AI implementation consultant - we don't stop at the roadmap
Most AI consulting ends with a slide deck. Twelve use cases, a maturity model, a recommendation to "pilot." Then the consultant leaves, and the firm discovers that implementation - the integrations, the data, the edge cases, and above all getting the team to actually use the thing - is where it all falls apart.
We're implementation people. We do the audit, yes - but then we build the first system in weeks, connect it to your apps, help your team adopt it, and run it on a monthly plan. The deliverable isn't a roadmap. It's a tool your people use on Tuesday.
We built our first tools for the Fame team - a 6-year-old people business with 100+ clients. They now give those hours back every week.
Read the Fame story →Why implementation is where AI fails
Three reasons, and we've designed around all of them.
Integration
The demo worked on a clean spreadsheet. Your data lives in HubSpot, Xero, a practice tool and three Google Drives, half of it slightly wrong. Implementation means connecting to the apps you really use and handling the mess - which is engineering, not prompting.
Judgement
AI is confidently wrong on a regular basis. A system with no human in the loop ships a hallucinated figure to a client eventually. We build every tool so that AI drafts and a person decides - the review step is designed in, not bolted on.
Adoption
This is the one nobody budgets for. A tool the team doesn't trust or understand gets quietly abandoned in week three - usually because someone's worried it's there to replace them. We build to make your people more powerful, we say that out loud, and we support the rollout so the team actually adopts it. (We wrote the playbook: how to get your team to use AI.)
How implementation works with us
- 01
Scope
2 weeksAn ops audit: we interview your team and map your apps to find where the hours go and where AI genuinely pays - usually the repetitive copy-paste work between apps, not the flashy stuff. Then we choose what to build: a roadmap with the first three builds scoped, priced and ranked by hours recovered.
- 02
Build
2-6 weeksWe build and test the first system, connected to your real data and apps, with human review wherever judgement matters. You approve it before it counts.
- 03
Roll out
We train your team and manage the change so people actually use it: a live walkthrough with your team on real work, write the one-page usage policy with you, and check in at two weeks - the point where most rollouts quietly die.
- 04
Improve
MonthlyWe watch usage, handle model and app changes, fix what breaks, and keep improving it. Usage tracking shows you who's using it.
What we implement
For people businesses - agencies, accountants, lawyers, consultants, recruiters:
Document and transcript processing
Call recordings, intake forms and PDFs turned into structured records in your systems
Reports that draft themselves
Client reporting drafted from your data, reviewed and sent by a person
Triage and routing
Requests, emails and tickets classified and sent to the right person
Knowledge assistants on your own data
Answers from your SOPs, past work and documents, with sources
Agents that keep your apps in sync
CRM, accounting and practice tools kept consistent without rekeying
Client-facing updates
Portals and status pages that answer "any update?" automatically
Why implementation needs an engineer who stays
Models get deprecated. APIs change. Prompts drift. A system that worked in March quietly degrades by September unless someone's watching. That's why our implementation ends in a run plan, not a handover document: we're on the hook for keeping it working, and you're never dependent on a consultant who's moved on.
What it costs
| Ops audit | $2,500 / £2,000 · two weeks · credited to your first month |
| Run · starter | $2,000 / £1,600 a month · one team, one painful workflow: first tool built and live, then one build at a time, small changes in 5 working days, monthly hours-back report |
| Run · growth | $3,500 / £2,800 a month · 10-30 people: two builds running at once, small changes in 2 working days, weekly check-in with a named lead, live hours-saved dashboard |
| Run · embedded | $7,500 / £6,000 a month · 30+ people: continuous backlog in two-week sprints, same or next-day turnaround on small changes, we sit in your Slack and task tool, weekly planning and a quarterly roadmap |
Your first build, training and change management are included in every monthly plan · 3-month minimum on run plans · AI and API usage billed at cost · it runs on your own accounts, so you own it - you pay those tool costs directly
People first, by design
We're the AI implementation partner that isn't trying to shrink your team. We build so AI does the grunt work and your people do the work only they can do - and because we say so, your team trusts the tools enough to use them. That's not just a nicer story; it's why the implementation sticks.
See it built for a real team
Asset review and approval tool
Clients watch, comment on and approve every episode and asset in one place, with automatic quality checks first.
See the build →02Audio editing plugins
Plugins for Pro Tools, Reaper, Cubase and Audition that bring client comments, cleanup suggestions and uploads into the editor's own software.
See the build →Questions, answered plainly
What does an AI implementation consultant do?
Takes you from "we should use AI" to working systems: finds where it pays, builds it, connects it to your real apps and data, helps your team adopt it, and keeps it running. We do all of that; most consultants stop at the roadmap.
How much does AI implementation cost for a small business?
Our two-week audit is $2,500 (£2,000), credited to your first month; run plans are $2,000-7,500 a month, with your first build included. All prices are published.
How long does implementation take?
Two weeks to audit, then the first system in 2-6 weeks. We'd rather ship one working tool in a month than twelve use cases in a deck.
How do you make sure our team actually uses it?
We design for it: human review built in, a live walkthrough on real work, a one-page usage policy, a two-week check-in, and a people-first promise that lowers the "is this replacing me?" fear that kills most rollouts.
What about our data and security?
Your data stays in your systems - what we build runs on your own accounts, not ours. We write a plain-language AI usage policy with you covering what never goes into AI tools, and we use settings and providers that don't train on your data.