Everyone wants breakthrough results; few are willing to take the risks that often are the price of such results. Innovation without discipline burns cash. A discipline that does not have imagination is simply getting in your way. The answer, we’ve long argued in Harvard Business Review, is a repeatable operating system for change — clear guardrails, fast feedback and a portfolio of bets. Five business executives demonstrate how to translate that theory into action.
Tie bets to revenue — and show by doing
David Walmsley, the chief digital and technology officer at Pandora, focuses on innovation that moves the top line, not merely on back-end polish. His mantra: Test fast, learn in public, and scale what works. That bias toward action generates momentum, while limiting risk.
The approach echoes patterns the BCG research uncovered among high-performing innovators: We have top-innovator leaders shifting funds to proven bets more often than those of laggards. For consumer businesses, a 1-2 point lift in conversion or average order value is a compounding of EBIT impact. Walmsley’s playbook: Set realistic boundaries, ship experiments and measure outcomes that matter to the CFO.
Engineer Reliability First, Then Push the Edge
Racing technology, says Steve Riley, who heads IT operations and service management at the Mercedes-AMG PETRONAS F1 Team, is not about the shiny first; it’s about the kind of performance that says 0% risk on the curve. But the margins that determine podiums often derive from experimental with new tools.
The reality for his team: autoclaves, machine tools and test rigs are now digitally linked and remotely controllable — an I.T. footprint that hardly existed a decade ago. Which makes IT an enabler of manufacturing innovation, not simply a helpdesk. Riley’s rule of thumb: treat infrastructure like a utility, but put your best people on the “value-add” projects that move race pace forward and manage their risks with explicit thresholds, runbooks and rollback plans.
Construct guardrails that speed up, not slow down
Innovation only scales once you have hardened the basics, says Kirsty Roth, who is chief operations and technology officer at Thomson Reuters. That’s modern cyber controls, privacy-by-design, and transparent AI ethics. On that foundation, risk becomes less of a brake and more of a safety harness.
Her approach is mapped to the NIST AI Risk Management Framework: establish a risk appetite, codify “red zones” (for example, sensitive data and model explainability), and have “safe-to-try” lanes where teams can — within the business risk appetite — experiment without waiting on endless approvals. Gartner research agrees: Companies embedding trust, risk, and security management into product teams are more likely to roll out more AI use cases beyond pilot stages.
Think of risk as information—learn in short loops
The London Interdisciplinary School’s CEO and co-founder Ed Fidoe is teaching leaders to learn from biology and systems thinking. Some spread risk among many small bets; others concentrate resources where evidence is strongest. The goal is not to get rid of risk but to try to turn it into feedback.
In practice, this means pre-mortems to surface modes of failure prior to launch, thin-slice MVPs to capture early signal, and explicit “kill or scale” gates accompanied by risk-adjusted ROI. Systems maps allow you to think through the ripple effects on customers, suppliers, regulators and even culture — so your innovation fits the ecosystem rather than rages against it.
Make learning a KPI — and name learning
For Nash Squared CEO Bev White, culture is paramount. Her organization’s Digital Leadership Report shows that approximately one in five AI initiatives are already operational at either mainstream or scaled levels—proving that you can’t build at pace without certainty of risk. The counterpoison is governance that constrains downside and a learning culture that compound upside.
White recommends establishing non-negotiable levers of control — data lineage, human-in-the-loop for sensitive use cases and clear ownership — as well as mechanisms that reward smart risk-taking. When projects fall short, teams should document their lessons, update playbooks and redeploy talent fast. The only way not to cash in is to make the same mistake twice.
A pragmatic playbook for change
Begin with a portfolio: reserve a share of budget for core upgrades, another portion for adjacent bets and a measured stake for transformational bets. If evidence, not politics, suggests we need to rebalance more often — quarterly, let’s say — so be it.
Codify guardrails: Publish your risk appetite, data boundaries, model governance, security standards in plain language. Empower teams with the knowledge of what is safe-to-try.
Instrument outcomes: track the leading indications of adoption (cycle time, escaped defects) and business results (revenue lift, cost-to-serve, risk reduction). Sunset initiatives that don’t move the needle you defined up front.
Close the loop: run blameless post-mortems, keep reusable patterns, replicate winners through shared platforms. As HBR also notes, transformation will win when learning no longer is an after effect but becomes the operating rhythm.
None of the leaders above romanticizes risk; none fears it. They turn into ideas into controlled experiments, they put up guardrails to go faster, and they let data — not dogma — decide the next bet. That’s how innovation compounds.