If your last innovation program drowned in reporting, the obvious fix is to strip the governance out. Experian tried that with its disruptive bets, and the overhead turned up somewhere else.
At the start of the year, Experian's CEO and board backed a new initiative to find horizon-three bets that could disrupt the company's own business. The question behind it was how durable that business really is, and whether Experian is moving fast enough.
Mike Charyk, a founding member of Experian's global innovation team, describes the company as one of "the classic incumbents," facing competitors that already use new technologies. Ten global tiger teams had eight weeks to make their case.
Their earlier approach to innovation had leaned so heavily on governance that reporting became the main focus. To avoid repeating that, the new initiative went to the opposite end: minimal reporting, open formats and informal reviews with the CEO.
Mike says they "swung the pendulum way too far in the other direction." In practice, the open setup moved the overhead to other places: unclear expectations, last-minute requests and a long wait for funding.
The experience taught Experian how to restructure its disruptive innovation work around four structural components: clear definitions, team design, clear outputs and reviews, and a fast decision gate.
1. Define what "big" and "fast" mean before teams start
This is the part Experian got right from the start. Its existing innovation system worked well, but mostly produced core innovation. To push teams toward bigger bets, the company gave the work a CEO mandate, which made it a visible priority the operating committee could ask about.
The mandate also came with clear definitions, so that, as Mike puts it, "when we say big (opportunities), people know what we mean." A big idea meant at least 250 million in annual run rate within five years. A fast path to market meant 12 to 18 months, compared with the three to four years Experian usually needs. Those two numbers gave teams and leaders one shared picture of what Experian was looking for.
Experian also set a minimum standard for evidence. Every team had to speak with customers or industry experts before pitching, on top of any desk research.
Experian split the initiative across 10 global teams, each focused on one area of the business. In the first sprint, every team explored what agentic AI could do for its area. Because the lines between areas were left open and nothing shared was defined upfront, each team designed its own full technology stack: tools to manage agents, workflows and governance, plus controls for metering agent use. As a result, Experian ended up with 10 separate AI stacks.
Only the top layer of each stack, the workflows and the market problem, was unique. The layers underneath, from agent management to governance and metering controls, were the same in every stack, built 10 times over.
For Mike, the 10 stacks were the clearest sign that something was missing. His advice: define the shared components before the first sprint and tell teams to assume they're already built. With the foundations covered, each team can "focus on the market problem and what you can do that's unique." Mike believes the teams would have moved faster that way.
3. Give every stage an output and every review a purpose
To keep governance light during the eight-week sprint, Experian replaced formal reviews with informal check-ins. The CEO met each team for 45 minutes to hear how things were going and ask what they needed from him. With a CEO in the room, though, Mike mentions, "you're going to be as buttoned up as you can” — so teams arrived with 40-page decks and clickable prototypes.
There’s nothing more formal than an informal CEO review.
By slide two, the CEO is already asking questions, and the rest of each deck stays closed. The same thing happened at a summit for the company's top 100 leaders. Experian wanted to give the teams visibility and set up their booths side by side. The teams treated it as a contest, and the polished demos turned the summit into what Mike calls "a bit of real innovation theater."
In both settings, teams were guessing what to bring and what the meeting was for. As Mike puts it, "what they wanted was clarity." Three changes give them that clarity:
Output templates for each stage. Experian now gives teams a plan for each two-week block, a picture of what a finished output looks like and the sections a deck should include. Teams know exactly what to hand in. Mike advises setting these up "right away" to "prevent all of that extra back and forth."
A stated purpose for every review. Sharing learnings, challenging each other and making a funding decision each call for a different setting; a room of peers behaves very differently from a room of the top 100 leaders.
A peer forum separate from senior showcases. Experian now runs smaller peer sessions where teams share early ideas and what they still need to learn. Mike believes these sessions would have caught the duplicate AI stacks much sooner if they had started earlier.
4. Build the decision gate and funding route before the work starts
Teams finished the eight-week sprint on time with their pitches ready. The decision process had yet to be designed, so the effort paused for six weeks of individual operating committee reviews, scoring and funding debates. The work slipped back to the company's normal pace. In Mike's words, teams "were sprinting, and then they were just waiting."
Mike's advice is to make sure "that gate itself is also designed for speed": have the funding mechanism ready before teams arrive, and make the decision on a fixed date against fixed criteria, so teams get an answer as soon as they reach the gate.
In the end, all 10 teams moved forward, each with 90 more days of runway. Mike sees this as a sign that the culture is still learning to stop ideas, and he is "optimistic that a few will stop for good reason" at the next checkpoint. For now, in his words, "the exploration continues."
Experian started out treating governance and speed as a trade-off: more of one meant less of the other. The sprint showed that the light version cost time too, just in different places.
What changed Mike's view was what the governance is for. When it's designed to give teams clarity rather than to produce reports, he says, "that's actually a beneficial thing."
