Macquarie Bank has recovered 130,000 productivity hours in seven months after deploying Google's Gemini Enterprise across its entire 5,000-person workforce. Close to 80 percent of staff now use the tool daily — an adoption rate that dwarfs industry benchmarks and makes Macquarie arguably the most aggressive AI adopter in Australian financial services.
What makes this story worth unpacking isn't the number itself — 130,000 hours is impressive but ultimately just a metric. It's the adoption model. Macquarie operates under APRA regulation, Privacy Act obligations, and the full weight of Australian financial services compliance. If they can get 80 percent of a regulated workforce using AI daily within seven months, the "we're too regulated for this" objection that many Australian businesses lean on starts to look more like an excuse than a constraint.
What Macquarie Actually Did
Director of Product for Enterprise AI Enablement Shahnwaz Ali laid out the approach at Google Cloud Next '26 in Las Vegas. The bank didn't start with a restricted pilot for the tech team. It offered Gemini Enterprise to every employee, paired with structured training and an internal certification pathway.
The results of that bet were immediate. Ninety-nine percent of employees completed the bank's "Using Generative AI at Macquarie" training, and approximately 3,000 attended live Gemini Enterprise demos. Weeks after launch, an innovation lab hackathon produced hundreds of solutions — built not by engineers, but by subject-matter experts in legal, marketing, compliance, and sales.
"They now can focus on what matters the most to us, which is delivering exceptional digital experiences to our clients," Ali told the conference.
The bank developed two categories of AI agents. Personal Agents handle individual productivity — document summarisation, research, content drafting. Enterprise Agents tackle complex business workflows. Ali demonstrated two onstage: one supporting a 30-person team handling sensitive matters by collating information and suggesting appropriate response wording, and another helping the legal team assess exposure during incident response.
The Risk Team Move That Made It Work
Here's the insight most organisations miss. Macquarie anticipated that risk, legal, and compliance teams would be the natural friction point — and instead of navigating around them, they ran straight at the problem.
"We obviously anticipated challenges with our risk partners — so, the legal team, the risk team, the compliance team," Ali explained. "The key part was taking all of our risk partners onto the pilot, so they were not the people sitting at the side trying to review something. We said, 'You are our first set of people. We want to get you onto the tool.'"
The logic is sound. When compliance officers use the tool themselves, they develop an intuitive understanding of its capabilities and limitations. Their risk assessments become grounded in first-hand experience rather than theoretical concerns. As Ali put it: "When they got onto the tool, they got a better understanding of how they're going to risk review it, because for us in a regulated industry that's super critical."
This matters well beyond banking. APRA recently warned the financial sector that AI adoption is outpacing governance — a finding that echoes what we covered in APRA's formal governance letter to banks, insurers, and super funds. Macquarie's approach offers a direct counterargument: broad access and strong governance aren't opposing forces. They're complementary.
How Macquarie Compares to the Big Four
Macquarie isn't operating in a vacuum. Every major Australian bank is racing to deploy enterprise AI, and the competitive context is instructive.
ANZ has rolled out AI agents within a Salesforce Agentforce-powered CRM — the first Asia-Pacific deployment at scale — consolidating data from 20 systems and reportedly saving business bankers approximately one working month per year. CBA is building "virtual relationship managers" for its business bank and using generative AI to combat phone scams. Westpac partnered with Accenture to deploy AI agents that reduced complex code migration tasks from six days to one hour. NAB developed multi-agent solutions for personalised customer content and code migration over a year ago.
But none of the Big Four have publicly reported anything close to Macquarie's 80 percent daily adoption rate across all staff. That's the distinction. The Big Four are deploying AI in targeted functions. Macquarie went organisation-wide from day one.
Chief Data, Digital and AI Officer Ashwin Sinha framed the philosophy clearly: "If an AI initiative doesn't result in better features, a more seamless customer experience, or more reliable service for our customers, we question its value. Everything comes back to creating happier customers who want to engage with us more," he said in the bank's announcement.
What This Means for Australian Businesses
The broader data on enterprise AI adoption makes Macquarie's numbers look even more striking. According to McKinsey's 2025 State of AI report, only 31 percent of organisations globally have scaled AI programs beyond pilots. Deloitte's 2026 enterprise survey found that just 25 percent of respondents have moved 40 percent or more of AI pilots into production, and only 30 percent of organisations have redesigned key processes around AI.
Macquarie isn't just ahead of Australian peers. It's ahead of global benchmarks.
Three elements of their approach are transferable to businesses of any size. First, broad access rather than restricted pilots — scarcity encourages shadow IT, while broad access builds organisational buy-in. Second, certification that signals institutional commitment and builds internal confidence. Third, embedding risk stakeholders in the pilot itself, which converts potential blockers into advocates. As CSIRO research recently found, Australian firms that lean into AI adoption are hiring 36 percent more workers, not fewer — the fear of displacement is less warranted than the cost of inaction.
None of these moves require the technology budget of a major bank. They require organisational decisions more than procurement decisions.
What to Watch
Macquarie's next move will test whether this adoption model translates into competitive advantage. The bank launched Q, a customer-facing AI agent, in January 2026 — extending AI from internal productivity into direct customer interactions. That's a meaningfully different risk profile, and how Macquarie navigates it will signal whether the governance model they built internally scales to customer-facing applications.
The 130,000 hours figure will also face scrutiny. Productivity hours recovered is a useful metric, but it's an input measure, not an outcome measure. The real question is whether those hours translate into revenue growth, better customer retention, or faster product development. Macquarie's leadership has been explicit that everything comes back to customer outcomes — they'll need to show that connection over the coming quarters.
For Australian business owners watching from the sidelines, the lesson isn't that you need Gemini Enterprise specifically. It's that the organisations moving fastest on AI aren't the ones with the biggest budgets — they're the ones that solved the governance problem first.
Sources
- Macquarie Bank saves 130,000 hours in seven months of Gemini Enterprise use — iTnews
- Macquarie Bank Democratizes Agentic AI, Scaling Customer Innovation with Gemini Enterprise — Google Cloud
- Macquarie Bank recovers 130,000 hours with Gemini Enterprise — AIVY
- Banks including CBA and Westpac are racing to scale AI — Australian Financial Review
- Big four banks accelerate push into agentic AI — Capital Brief
- ANZ rolls out AI agents for business bankers — Computer Weekly
- Macquarie reimagines customer-focused banking with launch of Q — Macquarie Group
- State of AI in the Enterprise 2026 — Deloitte
