National Australia Bank has deployed an internal AI tool called Ruby — built on Anthropic's platform — to all 40,000 of its staff, while simultaneously planning to cut 2,500 roles by September 2027. CEO Andrew Irvine announced both initiatives within weeks of each other, making NAB the starkest example yet of an Australian institution investing heavily in AI and reducing headcount at the same time.
This isn't a pilot tucked away in a digital lab. It's a full-workforce deployment at one of Australia's four largest banks, running alongside a restructure that will eliminate one in every sixteen positions. The tension between those two announcements tells a story that every Australian business will eventually have to navigate — and NAB is writing the playbook in real time.
What NAB actually built
Ruby is NAB's branded internal AI tool, built on Anthropic's platform and designed for personal productivity across the organisation. CEO Andrew Irvine revealed Ruby during a media call accompanying the bank's FY26 results on 5 November 2026.
"We have now deployed an AI solution that we've branded Ruby to our internal colleagues," Irvine said. "Over 40,000 of our colleagues will have access to Ruby, and that's built on an Anthropic platform that we're very excited about."
Beyond the personal productivity tool, NAB has assembled roughly 50 "business-led squads" — cross-divisional teams working with strategic partners including Anthropic and AWS — to pursue specific AI use cases. These squads were all created within the past twelve months and are aligned to three strategic priorities: growing business banking, driving deposit growth, and strengthening proprietary home lending.
Irvine pointed to business bankers writing credit memos as a practical example. "We're leaning in hard and deploying capability there," he said. "A good example would be helping our business bankers write credit memos and speed up that process very, very significantly."
The results NAB is claiming are percentage-based rather than dollar-denominated — a notable contrast to its Big Four peers. According to its FY26 results, agentic AI has delivered a 20 per cent reduction in the time needed to approve business loans and an 18 per cent improvement in the bank's change delivery cycle. The broader transformation program delivered $463 million in productivity benefits against $1.8 billion in technology investment spend.
The other announcement: 2,500 jobs gone
Two weeks before unveiling Ruby, Irvine told the ASX that NAB would reduce its internal workforce by up to 2,500 roles — roughly 6 per cent — during FY27, with 1,500 of those cuts landing before Christmas 2026. The bank also flagged it would "materially reduce its use of consultants and external contractors" and recognised a $300 million pre-tax restructuring charge.
NAB framed the cuts as operational simplification: "reducing duplication and speeding up decision-making." Irvine insisted there would be "no material impact to customer-facing banking roles."
But the Australian Financial Review reported that the cuts are concentrated in technology and AI teams, with approximately 3,200 positions being closed and staff forced to reapply in a competitive process. Around 180 roles are also being offshored to NAB offices in Vietnam and India.
The Finance Sector Union was unsparing. "NAB's public statement today was nothing but corporate spin but its intention is clear — NAB sees its workers as a barrier to profits and a problem that can be swept aside with ruthless cuts," said FSU National Assistant Secretary Nicole McPherson. The union noted that management "refused to confirm whether the cuts were due to AI" when questioned directly.
Labor Senator Tony Sheldon went further, telling a parliamentary select committee that NAB workers were being given "the AI chop" and criticising the bank's lack of transparency about the role AI plays in its restructuring.
How NAB compares to the other Big Four
NAB is not alone — every major Australian bank is racing to embed AI across its operations. But the approaches differ in telling ways.
Commonwealth Bank is furthest ahead by most measures. Around 80 per cent of CBA staff actively use AI platforms including ChatGPT Enterprise and Copilot, 8,000 engineers use AI coding tools daily, and the bank has quantified AI-generated benefits at $200 million in gross value during FY26 — with expectations to double that in FY27. CBA has also invested $90 million in employee AI upskilling and built a centralised AI platform with partnerships across OpenAI, Anthropic, and AWS.
Westpac reports that 95 per cent of employees use AI tools monthly, with 72 per cent integrating AI into workflows. The bank has deployed agentic AI for mortgage assessments that saved 12,500 banker hours in its first month alone — and, as we reported previously, large Australian enterprises are increasingly measuring AI value in hours reclaimed.
