bentleystrategy.com

Waiting on AI Isn’t Prudence. It’s Forfeited Growth.

AI doesn’t compete with your data and martech investments. It multiplies their return. A five-step playbook for making the business case to your CFO and building the learning environment that lets your team move while the technology is still evolving.

Your CEO wants to know your AI strategy. Your board wants results this year. Your competitors may already be learning faster than you are. And your CFO just raised an eyebrow at another request for your marketing data and technology budget.

That's the squeeze. Every senior marketing leader I talk to is standing in the same spot: expected to produce an AI strategy, held to this year's number, and asked to do it without the funding for the data and technology foundation that all of it requires.

If those decisions are made using short-term ROI assumptions, most AI initiatives won't survive prioritization. That's because AI and the data and technology it depends on are not incremental investments. They're long-term, interconnected capabilities that determine whether an organization can compete in the years ahead.

AI doesn't replace the foundation. It raises its value.

AI and marketing technology are usually framed as competing for the same budget: spend on AI, or spend on the data platform, the intelligence work, the systems that connect marketing and sales. Companies pick one over another, or put small commitments against each, which rarely amounts to visible impact. That misses the point. AI isn't another item on the list. It's the multiplier that sits on top of it, and that should change how the budget gets allocated.

For years, the return on customer data, segmentation, and engagement platforms was constrained by one thing: human capacity. Organizations could invest in better technology and still act on only a fraction of what it enabled because people had to extract the insights, design the campaigns, and deploy them.

AI raises that ceiling. The same data drives more decisions and experiments. The same engagement platforms support more personalized journeys without additional headcount. Generative AI lets teams produce the volume of content personalization requires.

The argument for your CFO isn't simply the ROI of AI itself. It's that AI increases the return potential of the data and technology investments underneath it.

CFO skepticism is legit

The skepticism CFOs carry toward marketing technology is warranted. Many have lived through expensive platforms that underdelivered, integrations that took years, and renewals that grew faster than revenue.

But it would be a mistake to evaluate AI in the same way as those earlier investments. Incremental technology improves existing workflows. Disruptive technology changes workflows, economics, and what scale even means, often before anyone can model the return with confidence.

Most failed marketing transformations aren't technology failures. They're system failures: fragmented data, disconnected tools, weak adoption, and unclear ownership. AI deployments make those weaknesses more visible and painful, not less. Until now, humans were the glue, providing context, connecting data between tools, and patching botched automations by hand. AI agents built on an incomplete, disconnected foundation will get bogged down or make the wrong calls, and your CFO will be right to be disillusioned by yet another marketing technology promise.

The cost of doing nothing

Here's what gets left out of almost every investment decision I've watched leaders make: the cost of the option they think is free.

Standing still feels safe. It shows up nowhere on the budget. There's no line item for the deals you didn't win because a competitor's sales team knew which accounts to call and yours didn't. But that opportunity cost is real, and compounding. While you wait, a competitor funds the foundation, puts AI on top of it, and starts out-prospecting you, reaching the right buyer earlier, with a sharper message, at lower cost per touch. They're learning faster on every cycle, and the gap widens each quarter. Falling behind on the data-and-AI curve doesn't cost a fixed amount; it costs more every period you lose ground.

A leader who can put that number on the table, even a defensible estimate of the revenue at risk from inaction, changes the conversation entirely. Now you're not asking the CFO to fund a cost. You're asking them to avoid a loss. Those are very different conversations.

"We'll wait until it's ready" is a familiar and costly mistake

There's a sentence I hear from smart, experienced leaders when the AI conversation gets concrete: "Let's wait until the technology matures." It sounds prudent. It is, in fact, one of the most reliably expensive decisions a business can make, and we have a rich history of disruptive innovation showing that late adoption is rarely a winning strategy.

In Innovation and Its Enemies, the late Harvard professor Calestous Juma tells the story of the natural ice trade. In the late nineteenth century, harvesting ice from frozen lakes was a serious business. At its peak it employed around 90,000 people and was capitalized at roughly $28 million, well over a billion dollars today. Then came mechanical refrigeration, and the early machines were genuinely bad: expensive, unreliable, prone to exploding. The ice men found the technology wanting and waited for it to prove itself. It did, fast. By 1914, factories were making more artificial ice than was harvested from any lake; by the 1930s, commercial ice harvesting was essentially extinct. The people who waited didn't get to adopt refrigeration on a comfortable timeline. They were displaced by the ones who adopted it while it was still messy.

