Decoding ROI from AI

What you see in the initial use cases for AI is really the productivity and efficiency improvements for the people in an organisation, providing those general-use productivity tools and giving them those tools so they can make their work more efficient. But it’s not where people are going to see the returns that everybody’s looking for in the world of AI.

Leaders who are seeing positive financial results from AI are much more likely to use it across their value chain and in more sophisticated ways. These outliers are about twice as likely as their peers to have scaled or embedded AI across major business functions, including strategy, marketing, supply chain, and support services like finance, IT, and HR.

When companies put AI in more places and push it deeper into day-to-day execution, they see better results. The organisations that go after the hard problems, they put the more sophisticated deployments in place. Those organisations are seeing outsized return, both top line and profitability.

That’s a really important insight because what it tells you is that the power of AI, the ROI for AI, is in the hard problems: the large-scale work transformations that are workflow-oriented, value-oriented, not just task- or individual-oriented.

So how can your company maximise AI use in more impactful ways? It starts with a phased approach. Identify a small number of high-volume workflows that can deliver real business value. Define AI guardrails early and assess where it can handle repeatable judgment calls, leaving humans to focus on exceptions.

Put it all together, and you make AI integral to how the business runs. And that is the key to realizing top-tier performance.

Thinking about AI today, I think sophistication is much more important than ever before. The last few years have been about experimentation and efficiencies, but the tools and technologies are rapidly advancing so that the innovation is making us think much more deeply about our businesses, which inherently needs a sophisticated solution.

Our research tells us that leading companies are almost two times as likely to be operating at much higher sophistication levels with AI. The companies that are winning are not thinking about AI as a chatbot or going after robotic process automation. They’re fundamentally rewiring their processes, using data differently, and unleashing the power of autonomous agents to change the way they do business every day.

What that means is they’re building in guard rails. They’re looking at how to use this throughout every aspect of their organisation.

Sophisticated use cases are hard, but they also have the most upside. When I think about the pharmaceutical industry, bringing a new product to market is highly complex, from regulatory, medical, and legal, compliance. We also have to market. So taking AI to bridge a gap between marketing and compliance and scientists, that’s where AI really stands out.

This is a complex problem, and it crosses the entire organisation. One AI model can support every aspect of an organisation. So creating reusable assets to basically supercharge what teams are doing across different business units is where the real value is. That way, I can control the guard rails, I can build in the right compliance and controls so that it’s solved once, for everyone.

So when I think about where to start, I go past efficiencies and look for the biggest, hardest problems; the most complex data; the most complex process with the most humans involved. AI is built to help in those situations. We have to build it responsibly, we have to get the workforce on board, and we have to build a system that can fundamentally think about our processes differently. That’s the revenue upside for your enterprise.

For companies to see returns on their AI investments, they must put in place strong foundations, starting with innovation. Companies can begin by using AI to discover new value pools that are forming around shifting customer needs. They can then capture growth by quickly developing new products and services around those needs.

Leading companies that outperform their industry peers, when it comes to AI, generate twice as much revenue from products and services they launched in the past three years. They are also 2.6 times as likely to say AI helps innovate their business models. This shows that being serious about innovation really pays off.

These high performers also promote innovation across their workforce. They appoint innovation owners who have the task of supporting AI innovation projects within specific business units, and they use performance incentives to encourage employees to take a proactive test-and-learn approach to AI. And they set up the right infrastructure so that talent can experiment with AI creatively, but also safely, in dedicated sandboxes.

Lastly, top performers take a disciplined approach to managing innovation. They run structured reviews to decide which innovation projects to progress, to prioritise, or to end. And they do that by using clear metrics to scale what works and stop what does not.

So leaders deploy their resources into ideas with real commercial impact. In this way, top-performing companies stimulate AI innovation across the enterprise, setting them up to compete and win in promising new value pools.

AI can drive business innovation, from reshaping customer experience to developing new revenue streams. Our research shows that companies getting measurable returns from AI are scaling proven use cases across the value chain and enterprise functions.

But as AI scales far and wide, so can risk. Take chatbots, for example. If chatbots are not designed the right way, it can easily lead to privacy breaches.

If you think about training AI models, when you scrape the data, it can easily lead to IP infringement issues, legal harm, reputational harm, financial harm. High-performing organisations really understand this at a deeper level. They are 1.7 times as likely as their peers to have a documented Responsible AI framework guiding their strategy as well as execution.

Right-sized governance and risk management support AI innovation and growth. So, as companies think about deploying AI at scale and ingraining AI into the day-to-day operations, ensure the right guard rails around how the AI can be used, where it should not be used, what are the policies around it. That will unlock the value that they’re looking for.

A fundamental underpinning of safe use of AI is trust. Whether you’re an employee, whether you’re a manager within an organisation, or even those charged with governance, if you know that there are safeguards and guard rails around AI, one is more likely to adopt, experiment, and unlock the value that organisations are looking for.

Leading firms tend to have role-based data and AI access controls to protect privacy. They also create systems so that teams don’t have to come up with governance approaches on a case-by-case basis. High performers set up a standard process to gauge risk for each use case and add controls to product and delivery processes right from the very beginning. This allows them to replicate use cases across functions and markets. It also cuts out late-stage rework.