The sector-wide pattern is unmistakable. As the ABC reported, more than 7,000 banking jobs have been cut across the Big Four since last year. Swinburne University governance expert Helen Bird told the ABC that "when we think about the scale of the arrangements they're planning, it definitely has a look about it as though AI is replacing staff."
The question NAB won't answer
The most conspicuous gap in NAB's narrative is the one the Finance Sector Union keeps probing: is AI causing these job cuts?
NAB says no — or at least, won't say yes. The restructure announcement framed everything around "simplification" and "alignment." Australian Banking Association CEO Simon Birmingham told the ABC that AI is not directly behind this round of losses, but acknowledged more workforce change is inevitable: "We can either allow it to have an impact where we don't adopt and change and we therefore don't see the productivity benefits, or it's an impact where we try to seize as much of the AI work for Australia."
This echoes a tension we've explored before. CSIRO research found that AI-adopting firms are actually hiring 36 per cent more workers, but that headline obscures what's happening inside specific roles. A bank can simultaneously hire more data scientists and cut more operations staff. The net number may go up while thousands of individual careers go sideways.
What's clear is that Australia's regulatory framework hasn't caught up. As we covered in detail, the Fair Work Act doesn't require employers to disclose when AI is a factor in restructuring decisions. Until it does, workers and unions are left to infer — and organisations like NAB face no obligation to connect the dots publicly, even when the timeline makes the connection hard to ignore.
What this means for your business
NAB's playbook offers practical lessons regardless of your scale.
The single-platform approach works. NAB didn't let 40,000 people choose their own AI tools. They built one approved platform, branded it, and rolled it out organisation-wide. This is exactly the pattern that Macquarie Bank followed with Google Gemini — and it solves the shadow-AI problem that trips up most organisations. If you're running a team of ten or a company of five hundred, the principle holds: pick a tool, sanction it, and make it the default.
Align AI projects to business priorities, not technology curiosity. NAB's 50 squads aren't exploring AI for its own sake — they're mapped to business banking, deposits, and home lending. The use cases that get funded are the ones tied to the metrics the CEO reports to the board. If you're building a business case for AI adoption, start with the KPI your board cares about and work backwards.
Be honest about the workforce question. NAB is taking heat not because it's cutting jobs — banks restructure constantly — but because it's deploying AI and cutting jobs in the same breath while insisting the two are unrelated. Whether or not AI is directly causing NAB's redundancies, the refusal to address the question openly is eroding trust with employees, unions, and parliament. Smaller organisations have an advantage here: you can be transparent in ways a publicly listed bank apparently cannot.
What to watch
NAB has committed to investing at a similar level in FY27 and expects underlying cost growth of less than 3 per cent. That's a signal: the productivity benefits from Ruby and the AI squads are expected to compound. Watch whether NAB follows CBA's lead in eventually quantifying AI returns in dollar terms — the pressure to do so will only intensify.
The parliamentary scrutiny is also worth tracking. The Joint Select Committee on Artificial Intelligence is actively questioning bank leaders, and APRA has already signalled that AI governance in financial services needs to improve. If regulators or legislators mandate disclosure on AI's role in workforce decisions, NAB's careful separation of its two announcements won't hold.
For now, Ruby is live, 50 squads are building, and 2,500 people are updating their resumes. The tension between those facts is Australia's AI story in miniature.
Sources
- NAB deploys AI for staff called 'Ruby' that is built on Anthropic — iTnews
- NAB simplifying operating model and improving alignment to customer priorities — NAB
- NAB FY26 Full Year Results Summary — NAB
- NAB H2 2026 Earnings Call Transcript — Stock Analysis
- NAB's despicable decision to axe 2500 jobs — Finance Sector Union
- NAB asks thousands to reapply for jobs as search for 2500 cuts begins — Australian Financial Review
- NAB's local workers given the 'AI chop', says Labor senator — Australian Financial Review
- Is AI behind job cuts at NAB and the big banks? — ABC News
- CBA Chief AI Officer outlines bank's AI journey — FST Media
- CBA bolsters AI investment as business lending sees major growth — FST Media
- Westpac Data, Digital and AI Update — Westpac
