• AI

AI, the billable hour and the changing value of experience

For a long time, paying for digital product development by the hour has made sense. ‘Time and Materials’ is simple; a piece of work takes a certain amount of time, a team has a certain amount of capacity, and you pay for that time. But with the rapid development of AI, the relationship between time and what can actually be delivered is beginning to change.

Some tasks that used to take days can now sometimes take hours, giving teams more room to explore options, or get through repetitive work much faster. This gives experienced teams more room to focus on what matters: the tasks that need true human experience and judgement; understanding the problem, knowing what to produce, and making sure we produce the right thing.

It’s a question I’ve been spending a lot of time on at Brightec over the past six months: if AI changes what an experienced team can achieve in a day, what does that mean for how we price our work?

So, if the time it takes to do something is falling, does the value of what you’re buying fall with it? I don’t think it does.

At the end of the day you’re not really buying hours

At Brightec, we have a very senior team, with hundreds of years of experience between us, and several people have been here for more than a decade. That experience holds real value. We’ve seen a lot of problems before, we know the questions to ask, where issues are likely to occur, and when something needs more investigation. We can often get to a sensible answer much faster because we’re not starting from scratch each time.

AI is now giving that team better tools to work with.

If something that would previously have taken four days can now be done well in two, the opportunity isn’t simply to save two days. It’s to use that efficiency to create a better end product. We can explore more options before making a decision, test ideas that previously might not have been practical, spend more time understanding the problem, or iterate further once we have something working. 

Our goal should be faster and better, rather than simply faster and cheaper. 

However, doing more doesn’t automatically mean creating more value. If a job that previously took four days could now be done in two, it’s fair to ask why that doesn’t simply mean a lower bill.

For me, the extra capacity is only worth anything if it’s spent deliberately.

There’s no point adding work just because we can, but if that time can be used intentionally to reduce risk, improve the experience, or explore something that wasn’t previously practical within the time available, then it can create real value.

Knowing where that additional time is best spent is part of what an experienced team is for. The efficiency isn’t the value itself, it’s what that efficiency allows us to do.

So why not just value price everything?

Value pricing isn’t new, and we didn’t just stumble across it, but knowing that something is valuable and being able to confidently define and price that value are two different things.

A lot of the work we do with clients starts with a question rather than an answer. They rarely come to us with a clearly defined project, a fixed scope, and a list of requirements. Often, they come with a problem, an opportunity, or simply a sense that something needs to change.

And that’s not limited to product development. Sometimes we’re helping a client work through a strategic problem, sometimes we’re shaping what a product or service should be, sometimes we’re building something, and sometimes it’s a combination of all three.

A big part of our job is helping clients get one step further on their journey, and we can customise the shape of that work pretty significantly depending on what the client actually needs. That makes it difficult to say at the start of every piece of work exactly what the value will be at the end. Sometimes the value is in the journey, with the answer emerging through the work, rather than being something that can be confidently defined on day one.

That’s one of the reasons Time and Materials has been a practical model for a lot of digital product development. It allows both sides to deal with uncertainty and change without having to pretend that the end result is known before the work has started. It still comes with its own challenges, and it can become a constraint if greater efficiency leaves less room within the commercial model for further product refinement.

At the moment, we have all three models running across Brightec, and across a varying portfolio of work. We have Time and Materials on some of our more exploratory work, or larger maintenance contracts, fixed-price work where the scope and outcome are well understood, and team or capacity-based models where a client needs a consistent team but the priorities within that capacity are likely to change.

That might be a large, complex delivery where we know what needs to be built and can agree the outcome upfront. It might be a smaller, clearly defined piece of work where a fixed price gives everyone clarity from the start. Or it might be an ongoing relationship where the client knows they need a particular set of skills and capacity, but the work that team will be doing needs to have flexibility to change as the client’s priorities change.

They can all be the right answer, depending on the work.

The task isn’t picking one model and applying it everywhere, instead it’s about understanding the shape of the work and choosing the model that fits it.

What AI changes, and what it doesn’t

AI doesn’t make an experienced team less valuable. If anything, I think it makes experience more important, because it gets better at helping people produce things without getting any better at knowing what’s worth producing. The real advantage in product development comes from knowing what to produce in the first place.

I don’t think human judgement can be so easily commoditised in the way we’re seeing that happen with execution, and this is why I don’t think you can price experience out of the model, even as delivery gets faster. Ethical product development requires human oversight, and faster tools makes oversight more important, not less.

For me, the goal isn’t to slow down innovation. It’s to make sure that as we get better at building things, we’re also getting better at taking responsibility for what we build.

As AI plays a bigger role in how products get designed and built, keeping decision-makers close to the people affected by those decisions matters more. That’s part of the same argument as the pricing question, not a separate one: judgement is the part that doesn’t simply get cheaper as tools get faster, and this is what we’re actually pricing.

Beyond the billable hour

I don’t think the billable hour is going away, and it doesn’t need to. But as AI continues to change what an experienced team can achieve in an hour, a day, or a month, it’s worth being precise about what we’re actually charging for. 

There isn’t any particular driving force to move away from Time and Materials, or move towards value pricing everything. Instead, it’s about making sure the commercial model reflects the work we’re doing and the value we’re trying to create. 

But there is a bigger question for agencies too. If the way work gets delivered changes, then the agency model around it needs to change with it. How we structure teams, how we think about capacity, how we build our cost base, and how we make sure projects deliver the right level of return are all part of the same conversation. 

This is what I’ve been working on at Brightec: how we structure and price our work to reflect what an experienced team is now capable of delivering, while making sure the wider agency model works commercially. 

I don’t think there’s one answer, I think sustainability comes from being willing to adapt. The more varied the work we do becomes, and the more capable our teams become with AI, the more important it is that we choose the commercial model deliberately, on each piece of work, rather than reaching for whichever we’re most familiar with. 

AI might keep changing what’s possible in an hour. We need to get better at understanding what that hour is worth.