{{item.videoDuration}}
{{item.title}}
{{item.text}}
{{item.videoDuration}}
{{item.text}}
Key takeaways:
- AI strategy should go beyond tools and use cases to focus on growth and competitive advantage.
- AI can create new paths to growth, from strengthening the core to reinventing business models.
- Lasting advantage comes from AI-powered moats built on what’s unique to your business.
- The window to differentiate is open now. Leaders who make deliberate choices today can build an advantage that lasts.
Most companies have what that they call an “AI strategy.” But it’s typically about which tools to buy, how fast to roll them out for which use cases, and what costs to take out. That may be enough for incremental value. But it probably can’t give you a competitive edge. Most of your competitors are rolling out AI in similar ways.
What can put you ahead is differentiated AI strategy. By doing things differently, you can find new growth and dig new moats. A differentiated AI strategy covers not just how to use AI tools, but also how to change your business so AI can power new growth, while helping you navigate these top constraints:
- Today’s business against tomorrow’s. It’s the incumbent’s challenge: Even as you shift money and people toward a new AI-powered business, your old business needs support too, and its stakeholders and structures may resist change.
- Capital that’s spread too thin. Stakeholders may all want a piece of AI. But if you allocate capital evenly across every function, you probably won’t transform any of them. It’s better to pick the right spots to concentrate your capital, then scale success.
- Roadblocks to scale. Without the right operating model and architecture, even successful pilots won’t scale: People resist adoption, siloed and unreliable data impedes performance, and stakeholders don’t trust outputs.
- Vendor lock-in. If you don’t build an architecture that can flexibly route the right job to the right model for both cost and performance, you may suffer on both counts, as well as and find it difficult to switch to better models as the frontier advances.
- The burden of proof. AI is advancing so quickly, and much of what it can do is so new. It’s critical to use credible benchmarks that demonstrate measurable value to the CFO and other stakeholders.
It’s not easy to overcome these constraints, but the reward can be new growth that’s hard for rivals to match. Whether you’re a CEO, CFO, chief strategy officer, or on the board, here’s what you need to know.
How to grow faster and be harder to copy
For some companies, AI is delivering growth. From 2022 to 2025, S&P 500 firms (excluding AI foundation builders) that invested 1.6% of their revenue in AI had revenue growth that exceeded sector medians by 3.5%, according to PwC benchmarking data. Their total shareholder return (TSR) outperformed their sector's average by 20.2%. Lighter spenders underperformed by 4.1%. And 2025 is ancient history for AI. It can do far more today.
In our experience, AI-driven growth can happen in four ways, starting with lower risks and faster paybacks, then advancing to higher risk but also higher rewards.
It’s critical to move fast: AI natives, with no legacy to defend and no reflex to protect what AI makes obsolete, can be nimbler than incumbents. But the incumbents also have advantages: Data and relationships from your current customers can give you an edge, so long as you build on it faster than AI natives can close the gap.
Win the core. Whether it’s your brand, network effects, process power, or scale, there’s likely some core strength where AI could grow your lead further, even as AI natives strive to catch up. One restaurant chain we assisted, for example, has scale as one of its strengths. To help double its footprint, it used AI to clear its biggest bottleneck: ramping up its workforce.
Grow adjacently. Once AI collapses your cost of serving customers, you can expand into adjacent segments and geographies. But AI natives may enter too, so speed is vital. We helped an auto insurance company stand up an AI-powered platform, so it could overcome a shortage of claims estimators and seize an “adjacent" opportunity fast.
Build AI-native offerings. All-new revenue lines may be in reach. You might, for example, create products by embedding intelligence, or sell data or expertise in packages. We helped a healthcare company use AI to unify over 50 data domains and turn insights into products for new research, pharmaceutical development, and strategic partnerships.
Reinvent the model. You sell (for example) subscriptions, uptime, or outcomes instead of products, equipment, and one-off sales. By rewiring the economics, this can create a new moat. We helped one tech company move from selling siloed hardware to services and outcomes, growing recurring revenue to over $1.4 billion and doubling cross-sell.
Where to start isn’t a matter of ambition. It’s a matter of how strong your current core business is and how much risk your business can and should carry. Ideally, you should advance up these four types of AI growth like steps on a ladder, with the first step creating new revenue to fund the next. But your competitive landscape may oblige you to take on more risk early, skip a step or two, and move faster.
