Chief change officers at 200-plus-staff businesses are signing off on AI headcount before proving what the function should do. One question changes that.
The meeting is already in the calendar. Someone has put together a slide deck. There’s a proposed org chart, a list of tools to licence, and a number at the bottom of the page that needs a signature. The ask is to approve the budget for the AI function. The problem is that nobody in the room has proven yet what that function should actually do.
This is the position a lot of CIOs and change leaders find themselves in right now. The pressure to build internal AI capability is real, and the instinct to hire for it is understandable. But approving a department before the work is defined is a bet, and it’s a large one.
The question the deck doesn’t answer
Before the budget gets signed, one question is worth putting on the table, what specific business outcome have we proven AI can deliver, and what evidence do we have that scaling this will create value?
The source of that question matters. It comes from the same thinking that shapes how AI actually moves from experimentation to practical use in a business. The shift happens when teams turn existing knowledge into repeatable, reviewable workflows, not when they commit to headcount and infrastructure based on a plan. A budget tied to proven results is a different conversation from a budget tied to intent.
Picture a large multi-site accounting firm. The Head of Technology has been asked to stand up an AI function. The proposal on the table is a 12-month search for a Head of AI, two senior engineers beneath them, and a tooling budget to match. The total cost, fully loaded with recruitment fees, salaries, and infrastructure, is significant. The timeline to first output is 18 months at best.
Nobody has yet run a single AI workflow through a real business process to see what it returns.
What the hire actually costs you
The recruit lag alone is a serious consideration. Finding senior AI leadership in Australia takes time. Factor in the search, notice periods, and the ramp-up once someone starts, and 9 to 12 months before the function is operational is a reasonable estimate. During that period, the business is either waiting or making decisions without the expertise it hired to get.
There’s a second cost that’s harder to see. A new Head of AI arrives to a blank page. They spend their first months working out how the business runs, which processes are worth automating, and where the data actually lives. That’s valuable work, but it’s work the business is paying senior-hire rates for, and it produces a plan rather than a result. The evidence on this is plain, AI moves to practical value only when teams turn existing knowledge into repeatable, reviewable workflows. The new hire has to acquire that knowledge first. That takes time the business is paying for before a single agent does a single useful thing.
The case for proving value before committing to structure
The alternative is to pick one workflow that moves the bottom line, define it tightly, and prove it works before building a department around it.
That means one defined task, one reliable input, one standard output, and one approval point. Run real examples through it. Compare the output with your accepted and rejected cases. Fix the gaps before adding more access, more steps, or more autonomy. This approach means the budget approval conversation changes from “trust us, this will work” to “here’s what it returned on this workflow, and here’s what we’d get if we scaled it.”
For the accounting firm, that might mean starting with one process common across all sites, say contract review, supplier onboarding, or compliance reporting, and running it through a defined AI workflow with a person approving every output. Four to five weeks of that, with a clear record of what the AI got right and where it needed correction, gives the board something real to approve against.
A Technology Partner handles this work from the couple of weeks. The process mapping comes first, because the AI needs to understand how the business runs before it can do anything useful in it. Then the build. Then the running of the workflow, with data pipelines connecting the business’s existing systems, and a data lake so the AI works from a full picture rather than scattered files. Everything gets documented into a shared workspace as it’s built, so the business owns the record at every stage.
What the business inherits
This is where the rethink-the-hire logic lands. If the accounting firm later decides it wants to hire a Head of AI internally, it doesn’t inherit a blank page. It inherits a running system, a documented record of how the business’s processes were mapped, which workflows were built and why, what the AI returned, and where the humans stayed in the loop. The new hire walks in with a foundation rather than a starting point.
That’s a different hire, too. A senior leader brought in to own and extend a working AI function is a more attractive role than one brought in to figure out where to begin. The candidate pool is different. The onboarding is faster. The time to output is measured in weeks, not months.
The Technology Partner model also means the business isn’t locked to any one AI model or vendor’s architecture. The data, prompts, and workflows sit apart from the underlying model, so as the field moves, the business stays on the current best option rather than being tied to whatever was current when the department was built. That’s the model-agnostic position, and for a multi-site business with a large and growing agent footprint, it matters more over time, not less.
What this means for you
The budget meeting doesn’t have to be a bet. The question to put on the table before the org chart gets approved is a simple one: what have we proven, and what does scaling that proof actually cost compared to building the department first?
For most staff heavy businesses weighing this decision right now, the answer points to starting with a Technology Partner, proving value on one defined workflow, and building the record that makes the internal hire, if and when the business wants it, a much easier decision to make.
If you’re in that approval meeting and you’d like a plan built around your business’s actual processes and goals, take our AI Roadmap Interview.
