Patrick Pugh is PwC‘s US and Global Alliance & Ecosystems Leader, helping organizations accelerate AI-driven business transformation.
Enterprise leaders are overwhelmed with AI choices and options. New models, agents, platforms and ecosystem partners are emerging at a dizzying pace, each promising to change how businesses operate.
Optionality and choice are typically a good thing, but when I speak with leaders, I see another side of it. They are being asked to move faster on AI while making decisions in a market that is changing underneath them. Which technologies deserve investment? Which partners will matter two years from now? Where should they commit, and where should they keep their options open? Where can they get an expected return?
When nobody knows exactly which technologies will win, the natural response is to hedge. Launch another pilot. Add another tool. Test another provider. Keep as many options open as possible.
But optionality has a cost.
The companies that pull ahead in this next phase of AI won’t necessarily be the ones making the most bets. They’ll be the ones with the discipline to make fewer, clearer bets and create the space for what works to scale.
Experimentation should lead to decisions.
Experimentation has been essential to the progress we’ve made with AI. Organizations needed room to learn what the technology could do, where it could create value and how people would use it.
But experimentation is a means, not an outcome.
There’s a difference between experimentation that helps an organization decide and experimentation that delays progress. When every promising technology becomes another pilot, organizations can spread their people, capital and attention across so many initiatives that few have what they need to become meaningful at scale.
That can take a toll on people, too. Employees are already navigating a constant public narrative about how AI will change their jobs, while being asked to adopt new tools and rethink how work gets done. When organizations layer one experiment on another without clear direction, that uncertainty can turn into caution. Instead of feeling empowered to use AI to rethink work, experiment and find new avenues for growth, people can become hesitant to commit to any one path.
The next phase of enterprise AI will require as much discipline as curiosity. Leaders still need to encourage teams to test and learn, but they need mechanisms to decide what moves forward, what gets investment and what stops.
Stopping matters. Every initiative an organization funds is a choice not to concentrate those resources elsewhere.
You don’t need to predict the winner.
The market may look very different a year from now. That uncertainty can create paralysis, with leaders preserving optionality rather than committing to the wrong model, platform or provider.
I think that’s the wrong standard.
The goal isn’t to perfectly predict which technology will win. It’s to get clear on the outcomes you want to achieve and confidently match the technology to those needs, rather than reacting to what’s generating the most attention in the market.
That starts by changing the question. Instead of asking, “Which technology should we bet on?” leaders should ask, “Which business outcomes matter enough for us to commit investment behind them?”
Technologies will continue to change. A clear business problem gives an organization something more durable around which to make decisions.
Fewer bets don’t mean less flexibility.
Strategic restraint can sound like an argument for moving too cautiously. It isn’t.
Organizations should move quickly when they have conviction. What they should avoid is confusing activity with progress. Making fewer bets means concentrating resources behind the opportunities where AI can create meaningful business value, while preserving flexibility as capabilities evolve. Leaders need to know where commitment matters and where flexibility matters.
That’s also why the ecosystem surrounding an AI investment is increasingly important. Organizations aren’t just choosing technologies; they’re choosing ecosystems. Every AI investment connects to a broader network of cloud providers, software platforms, implementation partners and data environments. Leaders should consider how technologies work with their data, workflows, existing platforms and partners, and whether that environment gives them room to evolve.
Build a discipline for choosing.
As the AI market gets more crowded, organizations need a repeatable way to separate an interesting possibility from an investment worth scaling.
I would start with four questions:
• Does this solve a business problem that matters? A compelling technology demonstration isn’t enough. Leaders should connect an investment to an outcome the organization values.
• Can we meet this need by expanding technology that’s already in our environment? Sometimes the better choice isn’t adding something new, but getting more value from what the organization already has.
• Are we prepared to scale it if it works? Before launching another pilot, ask whether the organization has the data, ownership, resources and willingness to change how work gets done if the experiment succeeds.
• What would cause us to double down, change course or stop? Define those conditions before momentum and sunk costs make the decision harder.
These questions won’t eliminate uncertainty. They’re not supposed to. They create discipline around making decisions despite it.
Intentionality is becoming an AI advantage.
AI leaders have spent the past few years learning how to experiment. The next challenge is learning how to choose.
The market will continue producing new models, new agents and new capabilities. There will always be another promising technology around the corner. Organizations can’t wait for that market to settle before deciding where they want to create value.
The leaders who succeed will give their teams permission to explore, but they’ll also know when exploration has produced enough information to make a choice. They’ll concentrate behind the opportunities that matter, stop investing in the ones that don’t and leave enough flexibility to adjust as the technology evolves.
In an era of nearly unlimited technological choice, knowing what not to chase may become just as important as knowing what to pursue.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
