Search LinkedIn for AI transformation roles in the UK and you will find, at the time of writing, over ninety-nine open postings: AI Transformation or Enablement Manager, AI Business Transformation Lead, AI Adoption Lead. £70,000 to £110,000 plus bonus. The specs read almost identically: accelerate adoption, embed capability, make AI deliver measurable business value across the group.
If your company has drafted one of these, the hard decision is already made. Budget exists. Leadership agrees the work matters. The only question left is the one this piece is about: is a hire the right way to get the capability, or is it a partner? The answer is less obvious than the job spec suggests, and it is not always the partner.
What the hire really costs
The salary is the visible part. The real arithmetic looks like this:
- Salary: £70,000–£110,000, plus bonus.
- Recruitment: a 25–30% agency fee on a six-figure specialist role, and three to six months to fill it, this is a new discipline, and the candidates who have genuinely done it are rare.
- Ramp: six to twelve months before a new hire has learned your business, won internal trust, and shipped changes that matter.
- Timing: the role is scoped against today’s AI. Nobody knows what the landscape looks like by the end of a year of fill-and-ramp, which is exactly why a year is the wrong unit of commitment and a few weeks is the right one.
- Concentration risk: one person asked to cover strategy, change and delivery at once, three different jobs in one description. That is a scoping problem, not a talent problem, and when that person moves on, the knowledge moves with them.
Call it £140,000 and the better part of a year before the work is visibly different. None of that is an argument against the person. It is an argument about sequencing.
Why the order matters
Two things happen in the months before a company has anything to show, and both work against whoever is asked to lead the change.
First, resistance to change is at its highest before anyone has seen anything work. A strategy, however good, asks people to take the benefit on trust. One visible quick win on a single workflow, a proposal that now drafts itself from the call notes, a pipeline that fills without the weekly push, lowers that resistance faster than any strategy document, because people can see it and the person next to them is already using it.
Second, the incoming AI Transformation or Enablement Manager then inherits a cleared path and a result to point at, rather than a sceptical room and a blank page. Their first quarter goes on extending something that works instead of arguing that something could.
That is the whole case for install first, hire second: prove something, then bring in the person to lead what comes next.
When hiring is the right answer
Hiring works when there is an operation for the hire to run. A company that already has AI-first workflows live in two or three functions, and needs someone to own the roadmap, govern the systems, and keep the capability growing, should absolutely hire. At that point the role has infrastructure under it, and the ramp is short because the systems already exist.
Hiring first, before any of that exists, is the expensive order of operations. It asks one person to build the operation and lead it at the same time, with no team yet and every function’s priorities competing for the same weeks. That is a hard brief for anyone, and an expensive way to run a pilot.
When a partner is the right answer
A partner makes sense when the capability does not exist yet and the company wants it installed rather than described. The comparison, honestly drawn:
- Time to first shipped change: weeks, not the fill-plus-ramp year. The diagnosis is days; systems go live in short cycles.
- The work is done by operators: people who have run the same redesign across many companies, building with your team rather than presenting to it.
- What you are left with: working systems, run by your people. In our engagements the systems are built so they can only be operated by a trained team, knowledge transfer is a property of the architecture, not a clause in the contract.
- Cost: scoped per function, a fraction of the loaded cost of the hire, and it stops when the work is done, no permanent salary line for a build phase.
The failure mode on this side is the big-consultancy engagement: a strategy phase, a deck, and a delivery team that rotates. If the partner does not install and hand over running systems, you have bought the binder, not the capability.
The sequence that actually works
The strongest version is not either-or. It is: install first, hire second.
Bring in a partner to enable one function, sales, go-to-market, brand and content, or operations, and prove the redesign where one accountable leader can say yes. One of our clients cut the admin load on their sales team by two-thirds this way. Then, when there is a live AI-first operation to govern, make the hire into a role with systems under it, evidence behind it, and a much better story for the candidates you actually want.
The job spec you drafted does not disappear. It gets better, because the role now has something to lead.
Does bringing in a partner mean we should cancel the open role?
Not necessarily. If the role is scoped to govern and grow an existing AI-first operation, keep it, and fill it once the first function is live. If the role is scoped to create the operation from nothing, the arithmetic above applies: install first, then hire into strength.
What about a fractional Chief AI Officer instead?
A fractional CAIO can add real judgment at board level, but the constraint in most companies is not judgment, it is installed capability. Titles govern operations; they do not create them. Build the operation first, then decide what kind of leadership it needs, fractional, full-time, or none yet.
How do we judge a partner before committing?
Ask three questions. What does your diagnosis look like, and what does it cost? Show us a workflow you redesigned, before and after. And: what does your team leave behind when you go? A good partner answers all three with specifics. A deck-seller stalls on the third.
The Hire-or-Partner Scorecard: the loaded-cost arithmetic, ten questions, three verdicts. Five honest minutes, free, instant download.
And if you want to test the thinking before talking to anyone: the AI-first Work Redesign Audit is a self-serve exercise, run it on one of your own workflows and see what the redesign step actually asks.
If you are weighing this decision right now, the fastest way to test the partner route is a discovery call, we will scope the diagnosis against the function your job spec was written for. And if the honest answer is that you should hire, we will tell you that too.