A founder is reviewing the next hire. The team says it needs more capacity. At the same time, it is adopting AI tools intended to make the existing team more productive.
Both investments might be justified. But they need to belong to the same conversation.
What work is holding the business back? How much of it can the current team absorb? Where does human judgment remain essential? And what would genuinely improve if the company hired another person?
Without those answers, it is possible to increase headcount and software spending while leaving the real constraint untouched.
Start with the result you need
“Save the team ten hours a week” sounds like a useful objective. On its own, it is incomplete.
Ten hours spent on what? By whom? And what will happen when those hours become available?
Consider a product team using AI to prepare customer research summaries. Producing the summaries faster could help the team make better decisions sooner. But if nobody has time to review the findings, or the next decision is weeks away, faster preparation may have limited immediate value.
The same applies to recruitment. Generating a shortlist quickly is useful only if the people on it are relevant, the evidence is credible and someone can make a decision.
Choose a result close enough to the work that you can assess it. That might be a shorter time to resolve a customer issue, fewer corrections before a release, or a stronger shortlist reaching the hiring manager sooner.
Then check whether the improvement survives the entire process.
Include the work that follows the output
The subscription or model bill is one part of the cost.
Someone may also need to prepare the inputs, connect systems, check the output, resolve exceptions and correct mistakes. In some workflows, those steps will be light. In others, they could absorb much of the apparent saving.
For sensitive work, consider the review and documentation needed to use the output responsibly. A draft produced in seconds can still require substantial professional judgment.
A useful comparison looks at the total cost of an acceptable completed result. It includes the people and tools involved, alongside the checking and rework.
Compare that with the previous process at a similar quality standard. Otherwise, it is easy to celebrate a faster first draft while overlooking a slower route to completion.
Give the recovered time a purpose
There is a difference between reducing spending and releasing capacity.
If salaries and working hours remain unchanged, completing a task faster does not automatically reduce payroll. It may allow the same team to serve more customers, improve quality, address neglected work or operate at a more sustainable pace.
Those are legitimate benefits. They should be named.
A founder might decide that recovered engineering time should go towards reliability. A commercial leader might use it for deeper account preparation. A hiring manager might spend more time assessing candidates instead of coordinating interviews.
Reducing an overloaded team’s working hours can also be a sensible outcome. Every recovered hour does not need to become another task.
The important point is to make the decision deliberately. Otherwise, a claim of “time saved” can sit in a presentation without changing how the business operates.
Rewrite the hiring brief around the remaining need
This is where AI becomes a hiring question.
Before approving another role, revisit the work itself. Which activities can now be handled differently? Which still require expertise? Where is the team short of ownership, judgment or delivery capacity?
A team may need someone who can resolve complex customer problems rather than produce more responses. It may need an engineer who can assess trade-offs rather than generate more code. It may need a talent leader who can define the brief and challenge assumptions rather than increase sourcing volume.
AI may change the shape of the role. It does not establish, by itself, whether the hire is necessary.
If the constraint is unclear decision-making, adding another person may create more coordination. If the constraint is genuinely specialist capacity, delaying a hire because a tool looks productive could be equally unhelpful.
Define the need before choosing the solution.
Assess judgment, including its limits
“Comfortable with AI” is a weak hiring requirement unless you explain what that means in the job.
Ask candidates to describe a workflow they improved. What changed? How did they check quality? Where did the tool fail? What did they keep under human review?
A useful answer should make the boundaries clear.
Someone who can recognise a convincing but incorrect output may be more valuable than someone who produces ten versions quickly. Someone who knows when a simpler process is sufficient may save more than someone who applies AI to every task.
The hiring decision should reflect the work you need done and the responsibility the person will carry.
Make the investment testable
Before increasing headcount or AI spending, write down the outcome, the current baseline, the quality standard and who owns the result. Decide when you will review it.
That gives you something concrete to challenge.
You may find that the same team can deliver more. You may find that the tools make a specialist hire more valuable. You may find that the process needs fixing first.
At Open Sourcing, that is where a useful hiring conversation starts: what needs to change, what is preventing it, and who is best placed to take responsibility.
Faster work is useful. A stronger business needs a clear account of what that speed delivers.