AI transformation partner
We find where AI pays. Then we get your team using it.
For companies of 50 to 300 people. We get your team using the AI tools you already own, build only what those tools can’t do, and run it with you for the year. Diagnostic first.
CREDITED IN FULL · REFUND IF THE READOUT MAPS UNDER $10K · WE DON’T RESELL LICENCES
01 · Diagnose
$5,000
Two to three weeks. Credited in full to Adopt.
02 · Adopt
Priced in your readout
Your team set up and trained on the tools you already pay for.
03 · Partner
Twelve months
Adoption checked monthly, the return reported against your baseline.
Two founding slots at $2,500
The published diagnostic · from the client’s own records
Our approach
Helix takes a company from AI experiments to an operation that runs on it.
Workflow first, AI second. We map how the work moves before anything is bought or built, get your team using the AI tools you already pay for, and build only what those tools can’t do. Anything that sends, publishes or spends waits for a named person to approve it, and we report against your numbers, not our slides.
AI your team actually uses.
Step 01 · Diagnose
The Diagnostic
A kickoff with leadership, 5 to 8 recorded interviews with the people who do the work, and a live readout in two to three weeks: every opportunity ranked and priced.
Learn moreStep 02 · Adopt
Your team, on the tools you already own
Copilot, Gemini, ChatGPT or Claude set up around each team’s real work, trained, with an owner on your side. Built only where those tools can’t do it.
Learn moreStep 03 · Partner
Twelve months, measured
A monthly adoption check, a return report against the baseline set in the diagnostic, and a re-diagnosis every quarter.
Learn moreThe problem
You bought the tools. The work still runs the old way.
Most companies have tried the obvious fixes. None of them changed how the work gets done.
The pilots never reached production
A demo worked well in a meeting, then sat unused. Nobody carried it past the proof of concept into daily work.
Seats bought, workflows unchanged
The licences are paid for. Nobody set them up around the team’s real work, so the work still happens the same way it always did.
Nobody mapped the work first
The automation went in where it was easy, not where it paid. The expensive steps are still done by hand at 9pm.
Nobody owns adoption
No one is measured on whether the tool gets used, so it quietly stops getting used. The setup was never the hard part.
The board wants a plan
Leadership is being asked what the company’s AI plan is, and there is no map to point to. A slide deck is not an answer anyone can act on.
The problem is not your team. It is the work nobody wrote down. That is what we map, and that is where we start.
We are not the 200-slide roadmap, the licence reseller’s workshop, the pilot that never left the demo, or the AI SDR that blasts your list. Workflow first, AI second.
How we work with you
A working system in your tools, not a deck in your inbox.
Set up where the work already happens, on terms written down before it starts.
Your tools. Your approval queue. Your numbers.
- Microsoft 365 Copilot, Gemini for Google Workspace, ChatGPT or Claude, whichever you already pay for
- We don’t resell licences, so the recommendation is whatever fits the work
- Anything that sends, publishes or spends waits for a named person to approve it
- The baseline is set in the diagnostic, so the return is measured, not claimed
Credited in full. Scope fixed in writing.
- The $5,000 diagnostic credits in full to Adopt
- Scope and price are fixed in writing before anything starts
- Full refund if the readout maps under $10K of recoverable value
- You own everything we set up or build, with a documented handover
The engagement
Three steps, and what lands in each.
In that order, with the terms shown at every step.
Step 01 · Diagnose · two to three weeks
We find where AI pays, and where it doesn’t.
A kickoff with leadership, then 5 to 8 recorded interviews with the people who do the work across two or three departments. The readout is presented live: what to do, in what order, and what each step is worth.
- Kickoff with the CEO or COO, with the interview list agreed on the call
- 5 to 8 recorded 60-minute interviews, plus a short staff survey
- The map of how work actually moves, against what the process document says
- Every opportunity ranked on value, feasibility, return and risk
- What the tools you already pay for can do today, and what genuinely needs building
- Where AI does not belong, named
- Where the freed hours go, agreed with you, so the return can be measured
- Two or three working demos on your workflows, each with a price
- You keep the roadmap either way
Step 02 · Adopt · from two weeks
Your team, using the tools you already own.
Most of what the readout recommends is not a build. It is setting up the AI tools you already pay for around each team’s real work, and training people until they use them. We build only what those tools can’t do.
What we set up
- Tool setup. Copilot, Gemini, ChatGPT or Claude configured for each team’s workflows and data.
- Training. Per team, on their own work, not a generic course.
- AI use policy. Who uses what, on which data, with which approvals.
- Licence review. Who needs a seat, and what you pay for that nobody uses.
- Builds, where needed. Only what the tools you own can’t do within about six months.
How it runs
- Every system ships with a trained owner on your team
- Anything that sends, publishes or spends waits for a named person
- Proof of concept first, then production
- Priced in the readout, scope fixed in writing
What you own
- Everything we set up or build
- The documented handover
Step 03 · Partner · twelve months
We stay for the year, and prove the return.
Adoption is where AI projects die quietly. For twelve months we check who is using what, tune it, add the next workflows, and report the return against the baseline set in the diagnostic.
The setup is the small part. Getting the team to use it is the rest, and that is what Partner is for.
- A written plan for each quarter
- A monthly adoption check: who uses what, read from the logs
- A return report against your baseline
- Tuning, and new workflows as the company grows
- Re-diagnosis every quarter
- Documented handover: you own everything we build
Microsoft 365 Copilot · Gemini for Google Workspace · ChatGPT · Claude · your CRM, inbox and data
The published case
They were paying for leads nobody followed up.
Enquiries
259
Called back
167
Visits booked
20
Visits attended
0
Enrolled
2
leads a music school had paid for, contacted once, then dropped
- We read four spreadsheets against each other and found a show-up rate of zero that nobody had calculated.
- We built a second brain, one dashboard, and the follow-up and win-back modules for the leads they had already paid for.
This was a small business. Our first two diagnostics at 50 to 300 people run at a founding rate, in exchange for a published case like this one, with real numbers. Our own company runs on the same approach, shown live on the call, not on a slide.
One ladder
The price is on the page.
Not a licence reseller’s free workshop. Not a $500 look-over. A priced roadmap with a number behind every step.
Diagnose · two to three weeks
Credited in full to Adopt. You keep the roadmap either way.
From two weeks
Scoped and priced in your readout, on your teams and your numbers.
Twelve months
Priced in your readout and agreed before it starts. A written plan each quarter.
Two founding slots at $2,500, in exchange for a named case study with real numbers. If the readout does not map at least $10K of recoverable value, you get a full refund. Scope and price are fixed in writing before anything starts. The full breakdown is on the pricing page.
Fit
Who this is for.
For you
- Companies of 50 to 300 people, in any industry
- You are paying for AI seats and the work still runs the old way
- Leadership or the board wants an AI plan it can act on
- You want the work mapped and priced before anything is bought or built
Not for you
- You want AI inside your product. We work on the operation around it.
- You want a slide deck. We set up and run the system with your team.
- You want AI rolled out without talking to the people who do the work.
FAQ
Common questions.
What does the diagnostic involve?
Do we have to buy new tools?
Why not the free workshop from our Microsoft partner?
What if an agent gets something wrong?
Who owns what afterwards?
What if we already tried AI and nothing stuck?
Next step
Stop running AI as an experiment.
Twenty minutes. Tell us where AI is stuck in your company, and we will tell you whether the diagnostic is worth running.
$5,000 diagnostic, credited in full · you keep the roadmap either way · you own everything we build