There are three trends shaping artificial intelligence (AI) driven business transformation – a shift to decision intelligence, agentic AI enhancing agency and composite AI making automation smarter.
That’s according to Tom Harris, partner and chief commercial officer at Esher Advisory and former SVP of digital business transformation at Mastercard, speaking at the Digital Transformation Conference London.
Shifting to decision intelligence
Until now, businesses have tended to apply AI in a “scattershot” way without specific use cases or clear decision-making processes, Harris said. The shift to decision intelligence focuses on making AI adoption and use more precise. “You can do it (apply AI) in a thoughtful way – derisk it, add value and start unlocking growth.”
Applying decision intelligence:
Harris cited three examples of applying decision intelligence alongside AI in an organizational context.
- Decision support in patient diagnostics: Human makes the decision, using deep learning for anomaly detection and generative AI for alerting.
- Decision augmentation in customer service: Human and system make the decision, using a large language model (LLM) for sentiment analysis and a knowledge graph for actions.
- Decision automation in wind farm operations: System makes the decision – learning, predicting and optimizing.
“These are three very different use cases which require three very different approaches.”
Enhancing agency with agentic AI
The rise of agentic AI is driving enhanced agency around AI implementation, Harris said. “If you think about how humans operate, we are very high agency. We have a high level of understanding, we can be proactive, we can achieve complex goals in complex environments and be very autonomous.”
There’s still a gap between what a human can do and what technology can do. AI agents have the capacity to fill this gap, Harris said. Agents can be designed to solve specific tasks either on their own (e.g. implementing guardrails/parameters over how decisions are made) or made smarter by operating them in series (e.g. putting rules in place in relation to regulatory frameworks). You can also attach agents to a knowledge graph to make them smarter in a particular subject area or put planning/sequencing around them so they know what to do when, Harris said.
“By connecting these together into your tool set you’re able to deliver a more sophisticated outcome,” Harris added.
Join us at All Access: AI in Business Transformation 2025 to future-proof your business and accelerate growth with the right strategy and the latest technology
Composite AI – making automation smarter
Composite AI is about making automation even smarter, Harris said. “What I really think will be exciting going forward will be heuristic and physics-informed models.” These are AI models that can be trained and learn without the need to be constantly fed data.
“We’re starting to do this now. This gets us into the world of very sophisticated models. By putting all these things together, we start to eliminate hallucinations, the value from AI becomes more predictable and the growth you can generate becomes more material.”
Harris said these three factors are key to maximizing the value of AI:
- Value impact: 75 percent of top 500 global companies will apply decision intelligence by 2026 (Gartner).
- Actionable: 33 percent of software applications will incorporate agentic AI by 2028 (Gartner).
- Trustworthiness: Composite AI market will be worth US$10 billion in 2030 (MarketsandMarkets).
[inlinead-1]