabashi

kristian kabashi

founder · the blank collar

kristian founded the blank collar in 2019 and leads product and go-to-market. he started the company after watching one too many AI pilots stall — the technology was fine, the layer around it was missing. today he works alongside the team to install that layer inside companies that can't afford enterprise overhead and won't tolerate enterprise speed.

the slogan

work is for bots.
life is for humans.

faq

the blank collar &
blankcollar.ai.

what is the blank collar?

the parent company. a swiss-based AI transformation firm operating across europe. we build blankcollar — AI woven into how your company works — and we run the programmes that install it inside companies that need to compound, not pilot.

how is blankcollar different from a chatbot or a "copilot"?

a chatbot answers questions. a copilot assists one person. blankcollar installs five layers inside your company — memory, skills, tools, agents, routines — that learn the way you work, act on your systems under accountable humans, and compound the longer you use them. it's not an app you bolt on; it's how the company runs.

why have most AI initiatives failed to date?

three reasons that compound. one — pilots are scattered across teams instead of focused on the highest-leverage 2–3 workflows. two — the AI doesn't retain feedback or codify the company's way of working, so it stays generic. three — companies build alone instead of partnering with specialists, and the success rate of internal-only builds is roughly half that of partnered builds. blankcollar fixes all three by design.

are you a software company or a services company?

both, on purpose. the platform is software you can run yourself. the programme is a partnership: embedded operators, codify sprints, transformation engagements. the data shows that platform + partnership outperforms either alone — so we sell both, and we let customers buy the depth they need.

what about the back office vs the front office?

most companies put their AI budget into sales and marketing. most of the measurable ROI shows up in operations, finance, support, and HR — back-office work that runs on documents, decisions and recurring tasks. we tell customers honestly where AI compounds first. usually it isn't where they were planning to start.

where are you based and where can you deploy?

switzerland-based. europe-focused. swiss hosted by default. on-prem and air-gapped deployments available for regulated customers. we operate where the regulation actually governs how companies run — FINMA, GDPR, DSG, HIPAA — not where it's most convenient to host.

can we use the platform without the programme?

yes. the self-serve startup plan is fully online — you install yourself in an afternoon. the business plan can be self-installed too, though most customers take the codify sprint to get more value faster. the enterprise programme is for companies that want to compress an operating-model change from years to months.

what if we leave?

your data is yours. your skills are yours. your agents are yours. the platform is portable by design — markdown source-of-truth, open standards, exportable in 24 hours. we are explicitly designed to be left behind, not locked into.

work with us.

90 minutes. a working prototype on your sources. a written scope within 48 hours. you decide if we're the team.