Start with a working session
It's one focused day. We look at how your work runs, then pick one real workflow together. If your team is new to AI, we start with the basics on your own work: briefing AI well, writing down the context it needs and organising work so it builds up. We tell you upfront who should be in the room.
What you walk away with:
- A working prototype on one of your workflows
- Where you are on the path: AI-absent, AI-enabled or AI-native
- A build, buy or partner call for each part of that workflow
- Who is accountable for each decision
- Our advice on where to go next
The session is credited in full if you go on to work with us.
Then two steps on one path
The working session tells you which step you start at.
| AI-absent → AI-enabled | AI-enabled → AI-native | |
|---|---|---|
| You're here if… | You have the tools, maybe even daily users, but the work hasn't changed | Your first critical workflow runs on agents, but most work still waits to be asked |
| What we do | Put a governed AI workspace in place, and build your company's shared brain: your prompts, skills and knowledge in one place, checked against your policies and used every day | Rebuild your critical workflows so the brain drives the work, with agents as coworkers that escalate to your people based on your policies |
| Ends with | Your first critical workflow started and run by agents | The workflows that run your business, started and run by agents |
More detail: AI-absent → AI-enabled and AI-enabled → AI-native.
What a first workflow looks like
It has to recur and matter to the business, and today someone has to start it by hand. One-off prompts and demos don't count.
| Function | Agents start and run | A person decides at |
|---|---|---|
| Sales | A new inbound lead comes in: agents research and enrich it, draft a reply and book the meeting | Outreach to strategic accounts |
| Support | A ticket comes in: agents triage it, resolve it from the brain if it's tier-1, or draft the reply | Refunds or credits above a set amount |
| Finance | An invoice arrives: agents match it to its purchase order and flag any exception | Payments and write-offs |
| People ops | A new hire is added to the HR system: agents set up accounts, training and a schedule | Access to sensitive systems |
| Reporting | Every Monday, agents pull the numbers, draft the report and post it | Anything going outside the company |
Who does what
Your team holds authority, agents do bounded work, specialists bring judgment. Day to day, that looks like this:
| Who | What they do |
|---|---|
| Your team | Decides which workflows matter, who is accountable for each decision and what agents may approve. Brings the expertise the work depends on. |
| Agents | Start and run the workflows they're permissioned for, and escalate to your people based on your policies. |
| Our forward-deployed team | Sits with your people, rebuilds the workflow inside the systems you already run, and teaches your team to run it. |
How the engagement runs
You always know what the current step ends with. The plan grows as we learn how your work runs, and what the agreed workflow needs doesn't become a change order. You get something usable from the first session, and once we've seen your workflow we tell you how long it will take to reach production.
On price, talk to us. We scope it with you in the working session.
After the engagement
We build it so you can own it, and we operate it for as long as that's useful. Your team learns to manage what we installed, and can swap or change agents at any time. Bring your own agents or use ours. We can run the system and your brain as a managed service, or you can self-host.
Underneath is Station, the system we deliver on, configured for your use case. It has spend caps, a kill switch, full action logging, and a clear line between what the AI controls and what it only observes. Station runs Splice, too.
Questions about the process
What if we find something new mid-engagement?
If the agreed workflow needs it to work, it goes into the plan. We don't turn every discovery into a change order. Genuinely new work gets scoped with you, in the open.
Who's accountable when the AI gets it wrong?
A person on your side. There's an accountable person behind every decision that matters, and agents can handle some approvals that person has permissioned. Before anything runs, we agree who is accountable and what each agent may approve.
How fast will we see something?
You get something usable from the first session. Once we've seen your workflow there, we tell you how long it will take to reach production.
What does it cost?
Talk to us. We scope the work with you during the working session, which is credited in full if you go on to work with us.