AI is cutting the innovation cycle, from idea to market. If you think about the world today, there’s breakthroughs coming almost every day in the sense of AI or agents or data. So we have to think differently about how we rewire systems. One of the first steps is to break the boundaries of data.

We’re no longer in the world where data is locked in an application or a database. AI thrives when we combine that data and find new ways to solve a problem. We have to think outside the organisational boundaries. The light bulb moment is that the marketing team, and the digital commerce team, and the supply chain team all have to solve the same problem with the same AI.

Reusable assets, especially around data and especially around agentic architectures, is where the unlock comes in for an organisation. So our findings show us that companies that get the data right are 2.4 times as likely to create reusable AI assets around that data.

Rethinking our approach to data means thinking about it from the lens of an AI agent. How do we support that agent to make better decisions faster? How do we protect data that sets us apart from our competitors? What do we feed into the model to train it? And what do we hold back to just use it? So that tacit knowledge is what we’re after.

It’s knowledge that exists across an organisation that a human wouldn’t see. Humans should still make judgements. Humans should still be in charge of the strategy and where we go as a business. But AI is going to unlock, with the right data, knowledge across an organization that just is not apparent to us the way we work today.

To win in this world of AI, you’ve got to get something right. And it’s that data foundation that sets you apart: the connection of data across systems, the encapsulation of those data products, and the reusable AI assets to make you differentiated in the market.

For many companies, AI adoption means giving employees access to tools and encouraging them to experiment. While this approach has led to some isolated success, it falls short of producing the big financial gains that executives want to see.

A small proportion of firms are doing things differently, and it’s yielding better outcomes. PwC researched the AI practices and performance of over 1,200 companies and found that roughly one in five is getting far stronger financial results than its peers.

Strategic discipline is the driving force of their success. These companies are implementing AI strategy that closely maps to their top business priorities, rather than a collection of individual pilot projects. They’re designing AI use cases to promote efficiency and revenue growth, tracking business impact, and holding leaders accountable for the results. This disciplined approach is improving performance.

Compared with other companies, AI leaders are two-and-a-half times as likely to say that they’ve gotten better at creating new products and services and bringing them to market much faster. They’re more likely to have transformed their business and operating models, and they’re improving customer satisfaction and employee productivity.

So, what can your firm do to unlock the power of AI? Make sure that you start boosting your AI fitness by making early strategic choices. Fund, flex, and manage your AI portfolio, quickly weeding out the losers and scaling the winners. Through strategic use built on firm foundations, you, too, can master AI adoption and implementation and position yourself to outperform your peers.

Every organisation is at a different point in its AI journey. Many are asking the same question: how do we turn investment into measurable results?

According to PwC’s research, the top performers invest more than double the amount in AI as a share of revenue when you compare them to their peers. Companies capturing the most ROI and not just spending more on AI, they’re running it like a business discipline. They’re setting priority use cases and then tying them to revenue, margin, and risk. They’re selective about where they scale, and they move capital to initiatives that are delivering results.

Top performers are 1.3 times as likely as their peers to reallocate funding and resources as their priorities change. And critically, the most effective organisations fund AI projects that are directly tied to business objectives: so not just cost savings but revenue growth, risk management, and client impacts.

AI investment has to support quality, compliance, and new services. Tax strategy shapes how and where value is created. It should be built into AI investment from the start. Many jurisdictions offer research and development credits, incentives, or capital allowances for tech development and deployment. So structuring AI programs to qualify for one of these benefits can really lower the net cost and improve returns. Decisions such as where to locate people, data, and intellectual property should really take tax into account and consider how to protect margin and reduce risk.

So, how can your organisation turn AI investment into measurable results? By setting clear priorities, establishing the right operating model, and aligning across technology, finance, tax, and legal. These foundations have the power to transform ambition into outcomes in a way that’s commercially sound and globally coordinated.

To get the most out of AI, the workforce is absolutely essential. Leaders need to make sure that their people have the skills and the confidence to use AI in the right way and that those skills keep evolving as the technology evolves.

The organisations getting the most value from AI are investing in clear skills pathways so that their people grow alongside the technology. When that happens, AI stops being a technology experiment and becomes a real driver of business performance. In fact, according to our research, top-performing companies are 1.7 times as likely to provide ongoing, role-based AI learning. That means helping people understand how to use AI in the context of the job that they actually do.

A big part of this, and central to this, in my opinion, is building trust. Leaders need to be clear about how AI should be used responsibly and what good human-AI collaboration actually looks like. When people understand that AI is there to augment their skills rather than replace them, that’s when real collaboration starts to happen.

Just as important is creating a culture where people feel comfortable to experiment. That experimentation drives engagement, helps organisations learn faster, and ultimately improves performance.

Our survey shows that organisations that are getting ahead in terms of the results from AI involve cross-functional teams when developing AI solutions. That means bringing together teams early to get their views on where the technology can be embraced as they redesign the work that is being done.

When AI is built into systems with the right guard rails, it can automate routine decisions and processes. That frees people up to focus on the harder things: being creative, problem solving, and strategic decisions. When organisations redesign that flow of work, embrace the technology, equip their people with the right skills, that’s when we see organisations leaping ahead in terms of superior business performance.