From Experiments to Impact with AI Agents

Despite £24–32 billion invested in enterprise AI, 95% of organisations are seeing zero measurable return. That's the central finding of MIT's 2025 State of AI in Business report.* Just 5% of AI pilots manage to extract real value, whilst the overwhelming majority stall in the proof-of-concept fiction.

It's a staggering divide. Enterprises are keen to adopt AI. Ninety per cent have explored solutions and 80% have piloted them. Yet only a fraction manage to scale. The reasons aren't lack of talent, infrastructure, or even model quality. The true barrier is what MIT calls the GenAI Divide: tools that don't learn, don't adapt, and don't integrate into real workflows.

So whilst customer expectations for instant, personalised, seamless service accelerate, most AI initiatives remain trapped as demos, dashboards, or costly science projects. The result is straightforward: businesses spend, customers wait, and competitive advantage slips away.

From Reactive to Proactive: the shift that matters

Traditional support models don't scale. Headcount grows with volume. Budgets balloon whilst satisfaction drops. Support remains reactive instead of transformative.

AI Agents change this equation. Unlike static chatbots or brittle enterprise builds, AI Agents scale instantly without overhead. They don't just react—they advise, resolve, and continuously improve. They bring persistence, contextual memory, and adaptability. These are precisely the traits MIT identifies as the difference between failed pilots and successful implementations.

This is more than a technology shift—it's a business model shift. Human agents are firefighting problems. AI Agents are proactively solving them and elevating the customer experience into a source of loyalty.

Three Steps to Real Business Outcomes

Success with AI Agents doesn't require boiling the ocean. It requires clarity, focus, and staged growth.

1. Dream it

• Define what success looks like
• Set measurable outcomes tied to business objectives
• Align leadership and stakeholders on the "why"

2. Build it

• Identify "super moments" where AI Agents create disproportionate impact
• Launch an MVP within months, not years
• Prioritise business-first design choices

3. Scale it

• Expand into new channels and use cases
• Enhance capabilities to keep pace with evolving needs
• Augment the workforce by automating routine tasks and freeing humans for higher-value work

This phased approach moves organisations from experimentation to sustainable transformation.

Why now? Because the window is closing

The MIT report warns that the window to cross the GenAI Divide is narrowing fast. Enterprises are locking in learning-capable tools and vendor relationships that will be nearly impossible to unwind within 18 months.

The winners are already seeing results:

• £1.6 to £8 million in annual savings from reduced BPO contracts and outsourced processing
• 30% lower agency spend in content and creative
• 40% faster lead qualification and 10% better customer retention through intelligent follow-up

AI Agents don't just promise efficiency—they deliver measurable ROI. And they can be built in as little as two months, with new use cases live in just two weeks.

The message for business leaders is clear. The question is no longer if AI Agents will redefine customer experience, but how quickly you'll move to capture the advantage.

Sources:

• Challapally, A., Pease, C., Raskar, R., Chari, P. The GenAI Divide: State of AI in Business 2025. MIT NANDA. July 2025.

Find out your AI maturity

Curious about where your organisation stands on the journey towards the Agentic Era?

Take our AI Agent Maturity Assessment to evaluate your current level across strategy, data, technology, adoption, people, and governance. In just a few minutes, you’ll gain valuable insights into your strengths, development areas, and the concrete next steps to unlock the full potential of AI Agents in customer experience and beyond.