Welcome to the Agent Era.

Is this a bold statement of the future – or simply a glimpse of what’s already unfolding?

If we are at the start of the agent era – and I think we are – then it will solve one of the most difficult challenges in marketing today: shifting customer expectations.

Our research suggests that over 70% of customers expect brands to understand their unique needs and expectations. But it’s not just the demand that’s changing – it comes with a zero tolerance for irrelevance. Two out of three customers say they’ll switch brands if they feel underwhelmed – or unseen.

Relevance has become the currency of loyalty.

How do we earn more of it? Step forward, Agents. They have the potential to slip between customers and enterprise systems as a brand-new connective layer, delivering on-demand, delightful experiences. They can deliver the relevance we require.

But, for them to do so, we need to trust them to be more than assistants. We need to slowly let them be decision-makers.

The Agent Opportunity

This opportunity arises because technology is bringing us into a new era. It’s an era where:

• Visual Interfaces
• Conversational AI
• Real-time personalisation

…converge.

When this happens, we can create dynamic, context-aware multi-modal experiences for our customers. These aren’t just personalised, they’re actively co-created with intelligent agents at the edge of interaction.

And the complexity we’re dealing with makes them a necessity.

The Complexity Challenge

To deliver personalisation at scale, we need to deal with:

• Behaviour streams and real-time context
• Predictive signals
• Individual-level histories

These make for decision trees that are non-linear, non-intuitive. At scale, humans struggle to define the optimal path for every customer in this sort of environment, however much work we put into those trees in advance. At a certain scale, it becomes impossible.

That’s where the autonomous, experience-composing Agents come into play. They offer the opportunity to provide real-time, hyper-personalised experiences – if we can trust them to do so. And that trust will be a challenge. We have to both be confident in the experiences they provide, and still be able to balance business and experience KPIs.

That will require a shift in both management and mindset.

Agents are autonomous of meaning

But the opportunity is clear. Agents have the ability to derive patterns from large data sets without getting caught up in the “why”. They recognise the pattern, they act on the pattern. They don’t need the reassurance that a human seeks of understanding why the customer is behaving this way.

Think of a car brand, which has a long-running relationship with a client whose purchases tend towards the luxury end of the market. But, all of a sudden, we see them browsing information about cheaper, lower-end cars. This behaviour doesn’t fit with their established history and presence, and traditional models of personalisation would struggle to cope. But agents, powered by AI, can spot this pattern of behaviour, and adapt to it, providing the experience the customer wants.

Now, it turns out that the customer was browsing for a car to give their daughter as a graduation gift. However, the agent is able to provide the needed experience without the need for the “why”. It’s making the connection that says that the visible pattern is more important, at this moment, than the established customer history. And letting go of the “why” – that feeling of control – is a critical step in making the most of agents.

So, what do we need instead? We need conditions and brand guardrails.

From MarTech stack to Agentic Ecosystem

This model means some fundamental changes to how we think about our marketing:

• From rules-based flows to self-optimisation agents
• From static segmentation to real-time learning
• From channel to composable, adaptive layers
• From functional workflows to agentic orchestration
• From data to knowledge management

This shift could well redefine the core architecture of marketing towards data and agent ecosystems, combining to create real-time composition of personalised experiences and dynamic content creation.

In other words, experience is no longer deployed; it is composed.

Agents in action

How does this feel in action? Our proud father comes to his familiar site, looking to buy a car for his daughter. The agent spots the change in behaviour in real-time, creating a reasonable interpretation of his intent. It asepses the right visuals, tailors the message and tone, and adapts in real-time based on the implicit feedback from the customer.

Result? A delightful experience for our regular customer, a thrilled daughter – and both a sale and an enhanced customer relationship for us.

This question, right now, isn’t “can we build this?”. We can. The real question is, “Are we ready to lead through it?”

---

Daniel Gonsior is a managing director and CTO of marketing technology at Accenture Song. This Insight is derived from the work and insights of the author and three colleagues: Munayan Ray, Sierk Conradi and Tilman Schlitt.

---

Image by khemrajdotin | pixabay