A squad of AI assistants, with a person on the hook for the work
Create an account, describe your situation, and the assistants matched to it start asking the right questions the same day. Where the work needs judgement, accountability or a signature, a person takes it — and we tell you which is which before you buy.
- 5
- ways in, so you never browse a catalogue
- 100
- assistants documented and numbered
- 3
- service tiers, from self-serve to human-led
- 40
- of them in the first delivery wave
Start from your situation, not from our catalogue
100 assistants are documented behind these five doors. You never have to browse them — pick the sentence that sounds like your week, and the right ones come to you.
“I want to launch something of my own”
Solo startup
You have an idea and no team. The work is scoping it down to something one person can ship, and finding the first ten clients without an advertising budget.
- HAssistedSoloLaunchA plan to launch a one-person digital business in 14 days
- HAssistedMVPScoperMinimum scope for a first release, and the list of what you will not build
- CHuman-ledFirstTenClientsWinning your first ten clients with no advertising budget
- CHuman-ledSoloPricingPricing on value delivered, not on an hourly rate
“I run a business and I'm drowning in work”
Small business transition
An established business with more work than hours. We start with a readiness score and an audit of your actual week, and we finish by showing the return in hours and pounds.
- HAssistedAIReadyScanThe Bionic Readiness Score — the entry diagnosis for a business
- HAssistedTaskAuditAn inventory of the working week: what to automate, what stays human
- HAssistedQuickWin90A 90-day plan to the first measurable hours saved
- CHuman-ledROIProofProof of return in hours and pounds, with the measurement method stated
“I want to learn this myself”
Academy
You would rather own the capability than rent it. Prompt engineering on your own business's tasks, a twelve-week programme, and an honest account of where AI breaks.
- HAssistedPromptSchoolPrompt engineering from scratch, on tasks from the learner's own business
- CHuman-ledBionicPath12A 12-week programme: from a traditional business to a bionic one
- HAssistedToolScoutTools matched to skill level and budget, not to fashion
- ASelf-serveAILiteracyWhat AI is and is not: hallucinations, limits, and the cost of an error
“I'm building for people, not for profit”
Social enterprise and CIC
Community interest companies, grant funding and impact reporting. The asset lock changes how you can raise money, and CIC34 comes round every year whether you are ready or not.
- CHuman-ledCICFormerRegistering a CIC: CIC36, the community interest statement, Companies House
- HAssistedImpactLedgerA continuous register of impact — the CIC34 report writes itself
- CHuman-ledTheoryOfChangeInput to activity to outcome to impact, with a measure at every level
- HAssistedGrantMatchUKMatching to funds: National Lottery, foundations and local grants
“I have a project and I need a team”
Bionic studio
A person leads, with a squad of agents behind them. Every assistant in this pathway is human-led by design — this is the door for work that cannot be handed to a model unsupervised.
- CHuman-ledSquadBuilderAssembling the squad of agents for one specific client project
- CHuman-ledEvalHarnessThe quality tests an assistant must pass before it reaches a client
- CHuman-ledEscalationDeskThe thresholds at which an assistant stops work and calls for a person
- CHuman-ledBusinessTwinA digital twin of the client's business, for testing decisions before rollout
Not sure which door is yours? Start with a readiness scan and let the answer come out of your own numbers.
You buy agent hours, not advisory sessions
Every assistant carries a letter telling you how much of it is machine and how much is person. Nothing is sold as autonomous when it isn’t.
You run it
The assistant is yours to drive. Bounded, well-defined work where a wrong answer is cheap and obvious.
Fixed subscription
We run it, a person checks it
The squad does the work and someone reviews it before it reaches you or your customers. This is where most of the catalogue sits.
Agent hours, plus review hours
A person leads, the squad assists
Law, governance, funding, safeguarding, anything with your name on it. A human owns the outcome and the agents do the legwork.
Priced per project
Human-led work is the larger half of the catalogue. That is deliberate — the parts of a business that carry the most risk are the parts we will not hand to a model unsupervised.
Eight mechanisms, not eight hundred prompts
Anyone can hand you a list of chatbots. These are the working parts that make an assistant catalogue into a service you can hold to account.
Agent hours, not consultant hours
You buy a package of squad work — say 40 agent hours plus 4 human hours — rather than advisory sessions.
It breaks the revenue ceiling of a one-person practice without pretending the human isn't there.
Bionic Readiness Score
A repeatable index of readiness across data, processes, people and compliance, measured when you start and again at 90 days.
It turns “transformation” into a number you can put in front of a board, a funder, or a CIC report.
Business digital twin
A decision model of your business, used to test a move before you make it for real.
Consulting stops being an opinion and becomes a simulation with assumptions you can check.
Synthetic customer panel
A set of personas that pre-test a message before you spend budget putting it in front of real people.
A cheap sieve that catches the obviously wrong version early.
Limitation. This is not research data. Its output must never be presented, cited or reported as research findings — it narrows what is worth testing properly, it does not replace testing.
White-label assistant
Your own assistant, under your own brand, hosted and maintained by us.
Recurring revenue for us and a working product for you, without tying you into a long contract.
Impact Ledger
A running register of activities and their effects, from which the annual CIC34 report is generated.
Impact reporting is the single biggest administrative burden a CIC carries, and almost nothing serves it well.
Content provenance
Explicit labelling of the AI share in what we hand over, using the C2PA standard, with a disclosure policy to match.
It gets ahead of regulation and builds trust in a market where most competitors quietly imply they wrote it by hand.
Escalation Desk
Defined thresholds at which an assistant stops, refuses to continue, and tells you to bring in a person.
It is a safety condition first. It is also, honestly, where the human side of the service begins.
Measured at the start, measured again at 90 days
One dashboard, from the first scan to the follow-up measurement.
Readiness scan
AIReadyScan measures where you actually are across data, processes, people and compliance. You get a number, and the working week behind it.
Pick a door
Your score and your situation point at one of the five pathways. We tell you which assistants apply, and which parts need a person.
The squad works, a person reviews
Agent hours get spent on the real backlog. Assisted work is checked before it reaches you; human-led work stays with a human throughout.
Re-measure at 90 days
The same index, run again, next to the hours and pounds saved. If the number has not moved, that is the conversation we will have.
And who it isn’t for, yet
We would rather lose the enquiry than sell you an implementation that cannot pay for itself.
A good fit
- You have real, repeating work that eats your week — quotes, scheduling, reporting, follow-ups.
- You have customers and cash coming in, and the constraint is your own hours.
- Someone on your side can spare an hour a week to review what the assistants produce.
- You want to own the capability afterwards, not rent a black box forever.
Not a fit yet
- Your real bottleneck is finding customers. AI will not fix a sales problem, and a mis-sold implementation costs you money and costs us our reputation.
- Cash is tight enough that this month's invoice is the question. Come back when it isn't.
- Nobody has time to learn anything new right now — assisted work still needs a human to review it.
- You need a guaranteed outcome by a fixed date. We can scope that conversation, but not promise it here.
If you are in the first column, the cheapest way to find out what is actually worth automating is a readiness scan of your own week. It starts in the dashboard.
Tell us what your week actually looks like
Create a free account, describe your situation, and the assistants matched to it start asking the right questions straight away.