Over the past three years, we have spoken with 7000 people about AI, in conferences and boardrooms, workshops and training sessions, from major enterprises to SMEs, across cities and regional areas throughout Australia and beyond.
And while the participants in these events could not be more different, there’s a pattern to the conversations that has become clear.
So what are the lessons we have learned from three years of working with AI?
Every organisation worries they’re behind the curve.
Whether a business has implemented Copilot for everyone, has put all their people through training, or has started experimenting with agents, people are nervous that they are not doing enough.
Part of that challenge is that AI is easy to procure, simple to use but hard to embed in operations, so having AI on hand isn’t the same as knowing how, where and when it is used in your business — or if it is creating value.
The reality is more reassuring: nobody has completely solved AI implementation. Even the companies that appear successful from the outside are navigating the same fundamental challenges internally.
‘Start with the use case’ is not as helpful as it sounds.
The question “what will we use it for?” can stall the roll-out of AI for months, and it is easy to understand why. For a moment, put yourself in the shoes of a business owner 140 years ago, wondering how on earth this new-fangled electricity would make a difference in their work. Or the 1940s, when a leader of IBM suggested the world would only need about five computers. Or the 1990s, when many businesses thought having a website was a bit of a fad. Sometimes technology arrives before the use case and that’s the case with AI. Getting started is more important than starting in the right place.
People are better adopters than business — but also more anxious.
While executives discuss transformation and innovation, employees hear a different message. They worry about redundancy and irrelevance. Middle management feels pressure from both above and below. This disconnect creates friction that can slow or stop AI initiatives before they properly begin. At the same time, individuals are your best AI users, and are probably employing it in a dozen ways you haven’t considered, just under the radar. Creating a culture where experimentation is valued, people share what they do on AI, and where you are committed to improving their work they do, rather than removing the human, is a wise starting point.
What delivers value?
Success stories start small but they can quickly add up for companies willing to build on early wins.
1. Start with the friction.
The most successful implementations consistently involve small teams where people have a genuine knowledge of how they get work done and the steps in their workflows that are most annoying or slow. Addressing friction points — whether that is copying and pasting data back and forth once a month for reporting or having to scour several platforms every time a customer has a question — means your people will welcome the inclusion of AI. They will also start to think of new pinch points where AI might be able to help.
2. Look for sideways innovation.
You will probably see the best results when you use AI to extend current capabilities incrementally rather than attempt a revolution, however that shouldn’t stop you innovating. Thinking alongside the square can open up opportunities to add on things you can’t currently do with your time and resources. Use AI to standardise your procurement terms, for example, but then also get it to create a procurement team playbook for negotiation. Use it to find errors and inconsistencies across your whole website, and then also get it to reword your messaging based on specific audience groups.
3. Put more time into training than building.
The enterprises extracting real value from AI aren’t necessarily those with the most sophisticated models, but the ones where people genuinely understand what AI is, how it fits in their daily work, and what they can and can’t do within your governance structure. Complex pilots are great, but individual activity can help shift the dial.
4. Ask the right questions.
For users, that is moving beyond ‘do this for me’ to ‘how do I do this’, and engaging with the answers. For organisations, it is shifting from ‘how do we implement AI’ to ‘how do we implement AI well’. We all expect AI to solve problems, but the greatest value comes from using AI to see those problems differently.
After three years and seven thousand conversations, we have learned a lot about how AI works — and how businesses work with AI. We’re optimistic, but there’s a long way to go.
If you would like to join us on this journey, let us know.
