The Prompt: Are your brand guidelines ready to deliver behaviour?
All our current clients are trying to figure out how AI fits into their brand. Some are adding chatbots. Others use generative tools to quickly create campaign materials. A few have hired heads of AI and transformation and are working out what this means for their customer experience.
At The Frameworks, we’re asking ourselves the question: when an AI interacts with your customers, how does it behave – and how does that reflect on you??
It’s the question that led me to Sandra Peat.
Peat is a brand strategist with impressive experience. She led brand at Red Bull, was a strategy partner at JKR, and is now working on a PhD about “brand behaviour”: the repeated actions of an organisation, including what it says, how it says it, and what it does.
She started her PhD because, when she looked for academic research on how brands build meaning through their actions over time, she found nothing. The idea just wasn’t in the literature.
“I kept seeing it in industry,” she tells me. “The actions that brands take – what they do – carry important meaning. But we spend so much time focused on communications, and particularly advertising. I started looking at academic research and realised it didn’t exist.”
Peat’s work is especially relevant now.The question of how brands behave has become much more complex, since the people shaping brand meaning are no longer just human.
The thousand interactions problem
Peat points to Michael Eisner’s idea that the Disney experience is shaped by a thousand small interactions, each adding or taking away meaning. That’s been true for years. What’s changed is the speed and scale, and the fact that more of these interactions are now handled by machines that haven’t been taught the brand’s behavioural expectations.
“We’re living in a world where the entities creating brand meaning are both human and non-human,” Peat says. “That process of iteration and meaning generation is almost out of our hands. It’s not about what the AI did. It’s about what behaviours we’ve trained into this AI so that it creates the meaning we wanted to create.”
This is the gap between the usual way we build brands and what’s needed now with AI. Our traditional brand strategy tools – like tone of voice, content strategy, visual style – are all about controlling what gets said. They’re great for that. But, as Peat points out, they were never meant to guide how a brand actually behaves.
“They guide a quite narrow part of what I would consider the experience that delivers meaning,” she says. “They don’t control brand actions. They don’t address inconsistency. They don’t tell you how to make decisions under pressure.”
For those of us who have built brand platforms and content strategies, this is a tough truth. She’s right. We write the words, build the frameworks, and make recommendations around how the brand should appear. But when a chatbot handles a frustrated customer late at night, or an AI agent creates social content, the tone-of-voice guide only goes so far in shaping the customer’s experience.
Working at the intersection of brand and AI
Peat shared a case study from Klarna, the payments company. When customers missed payments, Klarna tried using a clearly non-human, transactional AI for follow-ups. There was no fake empathy. No small talk. No pretending to be a person. Customers actually had a better experience with the brand than when a human called. The embarrassment of being chased by a person was worse than the efficiency of a machine doing the same job without judgment.
“Sometimes you don’t want your AI to be human,” Peat says. “Does this moment of experience require empathy with your consumer? Or is it better to be systematic and transactional in that moment?”
Australian telecom giant Telstra made a similar choice when moving customer service to AI, but set the boundary differently. For broadband issues, the AI takes over. But if someone calls to cancel a deceased partner’s phone contract, a human always handles it. No exceptions or decision trees. These are brand decisions, not just tech choices, and they need leadership.
What stood out to me in both examples is that these decisions aren’t covered in standard brand guidelines. Right now, no section in a brand platform tells you when your AI should stop acting like a person.
Behaviour in, brand out
So what should fill this gap? Peat admits she hasn’t fully figured out the format yet, which is understandable since she’s writing the book on it. But she’s clear that brand strategy needs a set of defined behaviours – pulled out from a values statement – guiding how the brand acts at every touchpoint, both human and non-human.
Communication is a fundamental part of brand behaviour, but it’s not the only one. Connecting those wider actions back to your communications makes the whole thing more powerful.”
That defining brand behaviour sits right where brand strategy meets AI. This conversation should happen before building the chatbot, writing prompts, or launching AI content. But in most companies today, brand thinking comes after the technology is already in place.
Peat has seen this happen first-hand. “They created the AI and then put their brand meaning layer over the top,” she says of one big project. “That’s fine for tone of voice. But are you missing an opportunity, from a brand behavioural point of view, to deliver a more unique brand experience?”
The volume trap
I asked Peat what she thinks is the biggest mistake B2B brands are making with AI right now.
“Just pumping out AI slop,” she says. “You do need a lot of content. But if you’re using AI to produce more content without defining the behaviour you want, you end up with slop. What’s your behavioural lens? What value does this create for your customer? What brand message does this convey?”
Most AI-generated content in B2B marketing today lacks a behavioural lens. There’s a brief, a prompt, and then it’s published. All these interactions and touchpoints, without clear intention, are shaping the wrong kind of brand meaning – whether companies realise it or not.
Peat’s research shows that the brands most likely to succeed are those that decide, before using new tools, what behaviours they want their brand to show over time, in every interaction – human or machine. This is a new kind of strategy for many of us, and it needs to start now.
The Prompt is a content series from The Frameworks exploring how senior B2B marketers are building brands in an AI world. If you’re wrestling with these questions, we’d like to hear from you.
