THOUGHT LEADERSHIP · CAPACITY PLANNING
Should a Small Training Provider Hire Another Assessor or Add AI-Assisted Marking First?
Hiring adds real capacity but it’s slow and hard to reverse. A tool is fast to bring in and easy to walk away from. A framework that goes deeper than gut feeling.
By Eduface · July 2026 · 8 min read
This is one of those decisions that feels like it should have an obvious answer and, once you actually sit down to make it, doesn’t. Hiring adds real capacity and real judgement, but it’s slow, expensive, and hard to reverse if volume dips again next term. A tool is faster to bring in and easier to walk away from, but it doesn’t replace the human decision at the centre of marking, and it isn’t free either. Here’s a way to think about it that goes a bit deeper than gut feeling.
Why this is a real economic question, not just a budgeting one
There’s a famous, almost century-old paper in economics that asks a question which sounds strange at first: why do firms exist at all, instead of everyone simply contracting individually for every task they need done. Ronald Coase’s 1937 answer, in “The Nature of the Firm,” was that there are real costs to using the market for every transaction, costs of finding people, negotiating, monitoring quality, and a firm exists precisely where it’s cheaper to bring something in-house than to keep transacting for it externally each time (Coase, 1937). It’s a strange paper to reach for when thinking about hiring an assessor, but the underlying logic transfers directly: the question isn’t simply “can we afford a salary.” It’s “where does the actual cost sit, and which option genuinely reduces it.”
Quick answer
Neither option is inherently right. Hiring brings judgement in-house permanently with fixed costs. AI-assisted marking is closer to buying a capability on the market: pay for what you use, scale in both directions.
Applied to marking capacity, hiring an assessor brings judgement in-house permanently, with all the fixed costs that come with that, recruitment time, training time, a salary that doesn’t flex down if volume drops. AI-assisted marking is closer to buying a specific capability on the market: you pay for what you use, it scales with volume in both directions, and it doesn’t require months of onboarding before it’s useful. Neither option is inherently right. The question is which one actually reduces your real cost, given your specific volume and its predictability.
The questions worth actually working through
Is your volume growing, or is it lumpy?
A steady, predictable increase in marking volume makes a strong case for a hire, since the fixed cost of a salary gets justified by consistent demand. A volume that spikes around specific enrolment periods and drops off between them is a much weaker case for a permanent hire and a much stronger one for a tool that scales down as easily as it scales up.
How much of your bottleneck is judgement, and how much is throughput?
If your problem is genuinely that nobody on your team has the specific expertise to mark a particular subject well, a hire solves something a tool cannot. If your problem is that your existing, qualified assessors simply don’t have enough hours to get through the volume, that’s a throughput problem, and it’s exactly the kind of problem AI-assisted marking is built to help with, since it changes how long each first read takes rather than requiring someone new to be trained from scratch.
What’s your actual ramp-up time for each option?
A new assessor typically needs weeks to months before they’re marking independently and consistently against your standards. A well-implemented AI-assisted workflow can be live in days, and the review step it introduces is, if anything, a faster way to train a new hire’s judgement too, since they’re reacting to a structured draft rather than developing their approach from a blank page.
How reversible is each choice?
This is the question people underweight most. If volume drops next year, reducing a tool subscription is straightforward. Reducing headcount is a genuinely different, much harder conversation, both practically and in terms of the relationship you have with the person you’d be letting go.
These aren’t actually mutually exclusive
The framing of “hire or automate” implies you have to pick one, and in practice, most providers end up doing both over time, just not simultaneously and not for the same reason. AI-assisted marking tends to be the faster, lower-risk first move precisely because it’s more reversible and addresses a throughput problem immediately. A hire becomes the right call once volume has grown to the point where it’s clearly durable, not seasonal, and once you’ve confirmed, often because your existing assessors are now spending their time on the higher-judgement parts of marking rather than the repetitive first pass, that what you actually need is more specialist judgement in the room, not just more hours in the day.
Frequently asked questions
Is AI-assisted marking always the right first move for a small provider?
Not always, but it’s usually the lower-risk one to try first, since it’s faster to implement and easier to scale back than a hire. If your actual problem is a gap in specialist expertise rather than throughput, a hire may be the more direct fix.
How do I know if my volume problem is throughput or expertise?
Ask whether your existing assessors could clear the backlog if they simply had more hours in the day. If yes, it’s a throughput problem. If the honest answer is that nobody on the team is confident marking that specific subject well, it’s an expertise gap, and a tool won’t close that on its own.
Why does reversibility matter so much in this decision?
Because demand for training providers is often seasonal or lumpy. A choice you can scale back easily if volume drops carries much less risk than one, like a permanent hire, that’s genuinely hard to reverse.
Do most providers end up choosing one option over the other permanently?
Not usually. Many end up using AI-assisted marking to handle throughput immediately, and add a hire later once volume growth looks durable rather than seasonal.
Sources
Coase, R. H. (1937). The Nature of the Firm. Economica, 4(16), 386-405.
Try the reversible option first
Eduface’s AI-assisted marking is live in days, scales with your volume, and doesn’t require a permanent headcount decision to find out if it helps.