Unit6 × Scalably
One company.
One system.
Unit6 has been running with Atlas for half a year — in every channel, next to every person. The next step is not another tool. The next step is one place where everything that already works comes together: HQ.unit6.
Scope for October 2026 · prepared 20 September 2026
Where we are
A lot already runs.
In fifteen different places.
Slack, Asana, Drive, Klaviyo, Shopify, Gorgias, Sheets, Figma, Cloudflare… every task crosses three or four of them and fifteen clicks. The information exists — it is just never where the person is at that moment.
The work is done.
The system isn't one yet.
Counts measured live on the Unit6 server: activity 18 Aug – 17 Sep 2026. Only what was actually used is counted.
The principle
AI doesn't replace people.
It changes where their time goes.
The goal is not fewer people. The goal is the same team doing two to three times more — because the routine runs through Atlas, and the person decides, checks and signs. Whoever uses it, grows. Decisions about people stay yours.
decisions made without a human — the system prepares, a person decides and signs
new tools to learn — Atlas is already in Slack and Telegram; HQ is one more door to the same operator
data outside your control — your server, your data, written access rules before the first byte
How it fits together
One Atlas.
Many doors.
Slack, Telegram, WhatsApp — and from October, HQ. Every door opens onto the same operator with the same memory, skills and rules — access scoped to each person.
Not a promise — an inventory
This already works.
Today. In production.
Every line below was checked live on 16 September 2026. Live capabilities; Partnerships is a preview.
8 personal operators in use
Vukadin, Ilija, Katarina, Dimitrije, Filip, Aleksandar, Pavle and the design seat — each with their own Atlas in Slack or Telegram, used every week. Plus 10 team channels in daily use.
Production line
Composition, optimisation, colour cards, the master sheet, storefront filters — 403 designs registered in one morning this week.
Design studio + scout
Design generation, a weekly pitch of ten evidenced concepts, a nightly sweep of twelve competing stores.
Copy, email, ads
Email copywriter for four brands, email planning for six brands, Klaviyo reporting, the Unit6 ads intelligence tool.
Greenlight in Gorgias
Briefing, verdict and a suggested reply on every ticket — inside the tool the team already uses.
The daily machinery
Creator pipeline for nine brands, a partnerships preview, fulfilment digests for eight brands, meetings with minutes, the newsletter archive, the Facebook group.
Your team today · last 30 days
Eight operators.
One per person.
These are the people who actually worked through Atlas in the last 30 days — and how. HQ gives them one place to meet it.
Support runs inside Gorgias (Princess and the CS team), not through Slack — it is counted under "Support", not here.
October · 30 days
Not a plan.
A result in a month.
HQ.unit6
One address, passkey sign-in, access per person and per team — everyone sees their own.
Tasks — Asana replaced
Orders, Kanban, history, evidence, who-what-when. Atlas opens and closes tasks from chat. Asana goes off.
Ads inside HQ
The ads tool lives inside, no second login.
Learning per person
Each team member gets their lessons on the Scalably Learning Hub and passes them; a prompt gallery for their role; a build tracker: the team asks for a tool through Atlas, requests are tracked, new tools are scoped and priced before building, delivery stays on record.
Atlas in HQ
HQ is one more door into the same Atlas — same operator, same memory, same rules, access scoped to each person.
Usage per person
Who works through HQ and Atlas, and how much — visible to you. Then onboarding with the team, and you announce the rule.
Explicitly not in v1: replacing Slack (researched — a pilot after v1, not a promise), the e-commerce machine (its own phase), the remaining modules (they move in over the next waves).
In 30 days — something that works.
Not a presentation about what could.
October, in order
Four weeks.
Then it's yours.
The first three weeks are quiet in visible results — that is the centralisation. Week four is when everything appears at once.
Under the hood
What HQ.unit6
is made of.
The backbone is already in production at another client. Your systems stay where they are — HQ connects to them, it does not replace them.
After v1
Same team.
Multiplied capacity.
Continuity
Knowledge lives in the system, not in heads. Every task, decision and document stays on record. A new person passes the lessons, gets their operator and gallery, and works from day one. Nobody is a single point of failure.
The organisation, mapped
Every person → every recurring task → what Atlas does → what is still manual. Kept live in HQ, so you see the company, not a guess.
Modules, in waves
Number Artist watchers, Partnerships, Reports, Meetings, Google Drive — each moves into HQ on its own gate, with its own one-page addendum.
Usage per person is not surveillance — it is proof of value, in your own numbers.
Tools nobody else here has
A factory,
not a product.
Private models and GPU
Open models on our own hardware — a dedicated GPU for Unit6 when the volume needs it. On request, scoped and priced separately.
Many agents, one build
Multi-agent coordination is how Katarina's email tool went from a specification to production in a day.
Reported → analysed → fixed
A problem flagged in the admin channel is picked up, diagnosed and resolved, with a written record. The Cloudflare limit was flagged at 17:14 and fixed the next morning.
New in AI → yours in days
Constantly at the edge of research. When a better model or tool appears, it enters the Unit6 system — not a wish list.
Trust and data
Rules before
the first byte.
Your infrastructure. Every external service and data flow agreed in writing. Private models where it matters.
Liability first. For a law firm I wrote, before the first meeting, a contract that holds me liable for any slip with their data. The same is on the table for you.
What we don't build. Fake reviews, fake users, anything your brand wouldn't sign.
People. We give you the metrics. The decisions about your team are yours.
Why this takes a month
We don't build from scratch.
We adapt what already runs.
HQ.unit6 inherits the backbone of a company system already in production at another client: work orders, Kanban, evidence and audit, an embedded agent, documents, access per person. Same architecture, different vocabulary — a field becomes a brand, an order becomes a task. That is why the first result is a matter of weeks.
To be clear
What this is not.
- We don't run your business — the system prepares, you decide.
- We don't touch data without written access rules — and never ask for more access than the process needs.
- We don't sell software off the shelf — everything is adapted to you, or it isn't done.
- We don't replace your people — we give them leverage.
The technology is ours.
The company stays yours.
The road
Steps are gates,
not dates.
v1
HQ.unit6 live, tasks replace Asana, onboarding done, the rule announced.
Adoption
The team works through HQ. The organisation is mapped and stays live.
Second wave
More modules move in; value measured in your numbers.
Store management
The first store runs end-to-end on the machine, under your brand.
Shared upside
When the machine builds profitable stores, we share in them.
Each step is one page, signed when its gate is met. Never a renegotiation of the whole.
The same road, drawn
Every step
has a gate.
Where this leads
One approve.
A whole store.
The pieces already exist — catalogue, creative, email, ads, support, operations. The machine assembles them. In front of it, a scout that watches trends and suppliers and brings you one proposal at a time.
Next step
Scope.
Then contract.
The accompanying offer defines the October scope and its acceptance — what is in, what is out, when it is done. Your confirmation and the first payment start the sprint.
Pavle Lazić · Scalably · your AI partner
pavle@scalably.io