cost line,
or growth line?
most companies treat AI as a cost — a license here, a pilot there, never compounding. blankcollar moves AI from the cost column to the growth column. you stop paying the hidden tax of context-switching, voice drift and slow ramps. you start compounding instead.
how the math changes
when blankcollar is in your company.
these are not aspirational numbers. they are what comparable teams measure 90 days after install. assumes a 50-person knowledge-work company with $100/hr loaded cost.
the hidden tax compounds.
- $1,000,000 / year lost to context-switching — 50 people × 4 hrs / week × $100 / hr × 50 weeks.
- 8-week ramp for every senior hire; institutional knowledge walks out with every resignation.
- Voice drift across sales, marketing, product — every output sounds like a different company.
- 95% of pilots never deliver measurable P&L. AI stays a cost line.
AI becomes the growth line.
- 4 hours / week reclaimed per senior — every prompt starts with company context, no pasting, no drift.
- Week-3 productivity for new senior hires. memory survives turnover, by design.
- One voice across sales, marketing, product, support — codified from your own writing, not a generic prompt.
- Outcome-priced options at enterprise scale — we get paid when you get paid.
numbers above are illustrative for a 50-person company. drag the calculator below to model your own.
pick the depth.
same five layers underneath.
the difference between plans is how deep we install — and how much help you want. you can move up at your own pace. no rip-and-replace.
already paying for chatgpt enterprise or copilot? see the honest comparison → (most customers keep both running.)
- opinionated industry templates
- 100% self-serve · no contract
- flat monthly · no usage math
- upgrade to business in one click
- 6-week codify sprint extracts your playbooks
- 1–3 packaged routines live in 60 days
- swiss-hosted · audit-ready · multi-model
- per-layer meters · capped, predictable
- 12-month co-led transformation
- on-prem · air-gapped · bring your own AI model
- SSO, SCIM, dedicated CSM, SLA
- outcome-priced — we get paid when you do
a flat $2,500 / month.
plus you only pay for what you actually use.
the platform fee covers swiss hosting, audit, multi-model routing, support and the first envelope of usage. layers above that meter per unit, so price tracks how deep you go — not how many seats you buy.
$2.20 / GB · month
every GB of company memory stored. first 50 GB included.
$2 / skill · month
per active codified skill. first 25 included.
$0.05 / action
one tool action or agent step. first 10k actions included.
$15 / routine · month
per running scheduled routine. first 3 included.
put your numbers in.
see when it pays back.
drag the sliders. the math is conservative — we assume each senior reclaims four hours per week (most measure six). every dollar saved is senior time you stop renting from yourself.
a 20-person company typically invests ~$2,200 / month in blankcollar and unlocks $38,100 in senior capacity — payback in under 2 days.
model your monthly bill.
live estimate, not a commitment. frontier model tokens passed through at +20%.
one memory.
just yours.
for founders, execs, and operators who want the same sovereign memory — wired to their own sources. flat monthly, no usage math. swiss hosted.
- runs entirely on your laptop or home server
- open-weight models only (gemma, llama, mistral)
- nothing ever leaves your device
- guided setup · monthly releases
- for the privacy-maximum user
- hosted in zurich, encrypted at rest
- open-weight + european models (mistral, gemma)
- no us model providers in the routing
- priority support within one business day
- for most founders and execs
- hosted in zurich, encrypted at rest
- routes to claude, chatgpt, gemini — whichever fits the task
- agent features for long-horizon work
- priority support within one business day
- for power users who want the best answer
want the whole story? see the personal page →
three ways to get help installing.
kickoff session
90-min session with the founder. we scope, connect sources, tune to your voice. memory + skills live by end of week. from $2,400 one-time.
book kickoff →codify sprint
6-week sprint to extract your DNA. memory + skills + tools + first 2–4 agents + 1–3 routines live. embedded operator on-site. from $12,000 fixed.
book sprint →transformation
co-led 12-month programme. six capabilities, four phases, named team, outcome-pricing options. all five layers + trust plane as the spine. $80k–250k / phase, scoped after diagnostic.
see programme →pricing, answered.
what counts as an operation?
one ingest event (a document, email, or meeting transcript flowing into the memory) or one query from your AI tool. that's it — no sub-units, no hidden counts. a query that cites five sources still counts as one operation.
can i cap my monthly spend?
yes. set a cap during onboarding. when you hit it, the system switches to read-only — existing queries still work, new ingest pauses, and we email you. no surprise invoices, ever.
how does storage billing work when i delete data?
billed per gb-month. if you delete 10 gb on the 15th, you're billed for half the month on that slice. daily reconciliation, no rounded-up months.
what do frontier model costs look like in practice?
passed through at provider cost + 20%. a typical 200-person team on frontier spends $500–1,500 / month in model tokens on top of the two meters, depending on query mix. every invoice breaks down token spend per provider.
when does billing start?
after setup is complete and your first user signs in. free tier is free forever. paid tiers are pay-as-you-go with no monthly minimum — if your team uses zero operations this month, you pay for storage only.
start small.
compound from there.
free tier on memory. business via the codify sprint. enterprise via the transformation programme. you decide the depth — the OS is the same.