AI for executive search firms with a reputation to protect.
A candidate brief that takes your researcher a morning, done in fifteen minutes. A longlisting cycle that blocks a week of consultant time, collapsed to a day. Client-ready output throughout: your formatting, your tone, your name on it.
Cut the legwork. Free up your team to do what they do best.
Working with retained search firms across the UK and internationally.
Invisible AI
Executive search runs on trust, discretion, and a reputation for judgement that takes years to build. Anything that reaches a client or a candidate has to read as the work of the firm, not the output of a tool.
That sets a higher bar for AI than most consultants are working to. AI belongs in the parts of the work where it can be built, tested, and reviewed before the firm puts its name to anything: research, synthesis, drafting, sector intelligence, internal knowledge. It does not belong in the moments where judgement is the product: the candidate conversation, the shortlist call, the read of a board.
Most of what I do is making AI useful in the first half of that distinction without it leaking into the second. I have been working with these systems from the start, most of that time at the senior end of technology businesses, and the lesson I keep coming back to is this: the technology is rarely the problem. The problem is what happens between procurement and actual use, which is where most implementations quietly fail.
Your consultants are almost certainly already using AI. The question is whether it arrives at your firm deliberately, or arrives unevenly, invisibly, and below the standard you'd sign off on.
Why a licence is not enough
Claude, ChatGPT, Copilot in various configurations: access is no longer the constraint. Most firms have bought a licence, and most firms have stopped there. This is why you still see the same pattern: one consultant who has worked out what to do, and a team that hasn't.
Getting consistent output takes work with your own material. That means prompts built for each specific task, tested against examples the firm actually uses, and enough practical training that the team arrives at a shared standard rather than a collection of individual workarounds. Without that infrastructure, the licence gives you an uneven capability you cannot audit.
A capability you have bought is not the same as a capability you have. I build the gap between those two things.
Output you can sign off
Every report that reaches a client has a partner's name on it. So the question is not whether the AI is clever. It is whether you can check what it produced before you sign.
What you get is a set of tested prompts with the checks built around them. I test each one against your firm's real material and the standard you set for the task.
Factual statements carry a source, so any claim can be traced back to where it came from: sometimes a public link, more often an interview note or the CV. A consultant's assessment is marked as an assessment rather than blended into the facts. Where the evidence is missing, the draft says so instead of filling the gap with something plausible. Review gets faster, because the consultant is applying judgement to the flagged items rather than re-verifying every line.
Each workflow ships with test examples: real material from your firm, paired with the output you agreed was right. Rerun them when a prompt changes or the model updates and you can see whether quality has moved before a client does. The handover records who owns those checks and when to run them.
AI can still misread a source or invent a detail. The checks reduce that; they do not remove it, and anyone who tells you otherwise has not run this in production. That is why the review gate sits before the client, every time.
Proven AI workflows
In the order of a typical search.
01Pre-call sector briefing
02Weekly sector digest
03Proposal and pitch drafting
04Target company research
05Approach research and personalisation
06Candidate report generation
07Practice analytics
08Internal knowledge retrieval
09Content and thought leadership
Time-saving figures are indicative, drawn from comparable workflows; we calibrate against your firm specifically during discovery.
Risks worth managing
The risk is rarely dramatic. It is familiar work reaching the wrong standard, or the wrong place.
A client asks where a fact came from
A report reaches a client with a claim nobody can source, and the partner whose name is on it has to explain. Factual claims carry a source, and anything unsupported is held for review before sign-off.
Candidate information enters an unapproved account
A consultant uploads a CV through an account whose terms and data location nobody has assessed. Accounts are approved before use, covering tier, residency and retention. A paid subscription is not approval.
Waiting leaves current use unchecked
Partners defer a decision while consultants carry on using personal accounts unsupervised. Establishing where AI is already in use, and giving the team an approved method, does not require committing to a full build.
I give you practical guidance on tools and configuration. Your firm remains responsible for its data protection decisions, advised by its DPO or legal adviser.
I'm based in Bristol. I started my career as a researcher at Oryx and Novo, spending over three years sourcing candidates, pitching roles, and working with consultants to refine longlists. The work behind the work. The parts of the search process AI is now genuinely useful for.
I then spent fifteen years in senior product and engineering roles at American Express, Zoopla, Deliveroo and Intent HQ (AI Business of the Year at the 2025 National AI Awards), followed by the last four working closely with large language models. Prompt engineering at scale, production tooling, and helping teams move from AI experiments to AI that actually gets used.
Maneform exists because executive search is one of the few places where the structured, judgement-led nature of the work makes AI genuinely useful, and where the stakes around discretion and quality mean it has to be done carefully.
The next step is a conversation
Forty-five minutes, on a video call, no slides. With the people who know the problems your Consultants and Researchers face.
We walk through how your team currently produces candidate reports, sector briefings, proposals and approach research, and identify the two or three workflows where AI saves the most time without compromising client-facing quality.
I'll demo the most relevant workflows I've already built, so you can see how they work and what the output looks like.
Within two working days you'll have a written summary: the workflows we identified, two or three recommended next steps, and indicative pricing for each. If nothing from the conversation warrants a proposal, I'll say so.
What do you want to solve?
Or just email
If you'd rather skip the form, write directly. Tell me about the firm, the problem, and what "good" looks like.
Once an engagement is underway, I'm on your Slack or Teams during UK working hours. Quick questions get quick answers; nothing waits for a scheduled call.