By Lucy Hogarth, Co-founder of The Marketing Centre.
I keep hearing the same question from founders: “If we add AI into sales and marketing, will it finally smooth things out?” I wish it worked as simply as that.
But the truth is, if sales and marketing are already pulling in different directions, AI amplifies that tension more quickly. If you’ve got shared goals, shared data and someone clearly in charge, AI can pull things together.
For years, we built teams around human limits – how much information people could process and how quickly. Translated into customer handling, capacity could only stretch so far before you needed more people. AI removes a lot of those limits, meaning fewer handovers and fewer ultra‑narrow tasks. It can also take a meaningful slice of discovery, research and early qualification before a human ever steps in, so you can use your people where judgment and context truly matter.
The problem is that most teams begin their AI journey with one isolated win: a rep drafts emails faster; a marketer drafts content; ops automates a spreadsheet. Handy, but nothing fundamental changes until data, decisions and day-to-day work sit on the same backbone. When you get that right, your CRM stops being a weekly reporting chore and starts acting like an intelligence layer the whole team actually uses. Outcomes improve because everyone is responding to the same reality, in near real time, instead of looking at competing versions of the truth.
Finding the real bottleneck
At The Marketing Centre’s AI Futures Forum, we bring together SME leaders, operators and technologists to talk about AI in practice, not in theory. A recent roundtable introduced a simple lens – a quick health check for how AI plugs into a business. It groups the usual sticking points into three pillars and asks you to fix the biggest one first.
The first is decisions and risk. Who owns AI informed decisions in your company? When the model is wrong, as it sometimes will be, what happens next? Which calls are reversible and which need tighter gates? If responsibility feels fuzzy or risks keep surfacing late, this is your bottleneck. The answer doesn’t need a new feature, rather clearer accountability and the right level of human oversight where risk is asymmetric.
The second is the revenue engine. This is the messy reality of how money is actually made. Do sales, marketing and customer data run as one system, or does information splinter at the edges? Are you chaining together point tools, or running a connected workflow from first touch to renewal? If activity is running but results aren’t moving, then the engine is misaligned. You don’t need more campaigns; you need shared inputs and a single source of truth that both humans and models act on.
The third pillar is data, AI and adaptation. How quickly does insight become a decision that sticks? Do promising experiments scale, or stall in meetings? Is your governance proportionate to the risk, or do you slow everything equally? The goal is to shrink the loop: test, learn, decide, scale. Then do it again. When this works, revenue gets more predictable, AI spend links cleanly to outcomes, and nasty surprises show up early when they’re still affordable to fix. The aim isn’t to make the process more complex; it’s to make it coherent.
Putting it into practice
Founders don’t need another grand plan. They need a way to start. We typically see this as a three-month runway that turns intent into an operating rhythm.
- Month one: Intent. Write a one-pager that says where AI will (and won’t) be used in your revenue engine this quarter. Pick a single shared metric, say, qualified pipeline by segment, and name the person who approves AI-assisted decisions and exceptions. Keep it to one page so people actually read it and remember it.
- Month two: Plumbing. Move from individual point tools to a common backbone so activity, outcomes and learning live together. Redesign two end-to-end workflows, so the same data drives both humans and models. Hold a short, weekly review to decide what you’ll scale next week based on what genuinely moved the metric.
- Month three: Scale. Turn the plays that worked into runbooks with thresholds, approvals and clear escalation. Shift team design away from narrow tasks towards owning a number from first touch to close for a defined segment. Track one meta metric: the time from signal to decision, to scaled change. If this process isn’t getting shorter, your blocker sits in one of the three pillars above. Time to go back and remove it.
The point of AI in the revenue conversation
AI isn’t here to replace your team. It’s here to remove the alibis: misalignment, fuzzy ownership, and the kind of procrastination that hides behind “we’ll revisit next quarter”. A tool cannot take responsibility simply because it can’t. Your team does, though.
The companies that win won’t be the most technical. They’ll be the ones who choose a cleaner structure, set clear responsibilities, and let AI amplify the good habits they already have.
