How to stop your AI agents acting like everyone else’s

Designing brand behaviour in an agentic world.

In the latest instalment of In Conversation, Isabella Speight, who recently joined as a Client Strategy Director, sits down with Ben Harding, Managing Director at the London-based design and brand experience studio NewTerritory, whose clients include Aeroméxico, Delta Air Lines, Mercedes-Benz, Coca-Cola, Ford, Bang & Olufsen and Panasonic

With deep brand experience expertise across sectors including rail, hospitality, automotive, aviation and lifestyle products, NewTerritory is recognised as a global design and brand experience studio, founded in London.

The conversation explores how you make experiences feel unmistakably on-brand when everyone has capable AI agents in customer-facing roles.

Isabella (IS): As AI agents play an increasingly important role in customer journeys, how do we stop these experiences drifting into a sea of sameness?

Ben (BH): AI agents are already becoming frontline representatives of brands. In service interactions, product support, customer care, onboarding. As organisations begin to embed agentic workflows into customer experience, the instinct is to optimise for accuracy, speed or knowledge depth.

We see that as a real risk for brands. As soon as organisations standardise on similar models, similar platforms, and similar orchestration patterns, experiences start to converge. Everything works but everything starts to feel the same.

This problem goes much deeper than ‘tone’. True differentiation comes from how an agent makes choices in real, unpredictable human moments in ways that reflect your brand’s character. Your agents must reflect your brand behaviour - rather than simply following the same script or a boilerplate set of rules that all your competitors are following.

IS: You talk about “brand behaviour” rather than brand voice. What’s the difference in practice?

BH: Traditional brand voice guidelines help words sound coherent. But when machines act autonomously, we must also better codify brand behaviour, much like we already do in well-designed service.

Take a stressed traveller whose flight has been cancelled. Two agentic customer service representatives may have the same polite empathetic tone, know the same policies - but one may enforce rules too rigidly. The other may actually act with empathy, flexibility, understanding, not only reassuring customers but changing its rules to suit the circumstances. The latter is heavily contextual, much like a human would act in this situation, and is likely therefore a better representation of a brand’s ethos.

We deal with the same issues in designing “human” service, but in that realm it is simpler as human EQ is so heavily at play. With AI agents it isn’t (yet!). This idea goes to the core of branding agentic systems: encoding a brand's service behaviours into the choices agents make, rather than just the words they use.

IS: Traditionally we’ve thought of brand experience leaders as those that succinctly orchestrate physical, digital and service expressions of brand. How does that evolve with agentic systems?

BH: When done well, branded AI agents become live, ‘always on’ representations of a brand’s service culture. Embedded in every interaction, every touchpoint, every resolution. Brands mustn’t oversimplify this or consumers will end up just dealing with the same generic LLM-based voice assistants dressed up in brand colours.

The organisations that translate their brand ethos into machine behaviour, and those that curate consistently well-designed service across both human and AI will stand out as the future brand experience leaders.

IS: Last question, everyone is wondering how to move their organisation into the agentic space right now – in your opinion, what do brand leaders need to do first?

BH: Brands must see AI agent transformations less as pure technology rollouts and more as design challenges that sit at the intersection of technology, brand and service.

Practically, the first step needs to be codifying brand-led service principles. Defining flexibility parameters, human-in-the-loop moments and how they recover trust when things go wrong. These codified service principles must become an integral part of a revised set of brand guidelines, not just something that sits within a set of technical guidelines.

The brands that win here will have been the ones that were able to evolve their experience without eroding the reason why customers choose them in the first place.

NewTerritory strengthens senior team with key hires

NewTerritory designs LATAM Airlines’ new amenity kit and bedding collection for premium business cabin

The Echo Chamber: Beyond surface mastery in design

How to stop your AI agents acting like everyone else’s

Designing brand behaviour in an agentic world.

Read more