Innovation isn’t a department. It’s a discipline.

For leaders of established businesses, innovation is often framed as a choice between protecting what already works and embracing what comes next. Based on more than two decades experience in highly regulated financial services and time leading in fast-scaling technology business, I’ve come to see it differently: innovation isn’t a function, a project or a single breakthrough, it’s a discipline of continually questioning whether the way you operate is still creating the value your customers need.

Legacy teaches you to see differently

I spent many years at Macquarie Bank, Nab and AMP before jumping into Culture Amp, a fast-growing tech company built from scratch. People often frame that as a leap between two entirely different worlds. In some ways it was. But it gave me a valuable vantage point of two worlds – legacy institutions on one side, and digital-native companies on the other. And it’s taught me that innovation isn’t what most executives think it is.

The popular myth is that innovation lives in flashes: the breakthrough product, the disruptive app, the moment you rewrite an industry. I don’t dismiss those moments, I’ve had a few of them, including a billion-dollar product I helped build early in my banking career. But I genuinely believe innovation is something far less glamorous.

Innovation is customer obsession, supported by great execution that delivers solutions that beat customer expectations.

It’s continuous, not episodic. The waterline keeps moving, what’s good today will be average and expected tomorrow. If you treat innovation as a project with a start and end date, you’re already behind.

What I’ve come to see more clearly is why large organisations struggle to keep up. It isn’t a lack of talent, budget, or ambition. It’s where their cognitive energy and capital actually flow.

I once sat on an investment committee overseeing a $1.7 billion annual budget, and watched most of it swallowed by compliance and legacy infrastructure before a dollar ever reached anything a customer would notice. The budget was staggering. What was left over for genuinely customer-facing innovation was not. Multiply that pattern across an economy full of hundred-year-old institutions carrying decades of accumulated tech debt and acquired systems that were never fully integrated, and you start to see why so many transformation projects fail to move revenue.

They’re solving problems that have proximity to the organisation, not problems that belong to the customer.

The wrong target for AI

This is where my worry about AI becomes urgent rather than theoretical. My fear isn’t that artificial intelligence (AI) will fail to deliver value, it’s that established businesses will point it at exactly the wrong problem. The default instinct is to treat AI as a back-office efficiency play: a tool for shrinking headcount and improving the cost-to-income ratio that boards and shareholders obsess over. I understand that instinct. But I believe that is missing the whole point.

Cutting costs might produce a cheaper product for customers, but the far more likely outcome is that the savings flow upward to shareholders while the harder, more valuable question: “How do we use this to build something customers can’t imagine living without?”, goes unasked.

There’s another mistake I see constantly, and it has to do with control rather than cost. A lot of organisations are holding back on rolling out AI tooling to their people, worried about expense, worried about security.

But it’s a bit like the teenage sex analogy: believing safety is in prohibition, rather than providing education and creating a safe environment. Employees are already using AI as consumers, on their own devices, with company data, whether or not leadership has “allowed” it. Believing there’s still a decision to be made about whether your team uses AI is its own kind of madness. The real risk isn’t that people will use it; it’s that they’ll keep using it in completely unprotected ways because no one gave them a safe alternative.

But handing someone a licence isn’t the same as making them capable. I fundamentally believe AI won’t take your job, but the person who knows how to use it will. Universities, schools, and organisations all have an obligation to make people genuinely literate and proficient users. The mistake is thinking that once the tool is deployed, the job is done. It isn’t.

If leadership stops at rollout, they won’t capture any of the value they were counting on.
And the human cost of getting this wrong is very real.

I think about what happened when call centres and hospital cleaning staff were “optimised” into scripted, timed, monitored roles: engagement collapsed, talent fled, and the informal judgement that once caught problems early; a cleaner noticing something wrong with a patient; a rep sensing a customer’s real need, disappeared along with the autonomy that made it possible. Applied AI at scale with that same logic risks hollowing out purpose across entire workforces, even for the people who keep their jobs.

Not skilling your workforce in AI is frankly reckless. There is no future world where these skills are optional. For employment, AI skills will be the ticket to the game. For companies, having an AI skilled workforce is no guarantee of success, but it will be a condition of it.

The order of priorities matters

I’m not dismissing the use case for AI to deliver efficiency, and I’m not arguing that legacy is worthless. Sometimes decades of institutional memory is exactly the asset a younger competitor lacks.

What I’m arguing is about sequencing and priority: customer obsession has to sit at the top of the hierarchy, with cost discipline underneath it, not the other way around. At Culture Amp, our organising anxiety was “how could we be disrupted, how could we become obsolete in a customer’s mind.” Everything else, including the AI strategy question every board is wrestling with right now, followed from that.

I don’t think the businesses that win the next decade will be the ones with the cleverest algorithms or the leanest back office. I think they’ll be the ones willing to ask an unfashionable question before every strategic decision: does this actually make the customer’s life better, or does it just make our quarter look better?

For me, the answer hasn’t really changed over twenty-plus years, only the tools for solving it have.

Sally Bruce is an experienced executive with 35+ years of leadership across financial services and technology in Australia and globally. She has built a career spanning highly regulated financial services and fast-scaling technology, including 20 years at Macquarie Group followed by senior leadership roles at NAB, AMP and Culture Amp. Sally is currently a Non-Executive Director at Airwallex SVF and QBE Australia Pacific.