The Data Front

From models to messages to agents, and what it means for your team

So far in this series, we’ve looked at how to get started with chat AI, which tools to use, and how to write prompts that bring back better results.

This time, I want to zoom out. Because the tools we’re all experimenting with today are just one wave in a longer story of AI. Understanding that story can make the present feel less overwhelming and the future more manageable.

You don’t need to be technical for this. You just need to see how the waves fit together, and why they matter for you and your team.

Wave 1: Predictive and modelling AI

The first wave of workplace AI was invisible to most people. It sat behind dashboards, scoring systems, and models that quietly powered business decisions.

Examples may include:

• Customer segmentation in retail: grouping shoppers by behaviour to target offers or tailor experiences
• Risk modelling in financial services: using probability to decide whether to approve a loan or flag potential fraud
• Forecasting in supply chains: predicting demand to keep stock levels balanced

In retail, managers were trained to understand why customer groups looked the way they did, even though the segmentation came from models they never touched. In financial services, colleagues worked with fraud scores they didn’t build, but had to explain.

That was the defining feature of this wave: AI shaped decisions with indirect interactions.

Wave 2: Generative AI

The second wave is where we are today, and it feels very different.

Generative AI brought AI out from behind the dashboards and into the hands of everyday users. Instead of quietly scoring or predicting, it generates. Text, images, code, audio — whatever you ask for.

This is the first time most business users have spoken to AI directly. And that shift has been huge.

A few examples I’ve seen recently:

• A leader drafting a first version of a team update before editing it themselves
• An analyst reframing their insight summary for a non-technical audience
• A business user uploading handwritten meeting notes and turning them into actions

These are not experiments. They are practical efficiency-saving uses: saving time, sharpening communication, and building confidence.

For enablement, this wave is about teaching people how to shape outputs, not just interpret them. That is why we spent two articles on prompting.

Wave 3: Agentic AI (emerging)

The third wave is forming now. Agentic AI moves beyond responding to prompts. It takes action.

Instead of waiting for your request, agents can:

• Read your meeting notes, draft a follow-up, and send it
• Move between tools, pulling information from SharePoint, creating a slide, and drafting an email
• Run in the background, managing scheduling or weekly reports

It is a step beyond co-pilot. AI starts to feel like a colleague you delegate to.

But with that comes questions of trust, governance, and oversight. What do you want an agent to do? What should remain human? For enablement, this is where clarity and boundaries matter most.

Why this matters for teams

Understanding these waves matters because:

• It cuts through hype: not everything labelled AI is new
• It builds confidence: if you have worked with segmentation or forecasting, you have already experienced AI in action
• It prepares you: you can focus on generative AI now, without rushing onto adopt agents before the guardrails, and your teams, are ready

Practical takeaways

• Ask yourself: what wave of AI am I really working with here?
• Leaders: set realistic expectations for what generative AI can and cannot do
• Business users: focus on simple wins like drafting, summarising, and reformatting
• Analysts: explore how generative tools can make insights more accessible
• Data teams: play the translator role, helping colleagues navigate models, outputs, and risks

AI is not arriving as one sudden leap. It is evolving in waves, and you have already lived through some of them.

The opportunity now is not simply to adopt the latest tools, but to enable people to use each wave well. Shape outputs with prompts today. Build awareness of governance and trust for tomorrow. And prepare your teams to embrace agents with confidence when the time is right.