BLOG

What to work out before your AI pilots scale

Scaling AI pilots? Assess how tasks, roles and skills could change, and what the business needs to do to turn promising results into commercial value.

•
4 min read
Helena Turpin
Co-Founder, GoFIGR
What to work out before your AI pilots scale | GoFIGR
5 second summary
  • Promising pilots still need a plan for how the work will change.
  • Look at tasks, responsibilities and skills before making workforce decisions.
  • Test assumptions before committing to a wider rollout, and revisit them as evidence improves.

‍

AI has given us an extraordinary opportunity. The slightly more annoying part is that our competitors have access to many of the same tools.

If you’re already running AI pilots across the business, you probably don’t need another explanation of how much time someone can save writing an email. You’re trying to work out where AI could make a commercial difference.

Some pilots may be promising. Others still have plenty to prove. But if a pilot does work, there is another question worth asking before you scale it: what would have to change about the work, and the people doing it, for the business to benefit?

Perhaps you could serve more customers without growing the team at the same rate, bring new products to market sooner or offer a service that isn’t viable today. Each possibility means making choices about how work gets done. A successful pilot alone won’t tell you what those choices should be.

Start with the result the business wants

At GoFIGR, our team started with our company goals: get ideas to market faster, experiment more and build things we could trust. Then they began pulling apart the way they worked to see what AI could take on, what people should still do and how the whole thing needed to fit together.

Our developers’ jobs are night and day compared with the beginning of this year. Instead of spending most of their time writing code, they’re directing AI, questioning its output, setting standards and deciding whether the result is good enough. They still own the consequences when something goes wrong, and we’re still working out some of the answers.

One of our developers put his finger on why this is harder than it looks. Even after AI had built something and passed its own checks, he wasn’t always sure it really worked. He didn’t want to become the person who simply moved work along for somebody else to test, without owning the result.

That is a useful question to take into any discussion about scaling. If AI takes on more of the execution, who will know whether the result is good enough, and what will they need to do that well?

Follow the work beyond the pilot

A pilot can show that a tool performs a particular task well in a particular setting. You still need to understand how that task fits into the rest of the business.

For each pilot you are considering taking further, work through these questions with the people closest to the work:

Time saved may create room to handle more demand or improve service. Turning it into a lower cost requires further decisions about how work is organised. It helps to be clear about which outcome you are aiming for before adding up the benefits.

Want to see what this looks like for your own role?

Try the free AI impact assessment

Look across jobs before making workforce decisions

One leadership team we worked with expected “AI in, headcount out”. The analysis showed substantial automation potential spread across departments rather than neatly replacing whole jobs. Helping people work more effectively offered a more realistic near-term return, while they planned for the bigger changes.

Another organisation was less interested in immediate savings. They knew their workforce would look different in a few years and wanted a basis for decisions today: which skills to develop, where to recruit, what to budget for and whether every vacancy needed replacing like for like.

Those are decisions worth considering while pilots are still being tested. You can explore what would change if a pilot scales without assuming that it will, or committing to a workforce plan before the evidence is there.

There is also a question about how people learn. Experienced people might be delighted to give routine work to AI. For someone new to a profession, doing that work is often how they learn to spot an error or recognise a good decision. We can’t just remove the junior tasks and assume we’ll somehow still develop the experienced people we need in five years’ time.

Build a view you can test

You don’t need a six-month time-and-motion study to get started. Use the role, workforce and technology information you already have, alongside what your pilots are teaching you, to explore a few plausible scenarios.

For a priority area, compare how the work might look if the pilot stays limited with how it might look if it is adopted more widely. Consider the tasks that would change, the skills people would need and the likely effect on capacity or service. Include the extra work and cost of checking outputs, training people and changing the process.

Show your assumptions openly. Use ranges where the evidence is uncertain, and validate the biggest opportunities with the people who understand the work. Be clear about what a pilot still needs to prove before you invest further.

That gives leaders something they can discuss, challenge and act on. You may decide to expand a pilot, change its scope or stop it. You may also find a recruitment or development decision that needs attention now, even while the technology is still being tested.

Nobody needs a perfect prediction of every job in 2030. They need to make a better decision now, and be able to revisit it as things change.

Who can help you understand how AI could change your workforce

At GoFIGR, we help People and transformation leaders understand how AI could change their workforce. Our paid organisational assessments examine the tasks within roles and model what different AI scenarios could mean for skills and capacity, giving you evidence for decisions about recruitment, development and how work is organised.

Talk to us about the decision you need to make and the evidence you already have.

Book an AI impact briefing

‍

Helena Turpin
Co-Founder, GoFIGR

Helena Turpin spent 20 years in talent and HR innovation where she solved people-related problems using data and technology. She left corporate life to create GoFIGR where she helps mid-sized organizations to develop and retain their people by connecting employee skills and aspirations to internal opportunities like projects, mentorship and learning.

GoFIGR Newsletter

Sign up for the latest resources on how to keep your people engaged, happy and high-performing

By signing up, you agree to GoFIGR’s Terms of Service and Privacy Policy.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.