We have officially entered the era of “applied AI”. For many boards and C-suite leaders, understanding how to gain the most value from large language models and agentic AI is now the single most important strategic challenge of the day. From scaling up pilots to securing investment, driving measurable business impact to bringing employees along for the ride, the road to AI maturity is fraught with challenges.
To dive further into this topic, and understand how enterprises are finding successes with AI, we convened a panel of technology leaders to share their insights and advice.
From pilots to enterprise transformation
One problem routinely cited by executives looking to make good on their AI investments is that they can get stuck in ‘pilot purgatory’, having started a number of exploratory projects before finding they won’t work on a larger scale.
For Rahul Shah, global chief digital and information officer at Mars Pet Nutrition, the key is to break the process down into simpler steps. “We ended up saying: instead of immediately focusing on scale, let’s define the five big bets we’re going to make. Then we made the shift from pilots to scale, then from use cases to capability, and finally from information to decisions.”
To identify these “big bets”, our panelists agreed that the best way was to delve into how employees are actually working day to day and seeking out opportunities to lighten the load. “You can work top-down, but you can also work bottom-up,” says Ursula Soritsch-Renier, group chief digital and information officer at Saint-Gobain, “using pain points employees throughout the business encounter every day.”
Nigel Richardson, chief information and digitization officer at Reckitt, emphasizes that centering AI projects in people’s everyday work is the key to avoiding pilot purgatory. “Doing pilots is incredibly quick and easy—and you can do such impressive things really quickly. To really build something that is scalable is a whole different world. What we found useful is going deep into processes and end-to-end workflows and making sure it’s not just throwing new exciting tools in but really understanding how people work and how we can reinvent that work in the future using AI.”
Not everyone agrees, however, that pilots are the new AI-related pitfall. “I love pilots, I think pilots are great” says Bruno Zerbib, chief technology and innovation officer at Orange. “I hate the pressure of trying to please people by going fast so everyone feels we are progressing at the ‘right’ pace—the reality is there is no playbook. We’re all discovering and learning and the most important thing is being humble and not caving to pressure to come up with random milestones to prove we are a great ‘AI company.’”
“You can work top-down, but you can also work bottom-up”
Ursula Soritsch-Renier, group chief digital and information officer at Saint-Gobain
Practical takeaway: “Pick the right business problem, secure top-down sponsorship, and then make sure you really go into workflows in depth,” says Richardson. And don’t be scared of pilot purgatory, see it for what it is—a place to explore AI’s manifold possibilities.
AI as organizational transformation
The reason Orange’s Zerbib is cautious when it comes to rolling out AI programs at scale is because he recognizes the second key challenge facing leaders: bringing your people on the journey with you.
“We have to be very careful with the notion of going fast at the expense of doing things the right way,” he says. “At the moment, we are picking the right job lines [to augment with AI], the ones which we think will give us return on investment, and they are acting as trailblazers. We’re not going to solve world hunger, but we want to have great stories that people did not lose their jobs but, on the contrary, AI made their life more fun than ever.”
At Saint-Gobain, Soritsch-Renier acknowledges that the workforce is often less literate from a technology perspective, as the organization is an industrial business focused on construction materials and, as such, hasn’t been called to embrace technology at the same level as other industries. Here, there is a huge opportunity to build enthusiasm among more skeptical colleagues. “Our people are spending far too much time on administrative work,” she says. “If you can reallocate the same capacity, resources and effort you’ve used for processing accounts receivable into cross-selling or upselling, then there is opportunity there. As long as people are willing to evolve, learn and grow, there is no risk.”
Richardson agrees, citing particular wins in the company’s R&D department. “We were finding that 30-40% of our scientists’ time was being spent on documentation,” he says. “That was a huge bottleneck. So, we developed an agentic AI solution called Write-It and something that was taking days now takes minutes and frees up their time to do much more innovative work.”
“We have to be very careful with the notion of going fast at the expense of doing things the right way”
Bruno Zerbib, chief technology and innovation officer at Orange
For leaders looking to communicate this message to their wider workforce, Shah has a positive framing. “All the jobs which are there to coordinate information from one place to another are going to be diminishing,” he says. “But this will create more choices than we have ever seen. Your human judgement is going to become even more important.”
Practical takeaway: It is within the C-suite’s gift to transform the daily working lives of their staff by making work less mundane and more rewarding. Lean into this. And get the messaging right. For Zerbib, paraphrasing Nvidia’s Jensen Huang can be helpful for this: “You will not be replaced by AI. Your job will be replaced by someone who knows how to use AI.”
Building credibility for AI investment
Another common problem for leaders is dealing with the pressure to innovate or the hesitancy to invest from the board. Executives must therefore learn how to communicate the benefits of AI clearly and comprehensively.
“In my experience, the board just wants to grow the business,” says Shah. “Technology is just one lever. Often the board’s questions around AI are not about what use cases you’re developing but about how you are growing and protecting the business. Our job is to separate the signal from the noise.”
One clear way to do this is through regular communication. “We are very focused on the tangible business cases of our major AI investments,” says Richardson. “Every quarter we review all the AI initiatives and look at the benefits we said we’d get, what we are getting, and how we can continue to improve and learn.”
The clearer the business benefit, the easier it is to have the conversation, of course. At Saint-Gobain, one solid example of this for Soritsch-Renier is an AI tool which makes it easier for employees to read through tenders for big projects. “We used our tool to scan 12,000 tenders and we can now select leads which are 15% more qualified for us and from which we see a 10% higher conversion rate.” These are numbers which are bound to excite any board.
It is, however, key to be just as open about where projects are struggling as where they are succeeding. “We’re not trying to tell them fairy tales,” says Zerbib. “That can be very dangerous right now—honesty is super important.”
Practical takeaway: As with any important relationship, the secret to building trust and securing board buy-in is simple: “Communicate, communicate, communicate,” says Soritsch-Renier.
