If we’re being totally honest, most QSR CRM programmes run like a recipe.
Someone, at some point, sat down and wrote it out. If a customer signs up, send a welcome. If they haven’t visited in 30 days, send a re-engagement offer. If they’ve ordered five times, promote the tier benefit. If it’s their birthday, what the hell, add a candle to the push notification.
Good recipes stick around, and this one certainly has. It’s thoughtful, carefully tested and built from experience.
The problem is a recipe is just a list of ingredients and set steps on how to use them, and it doesn’t know who’s actually in the restaurant.
It doesn’t know that your “lapsed” customer just had a baby and their whole relationship with eating out has changed. It doesn’t know that your “high-value” member has been ordering for their team at work, not for themselves, and is about to change jobs. It doesn’t know that the person you’re about to send a win-back offer to visited twice last week, just through a different channel.
The recipe fires anyway. Right message, wrong moment. Right trigger, wrong person. Or more often: right person, right message, but a context the recipe was never designed to handle.
This is the ceiling of rules-based CRM, and most QSR brands are hitting it whether they know it or not.
“Agentic CRM” is a phrase that’s been doing the rounds for a while now, and I understand if it sounds like jargon dressed up as technology. Underneath it, though, the idea is genuinely simple. In CRM, agentic means the system has agency: it isn’t just executing instructions someone wrote in advance, it’s making the decisions itself. Which customers to reach. Which channel to use. Which message to send. What to offer, and when. All based on what’s actually happening with each person right now, not what a segment decided three months ago.
Rules-based CRM asks the question – what have we decided this customer should receive? Agentic CRM poses a different one – what does this customer actually need right now, and what’s the best way to give it to them?
The difference sounds subtle. In practice, it’s the difference between following a recipe and being a great cook. A great cook knows the fundamentals: flavour, technique, timing. But they also read the room. They taste as they go. They adjust for what and who they have in front of them, not what the recipe assumes.
A few things worth being clear about
Agentic CRM is an approach to decision-making, not a product you buy off a shelf. You can build it on top of the stack you already have, provided the data foundations are in place and the platforms you’re running support real-time decisioning.
It doesn’t replace your marketing team either. The decisions a well-configured agentic system makes are the ones your team currently makes manually, just at scale and in real time. The job shifts from managing workflows to setting strategy, defining constraints, and reviewing outcomes.
And it doesn’t fix broken data. If your customer data is fragmented, inconsistent, or stale, agentic systems will only accelerate bad decisions. The technology is only as good as what it runs on, which is exactly why the next section matters more than the first two.
What this looks like for a QSR team
Rules-based CRM is designed in advance: you define the journeys, set the conditions, build the logic, and monitor whether it’s performing. Agentic CRM is defined by outcomes and constraints, not pre-scripted paths. You tell the system what you’re trying to achieve, what you’re willing to invest to achieve it, and what the boundaries are, and the team’s job shifts to reviewing outcomes and refining the parameters.
In practice, that means instead of building a 12-step re-engagement journey, you define what re-engagement means commercially and let the system determine the best path for each customer based on their actual behaviour. Instead of segmenting by recency and frequency once a month, you have a live view of each customer’s value and risk. Instead of choosing channels campaign by campaign, the system learns which channel each customer actually responds to and routes accordingly.
The hard part is trusting the system enough to let it make decisions you used to make manually, and being rigorous enough about your constraints that it makes them well.
What it demands before you start
Before any of this works, your data has to be ready for it. Agentic CRM runs on signals: what a customer ordered, when they last visited, which channel they came through, whether they engaged with the last message you sent. Those signals need to be clean, consistent, and flowing in real time, because the system is making decisions based on what’s happening now, not what happened last Tuesday.
If your customer identifiers are fragmented across POS, app, and loyalty, the system is working with a partial picture. If your event data is inconsistent or delayed, the signals it’s reacting to are already stale. That’s not a reason to give up before you start. It’s a reason to start with the foundations, not the features.
The QSR brands that have made this shift most successfully treated data infrastructure as the first project, not the last. Unified customer view, stable event schema, reliable data flows, in that order, then the decisioning layer on top.
Where to begin
Pick one always-on journey that genuinely matters to the business. Re-engagement of lapsed high-value members is a good starting point: measurable, commercial, and complex enough to show the real difference between rules-based and agentic approaches.
Define what success looks like in revenue terms. Incremental visits. Incremental revenue. Acceptable cost per intervention. Run the agentic approach alongside your existing rules-based journey as a control and measure what actually changes, not just what looks good in a weekly deck. And invest the time in the data layer properly. The work you do now cleaning identifiers and instrumenting events is the work that makes everything else possible.
The gap between what’s possible and what most QSR brands are actually doing with their CRM right now is still wide, wide enough that the brands moving on this now will be in a different league by the time everyone else catches up.
The recipe is fine. Having a great cook is better.
We’ve put the practical detail on how to make that shift, the five-step method behind it, into the Agentic CRM Playbook.
