That, for me, is sparring with AI

In an opening video of my first subject at UNSW for the Graduate Certificate in AI Strategy Gordon Au-Yeung, Director at EY, raised the concept of ‘sparring with AI”. He explained this as the continued back and forth with AI to improve the outcome of what was prompted and returned. Effectively a feedback cycle where the feedback is adopted by AI. This really stuck with me.

Since then, I’ve been pondering the notion of sparring with AI. In fact, I sparred with AI while I drafted this very piece.

Most of what I have been exposed to in the AI world has been process automation, usually with “humans in the loop”: humans writing the rules and AI executing them, AI taking over parts of processes humans used to do, systems automating processes to limit clicks, or humans prompting back and forth to improve an output. And of course, AI acting on a predetermined decision framework and humans approving, as well as chatbots – masked as AI. That back and forth, along with agents, digital twins and other forms of human-in-the-loop AI, is incredibly useful. Absolutely. But is it AI and humans working together, or just automated hand-offs and word processing in a swankier set of 'programs'.

But sparring got me thinking. 

It’s a genius phrase. What and where are the real examples of AI and humans genuinely working together? And beyond that, can AI bring humans together to challenge in a systematic way and make better commercial decisions that improve margins, reduce costs and strengthen business performance?

Until recently, I hadn’t really encountered that. Or at least, not beyond the promise of it, or the examples above.

As a data-based sales and marketing practitioner, I’ve spent a lot of my career looking for patterns in data and trying to turn them into recommendations or decisions. We got some of the way there with BI, analytics and indeed machine learning, but ultimately most systems still show you the data, albeit in increasingly pretty formats, and leave you with the question: now what do I do about it?

And when you socialise the pretty pictures - sideways and up the line, people’s ideas are often based on anecdotes. The data can become merely a backdrop.It leaves decision-makers to find the golden data nuggets. Work out what they mean. Determine whether something is a trend or a one-off. Decide what lever to pull. Then hope pulling that lever doesn’t create an unintended negative butterfly effect somewhere else in the business.

In the end, there is still a lot of subjectivity.

Back in my education consulting days, when I ran the largest international student survey with Lauren, we had an enormous amount of data. Mining it to understand what we should recommend to clients was always a significant and labour-intensive exercise.Many whiteboards, many tools and many late nights.

Similarly, as a CEO, modelling scenarios (based on price changes, sales volume, staffing and expenditure) to build into a forecast was complex and required many iterations as invariably, I’d want to nuance whatever was presented back - maybe because I hadn’t been as clear, considered and complete as I could have been in the first place; or because another question emerged from the answer.

Either way, everyone was sent back to the drawing board. More time and resources. 

Throughout my career I have never experienced a budget or forecast cycle that met the deadline and was simultaneously robust.

I always thought: wouldn’t our time be better spent analysing the output, challenging the recommendation and working together to find a better commercial outcome, instead of plugging numbers into Excel or other systems and playing with formulas?

This is one of the things that attracted me to Divi.

Divi takes the data already sitting within a business and uses it to identify what you should do now and what you should consider doing next. It brings together 30 years of business logic and IP with predictive and agentic AI that understands your business, grows with it, and surfaces commercial recommendations for decisions that you can make.

But the interesting part for me is what happens next.

The AI brings a recommendation to the table. Humans bring customer knowledge, market experience, context and judgement. They challenge the recommendation, change the assumptions and strength-test different scenarios AI and humans both have a 'seat' at the table.

That, for me, is sparring.

In fact, it’s the knock-out.

Imagine a management meeting where you can test a live commercial scenario and see the implications while the people in the room, who know the business challenge it.

Changes in fuel prices, climate impacts on pricing, fluctuations caused by volatile market conditions - bird flu, war, etc or labour costs and how they affect the bottom line. Holding less stock. Changing a supplier. Changing invoice dates and immediately understanding how that flows through to cash flow.

Or a finance meeting where a scenario can be input and its impact on the forecast understood that day rather than later. 

Or building a budget without spending weeks assembling the mechanics before the real conversation can even begin.

That, to me, is where AI becomes much more powerful.

Not AI replacing humans, or “low-value” work. That is a separate rant entirely and a pet peeve of mine.

But AI that complements human judgement.

Not simply automating what we already do, but helping us ask better questions, challenge our own commercial biases and make better decisions together.

Perhaps, somewhat ironically, one of AI’s greatest opportunities in business will be its ability to bring humans back to the table.