Insight · Clients

Where AI helps — and where human coaching still wins

An honest look at what AI tooling can and can't do inside a coaching relationship, with examples from the platform.

15 May 20265 min read

Key takeaways

  • AI is best at the synthesis work — CVs, market research, practice reps, first drafts.
  • Coaching is best at the relational work — trust, judgement, hard conversations.
  • If the answer changes depending on who is asking, it is a coaching question.

Where AI earns its place — and where a human coach still wins.

The question we get asked most often, by clients and by candidates, is some version of the same one: how much of this is being done by AI now, and is the coaching still real? It is a fair question, and the honest answer is more interesting than either of the easy ones. AI has not replaced coaching, and it is not about to. It has, though, changed which parts of the work a coach should be spending their hour on — and that is worth being explicit about.

The honest frame

Most of the debate about AI in coaching is framed as a contest: either the machines are coming for the coaches, or the coaches will stay untouched because the work is too human. Neither is right. The useful framing is jobs-to-be-done. A career transition contains a stack of different tasks, and they do not all benefit from the same kind of attention. Some are pattern-matching and synthesis problems that a good model handles in seconds. Others are relational judgement calls that no model is close to handling, and probably never will. The work is to draw a clear line between the two and build the service around it.

Where AI clearly helps

Understanding where AI sits in the candidate experience helps clients set expectations and have an informed conversation with employees before they start the programme. Four jobs on the platform are handed to AI without apology, because it does them faster and, in most cases, better than a person would.

1. CV alignment against a specific role

Reading a job description, comparing it to a CV, and flagging the genuine gaps is a pattern-matching task at heart. AI handles it in seconds, at any hour, without getting bored on the fortieth role. What it must not do — and ours does not — is invent experience the candidate does not have. The output is a structured comparison, with every claim mapped back to a verbatim quote in the source. The candidate edits from there.

2. Market and sector intelligence

Synthesising what is happening across a sector, who is hiring, and where the live conversations are is a volume problem. A coach can do it, but should not have to. AI reads more, faster, and surfaces the handful of items the candidate actually needs to think about that week.

3. Low-stakes practice reps

Interview practice gets better with repetition, and repetition is what people most often skip — usually because they are embarrassed to do it badly in front of another person. An on-demand virtual coach removes the embarrassment cost. The first three rounds happen with AI; the fourth, against a person, lands far better than it otherwise would have.

4. First drafts of outreach and follow-ups

Drafting a networking message or a thank-you note is a small task that consumes a disproportionate amount of mental energy. AI is good at the first version. It must not be the final version — anyone experienced can spot a sent-as-is AI message at a glance — but it lowers the activation cost of writing one at all.

Where human coaching still wins

Four jobs we will not hand to AI, and do not expect to.

1. Holding someone through the first fortnight

The first two weeks after a redundancy are mostly about being heard by another person who is not paid to be optimistic. A chatbot, no matter how well-tuned, is the wrong instrument for that conversation. People know the difference, and it matters.

2. Reading the room in a real conversation

Coaches earn their fee in what they notice the candidate is not saying — the hesitation before naming a sector, the slightly over-rehearsed answer about why the last role ended, the topic that keeps being deflected. Models do not pick this up, because the signal is not in the words.

3. Challenging a quietly limiting self-narrative

Most stalled searches are stalled because the candidate is telling themselves a story that closes off options — "I am too senior for that", "I have never worked in that sector", "I am not technical enough". A coach who has earned the right to push back will, gently, do so. AI will agree with the framing and help the candidate execute against it more efficiently. That is exactly the wrong help.

4. Judgement calls on offers and negotiations

Whether to accept, counter, wait, or walk away from an offer is not a question with a generic answer. It depends on the candidate's finances, their family, the state of the market in their sector, and how they will feel about the role in six months. A coach holds all of that context at once. A model, by design, does not.

A useful test

AI question, or coaching question?

If the answer changes depending on who is in the room, it is a coaching question. If the answer is the same no matter who is asking, it is probably an AI question.

"What is a strong opening line for this cover letter?" — same answer for everyone, AI handles it. "Should I take this offer?" — different answer for every person, and it should be.

How we've drawn the line in Launchpad

In practice the split looks like this:

  • Where a programme includes a human coach, the coach owns the relationship, the difficult conversations, and every offer-stage decision. They run the weekly sessions and maintain session notes inside the platform so context is never lost across the programme.
  • Cara, our virtual coach, is on hand around the clock — for practice reps, wellbeing support, quick questions, and the late-evening "should I send this?" moments. She is explicit when something is better taken to a human coach.
  • AI tooling sits underneath both, doing the synthesis work — CV alignment, market intelligence, drafting — so neither the candidate nor the coach spends a human hour on tasks a machine can do well.

The simplest summary: where a human coach is part of the programme, AI is there to amplify that relationship and make the hour together more valuable. For individuals using Launchpad standalone, the same tools — including Cara — provide structured, self-directed support. The two are complementary, and that is the only version of this we are interested in building.

If you would like to see how this works in practice, we are happy to walk through the platform with you and your team. Get in touch.

Was this useful?
Share

How ImaginativeHR can help

Human coaching, sensibly augmented

ImaginativeHR's position is the one set out here: AI carries the repetitive work, coaches carry the judgement. Launchpad's tools handle CV drafting, job matching and practice, while around 100 ImaginativeHR career coaches do the work that only a human can.

  1. 01Cara and the Launchpad toolset for drafting, practice and search
  2. 02A named human coach for strategy, confidence and difficult decisions
  3. 03Clear boundaries on where AI is used, and where it is not

Keep reading

More thinking like this.

All insights

Trusted by leading organisations

ACAL BFI logo
Accor logo
ActionAid logo
Addison Lee logo
Assurant logo
Baxalta logo
BGF logo
BioWare Corp logo
Team member portraitTeam member portraitTeam member portrait

Speak to our team

We're here to help you create impactful outplacement programmes that put people first.