While it's easy to criticize these choices in hindsight, the ice men weren't fools. Waiting felt rational, because the technology really was bad and the forces pushing them to wait were real. The same forces freeze marketing teams today: the tooling is half-formed, no one wants to bet on the model that's obsolete in a quarter, and the corporate gatekeepers would rather wait for a clear winner. But you can't wait those forces out. By the time they're settled, your window for competitive advantage has already closed. A leader's job is to build the environment that drives progress forward while the technology is still messy. I've seen this tension play out repeatedly in the large-scale technology transformations I've been part of.

Creating the learning environment

At a large national retailer, I was asked to accelerate AI adoption across our marketing workflows, and the richest opportunity was in the creative team. Demand for personalized content was growing faster than our team could support. Designers were spending hours stripping backgrounds out of product photography and dropping the products into different real-world settings, work generative AI could do in a fraction of the time. The team knew it, and still wasn't moving. It wasn't doubt about the technology; it was the fog around it: new tools shipped every week, the team was already underwater, and underneath sat the same stack of fears: the output won't be good enough, customers won't accept AI imagery, and IT, Security, and Procurement will tell us to wait.

So we built an environment that took the fear out of moving. We stood up a small task force of the designers and PMs who were genuinely curious. We worked with Procurement to trial tools on real projects, so the gatekeepers were partners instead of blockers. We added a peer-review step to catch anything that looked machine-made before a customer ever saw it. Finally, we set up a small experimentation budget with fast-track approval, so trying a new tool was a same-week decision, not a quarter-long one.

That let us test quickly, drop what didn't work, and roll out what did. It also allayed another fear: betting on the wrong model. Instead of standardizing on a single model that might become obsolete in months, we adopted a designer-friendly platform that connected to dozens of models simultaneously. That let teams choose the right model for each task while experiments produced something more valuable than opinions: real operating data. Work that took hours finished in minutes, and the time to produce creative assets fell by 75 to 85 percent.

We didn't wait for AI to be safe. We built the conditions to learn it while it was still messy.

Making the case: the Five-One Playbook

If you accept that building an AI-ready foundation is worth funding and that waiting is the more costly option, the job becomes winning the argument. This is my Five-One Playbook, five moves in one week, to build a compelling case.

1. Frame it as revenue and profit outcomes, not tools. Talk about the levers of revenue and profit growth (pipeline velocity, conversion, cross-sell, retention) and how each improves with better data and technology. Then put a number on the other side: the revenue at risk if you stand still for twelve months while competitors don't. Leave the platform names in the appendix; leadership funds outcomes, so lead with the outcomes you can commit to delivering.
2. Bring a blueprint, not a shopping list. A request for five tools reads as five separate bets. A reference architecture, showing the foundation, how the pieces connect, and where AI plugs in and why, reads as a plan your CFO can underwrite.
3. De-risk by piloting now, while it's messy. The antidote to "let's wait until it matures" is a small, funded proof of concept that starts your team learning today. You're not asking for a bet-the-company commitment. You're asking to understand this technology before it tips. The crowded vendor market works in your favor: many tools are inexpensive and consumption-based, so you start small and your spend scales only as usage grows, along with the outcomes it delivers.
4. Co-own the ask with IT and Finance. The fastest way to kill a marketing technology request is to make it look like the Marketing department wants toys. Walk in with your CIO and CFO already aligned on the architecture and the business case, and the question stops being "should Marketing get this" and becomes "here's how the company accelerates growth."
5. Name an owner for adoption. Most marketing technology underdelivers on its potential because of training and adoption issues. Adoption is often an afterthought, with no one accountable for integrating the new technology into daily workflows. Put one person in charge of adoption, not the vendor and not "the team," and make sure that person has the leadership support to solve the issues that will inevitably come up.

The winners of the AI race will emerge quicker than you think

The organizations that win the next five years are building foundations and learning systems while AI is still messy, not after it feels safe.

Most companies won't lose because they ignored AI entirely. They'll lose because they tried to manage a disruptive technology with incremental processes.

By the time the technology matures and the business case feels obvious, the advantage may already belong to competitors who spent years learning faster.

Don't be the ice trade.