Staying on top: How to build new moats with AI
With the marketplace changing at AI speed, it’s not enough to achieve new growth. Your strategy has to plan for new moats too, so you can keep your lead. Your competitors, after all, will likely have access to the same AI models that you do.
What can make growth last is what’s unique to you. That can be data, if you turn it into a "flywheel”: As your AI runs inside your business, it learns from outcomes that others can’t see. Or your moat can come from AI that leverages your unique distribution network and physical installed base. It’s also often your people and how quickly AI enhances their decisions. And, especially in highly-regulated industries, your moat can come from trust: You can grow where less-governed rivals can’t follow.
Whatever your new moats may be, a differentiated AI strategy should identify them early and design them into your AI growth plans.
AI-powered opportunities: by sector
| Sector | Industrialized | Emerging |
CM Consumer Markets |
|
|
E&I Energy & Industrials |
|
|
FS Financial Services |
|
|
HI Health Industries |
|
|
TMT Technology, Media & Telecom |
|
|
Cross-industry Functional value pools |
|
|
How to win with technology: Decisions for strategy leaders
As you chart the future, you'll need to determine how to approach technology:
- Should we rent or own it? In general, you should own what can create compounding value: domain data products, evaluation sets, orchestration logic, tuned models, and model weights. Vendors can usually provide the rest.
- Where should it run? Inference at the point of work boosts speed and resilience, but centralization saves energy and computing power. Strategists will have to balance priorities and resources.
- What do agents own, and who owns them? An orchestration layer can let you decide what agents do while creating audit trails, human reviews, and more. Without this layer, vendors may make these decisions for their benefit, not yours.
- How do we leverage falling unit costs? Cheaper AI often drives higher AI spend as volume surges. Instead, make a strategic choice for each function to use savings either to drive new growth or cut spend, then create oversight to make it stick.
- How will we sell to agents? As consumers and business clients increasingly use agents (often embedded in everyday applications) to find, assess, and purchase products, your front office has to be ready to compete.
Making AI strategy work in the real world: 6 actions to take today
It may feel like a big lift to make AI strategy a growth engine while also navigating your constraints. But with swift action in six areas, you can move up the “scale curve”: achieving synergies and growing across industries as new markets become reachable.
- Ambition: Choose an AI portfolio that can win. Concentrate capital by picking a few spots where AI can drive growth. Starve the rest. Companies who concentrate bets (as part of bimodal strategy) see twice the revenue growth of peers.
- Organize: Pick an org and delivery model. To build a new business as you support the old one too, give one senior leader a mandate to choose where to lead, lag, or exit and head a new delivery model, whether hub-and-spoke or centralized.
- Execution: Deliver value with KPIs that matter. Create an operating layer to govern agents, control AI spend, modernize and monetize data, and track business-relevant benchmarks. Agentic scaffolding can help you create new, AI-native workflows quickly.
- Technology: Architect for where AI is going. Evolve and enhance your existing enterprise architecture to create an AI architecture that works within your existing ecosystem, avoiding vendor lock-in and enabling AI at scale. A control plane can let you orchestrate across different models (including open-weight ones), and swap in new models as desired.
- Talent: Focus on the future. If you redesign workflows for AI (rather than layering it onto old processes) adoption can surge. Add incentives and a narrative, rethink your workforce structure, and be ready to reinvest the workforce dividend.
- Trust: Build confidence to scale AI. A strong trust AI approach, supported by calibrated governance and rigorous AI assurance, can help you identify risks early, strengthen stakeholder confidence and drive adoption at scale.
The bottom line
AI will soon be in every company, but it will only create advantage for a few: Those who use it to fuel a new growth strategy, not just as a technology to deploy as quickly and cheaply as possible. And, since most companies face the same constraints with AI, what can set you apart is how you navigate them: turning the challenges that you and your peers face into a strength that can give you new moats, built to last.
The window to make strategic choices that can differentiate you for the long run is open. It won’t stay open for long.
{{filterContent.facetedTitle}}
{{contentList.loadingText}}
{{contentList.loadingText}}
Contact us
Follow us
