Every AI tool in your martech stack should be either helping you save money or helping you make it. Operational AI saves money by automating workflows, reducing manual effort or cutting production costs. Go-to-market AI makes money by improving targeting, personalisation, conversion rates and customer value.
The problem lies in quantifying the commercial impact. In the conversations we’re having with marketing leaders, very few feel they can measure the value properly. Is AI-assisted content production saving two hours a week or ten? How much is AI-driven personalisation lifting conversion?
Given that every tiny element of marketing, and indeed business investment more broadly, is predicated on ROI, that seems like a major gap. The challenge is that the tools and the thinking for measuring AI’s commercial impact haven’t kept pace with the technology itself.
Proving ROI on marketing technology has never been straightforward. What’s different with AI is that the commercial models are still evolving – on both sides. Vendors are working out how to price something that’s fundamentally new, while the buyers are working out how to measure its value.
That’s not a criticism, it’s just where we are.
What AI are you using?
We’re still early in the AI journey. Many businesses are experimenting with the technology – sometimes because they’ve been told to from on high – but it soon becomes apparent that scaling from proof of concept to production is when things get hard.
Meanwhile, those that are adopting the more widely used platforms, such as Copilot, are only now discovering how it interacts with their existing martech to determine where it can create real efficiencies at scale.
Even the most tech savvy marketing teams are still working through the AI basics – deconstructing legacy processes, aligning data sources and redesigning how teams and technologies work together.
The trickiest part remains measurement though. How do you build a framework for evaluating AI use cases and deciding which to scale?
New capabilities require a bulletproof business case to secure the right level of investment, but funding will be contingent on identifying the commercial impact. All too often the marketing and IT teams have no idea what that will be until the use case is scaled up across the business.
You end up with a chicken and egg scenario. You don’t know how many AI credits you’ll have to buy from your vendor until you can review how many AI credits you had to buy last time.
How are vendors pricing AI?
Making the business case to the CFO is further complicated by AI pricing. Research shows that top SaaS companies averaged nearly four AI pricing changes in 2025 alone, and AI-related uplifts on renewals are landing between 20-37%. Meanwhile, two-thirds of IT leaders report unexpected charges from consumption-based AI pricing.
You might sign up for X credits per year, but then your actual consumption is based on how many tasks you run. You might then find that your AI tool comes in gold, silver and bronze tiers and is priced accordingly, from a fully autonomous AI agent down to one where you have to prompt everything manually.
The approach to AI agents varies too. Some vendors build and deploy agents for you, while others provide tools for you to build your own. The commercial implications of each are very different.
On top of this, pricing is only part of the picture. Most enterprises now have general-purpose AI tools such as ChatGPT, Claude and Copilot running alongside platform-specific agents embedded in their martech stack.
The natural instinct is to back the major platforms first and integrate them into existing tools. But as task-specific agents multiply inside every vendor product, many teams end up with both. That means very few have a clear view of where they overlap, where they complement each other, or what the combined cost looks like. It’s another dimension of the same challenges of commercial visibility.
How do you measure and compare costs accurately?
Understanding the commercial picture around AI usually starts with mapping the AI landscape across your stack, including the tools teams are using informally, and understanding the business impact of each use case.
This is where you need to lean on your vendors. They have the resources, benchmarks and case studies to help build your business case, and the good ones will welcome the conversation.
The process usually means looking at how AI supports your business strategy and prioritising the tools that accomplish specific goals. And as part of that, understanding the company’s appetite for risk. That might mean a phased approach, starting with discrete use cases within your existing stack to build confidence and demonstrate value before scaling.
But for many, the first step is working with their vendors to build the kind of business case a CFO will back; one that maps AI capabilities to measurable commercial outcomes. What is the outcome you want to see? For instance, do you want to save £1m a year through background automation, or do you need your chatbot to deliver a 5% increase in customer retention in the first ten months?
On top of that, factor in the other investment needed to make AI effective, because none of this is delivered in a vacuum. There might be extra staffing resources required, as well as costs for implementation and ongoing maintenance.
Adding AI isn’t just a technology decision, it has cost implications across operations, IT and marketing that are consistently underestimated. In almost every engagement we’ve worked on, this is the line item that most typically catches people off-guard.
What most organisations discover at this point is that AI doesn’t fix your foundations, it exposes them. The data gaps, undocumented processes, and disconnected systems that teams have been working around for years become impossible to ignore when you try to scale AI across them.
That’s not a failure of AI, it is the technology doing exactly what it should: showing you where the real work is. The businesses getting the most value out of the process are those that treat it as an opportunity to build the foundations, especially if there are elements they’ve been putting off.
AI is either saving you money or making you money. If you can’t say which, or how much, that is the first problem to solve. The good news is it’s highly solvable, and you don’t have to do it alone. Start the conversation with your vendors, get honest about your foundations, and measure what you can now. The rest will follow.
