Wolf&Grayson EditorialChristopher Bannocks has spent his career designing a different arc for multi-year data strategies. The old arc is familiar in large organizations, where ambitious roadmaps demanded heavy upfront investment in foundations and promised value was deferred several years into the future. A former Chief Data, Analytics and AI Officer in global enterprises across financial services and consumer goods, Bannocks now advises boards and executive teams as a fractional data and AI leader.
The new arc Bannocks advocates inverts that logic, anchoring long-term ambition to near-term business outcomes so momentum, trust, and funding compound rather than erode. Data and AI programs stick when they are treated as execution of the business strategy.
“Most organizations have a business strategy,” Bannocks says. “The most important thing you can do is align with that. How does what you’re going to do in data enhance, support or accelerate the business strategy?”
That framing matters because it resets the scoreboard. The point is not data governance for its own sake, or higher data quality as an internal metric. The point is customer growth, pricing performance, risk control, retention, and revenue. Data is how the business gets there.
Data is the enabler, not the objective
In most companies, data is not the product but a supporting capability whose value is measured by how directly it advances commercial, operational, or customer outcomes. When data becomes the goal rather than the means, long-running programs tend to lose credibility with the business.
“Data governance isn’t the objective. Data quality isn’t the objective. Insights isn’t necessarily the objective,” he says. “The objective will be something more ethereal, like we want customer growth or we want better sell-through or we need to think about what we’re doing with pricing.”
That distinction shapes the program design. If the business strategy does not explicitly include data as an enabler, the initiative will be treated as optional, and optional work is the first to lose funding and attention.
“Data needs to be accepted as part of the business strategy, so when you look at the plan to deliver the business strategy, there should be a line in there around how data supports it,” Bannocks says. “If there isn’t, then you have to question how bought in the business is to your plans.”
The practical implication is that data leaders should spend less time perfecting enterprise roadmaps and more time building an unambiguous line of sight from data work to business outcomes.
Quick wins are visceral, not cosmetic
The trap in looking for quick wins is selecting wins that look impressive on a slide but take too long to land. “Problems that are too large end up not being quick wins,” Bannocks says. “They don’t create trust because they drag on and everybody thinks that data is really hard or difficult to do.”
Quick wins must feel immediate to the business and resolve something already bothering a decision-maker. “You need business leaders going, that was quick,” he says. “It doesn’t need to be easy, but it should be fast and it should solve a problem that is very present in someone’s mind.”
He points to a handful of practical ways leaders can generate quick wins early, focusing on moves that may look unglamorous but often deliver the highest leverage. The rule is that quick wins must deliver meaningful business value rather than cosmetic progress.
- Make data accessible. Too many organizations lock data behind over-engineered pipelines and scarce specialists. “Often data is locked up. Nobody knows how to get it out,” he says. Giving the business safe access to data so teams can explore and learn can shift perception in weeks.
- Move expertise closer to the work. Centralized data teams can become bottlenecks. Bannocks has seen rapid progress by embedding data capability into the business. Sometimes that means hiring someone for a business unit rather than a central function.
- Fix one number. Inconsistent reporting erodes credibility. “I get one number from there and I get one number from there and I never know what the real number is,” Bannocks says. Stabilizing a single metric across the enterprise can restore trust faster than building another dashboard.
Scaling AI requires adoption, then incremental data repair
When organizations talk about AI at scale, they often skip straight to platforms, models, and vendors. Bannocks starts with people. “One area is user adoption versus user fear, user understanding,” he says, describing the bridge leaders must build so teams can apply AI to their work.
But once a company aims beyond efficiency into revenue and customer advantage, the hard constraint returns: the underlying data. “Businesses underestimate the amount of work required to get their data in good shape,” Bannocks says, particularly when AI depends on a deep understanding of customers, products, and decisions.
The danger is reacting to that constraint with a sweeping data quality program that often ends in large budgets, multi-year timelines, and benefits postponed to the final act. “That’s when people get tired and switch off,” he says, describing the moment leaders are asked for “100 million over three years” with value arriving only at the end.
“Scale doesn’t all have to happen at once,” Bannocks says. “Scale can happen incrementally. I’m not trying to solve all data. I’m trying to save the very specific 20 fields I need to get this piece of capability out of the door. Then I’m going to solve the next 20 and the next 20.” The approach keeps AI delivery moving while steadily raising the organization’s data maturity.
The first 90 days are about alignment, not architecture
Bannocks argues that the early phase of a major data program is mostly human work. The technical plan matters, but it fails without executive agreement on what will, and will not, be prioritized.
“Alignment of executive leadership to the approach is the most important thing in the first 90 days,” he says. “You’re going to choose one or two things, which means you’re not going to choose eight or 10 other things.”
That is where friction emerges. Leaders must accept that their preferred initiative may not be first, not because it is unimportant, but because it may not yet have a business case that proves value quickly. “None of these things tend to be technical. They are mostly human related.”
He is equally clear about what the data leader should not do: decide what the business should care about. “It’s not your responsibility to define the priority of the business problems,” he says. That belongs to the CEO and business executives. The data leader’s job is to translate priorities into a sequence that can move quickly, identify what will go fast or slow, and avoid boiling the ocean.
The payoff is competitive advantage
The temptation to copy competitors is strong, especially when AI headlines make it seem like everyone is racing. “Are you trying to keep pace with that competitor or are you trying to leapfrog?” he says. “Keeping pace might take you 12 weeks, by which case they’ve got a new release out. You’re going to be chasing your tail.”
The point of multi-year data work, in his view, is not to replicate someone else’s tools. It is to build a durable advantage rooted in strategy, operating model, and responsible execution. That is also why governance cannot be an afterthought when AI starts to touch customers and revenue.
By establishing a clear set of principles for safe and responsible use, organizations can match AI ambition to risk appetite and make informed tradeoffs. “If you’re going to risk something, you need to have a return,” he says.
Follow Christopher Bannocks on LinkedIn or visit their website.
