The AI benefit is now a line in the results pack.
CBA, Suncorp and IAG put AI numbers in their FY26 results. When the country's most regulated institutions report the benefit rather than the ambition, the argument shifts from whether the compliance case can be made to why credit is still waiting.
Hi folks,
CBA, Suncorp and IAG all put AI numbers in their FY26 results this month. When the most heavily regulated institutions in the country start reporting the benefit rather than the ambition, the question for everyone else stops being whether it works and becomes why credit is still waiting.
The one thing.
Australia's most regulated institutions have stopped piloting AI and started booking it
- CBA booked about A$200m of measured gross benefits from AI in FY26. They expect that to double in FY27. About 80% of staff were actively using ChatGPT Enterprise or Copilot at 30 June. (CBA FY26 results)
- Suncorp disclosed 15 AI assistants handling more than 3m customer interactions a year, and 3,900 staff-built productivity agents. (iTnews)
- IAG will spend about A$200m on AI and AI enablement in FY27, with more than 60% of staff using AI regularly. (iTnews)
What this means for a fund:
- Three of the country's most regulated organisations are past arguing whether the compliance case can be made and are reaping the rewards of their investment
- Note what CBA did with the payoff. Technology spend stays flat at A$2.4bn in FY27.
In the mix.
- No sovereign wealth or pension fund can show AI has made them money
- SPEC Research assessed 27 sovereign wealth funds and 26 pension funds. CPP Investments runs 2,100 Copilot licences at 73% active use.
- Why it matters: measure the workflow you changed, not the licences you bought. CPP's John Graham: "Will it help us make faster decisions? Yes. Will it help us make better decisions? TBD."
- Banks are cutting the analyst intake that feeds their AI teams
- Junior classes are down about two thirds while 62% of AI talent comes from that same pool, per McKinsey.
- Why it matters: the junior seat is where judgement gets built. Cut it without replacing the training path and the bill arrives in five years.
- More than one in three Australians have already asked AI about their super
- Why it matters: your investors are already using these tools on their own money. That sets the expectation of what a fund can answer, and how fast.
From our week.
- Levercon has been backed by Airtree, and Jake and I are joining their Frontier residency. Announcement here.
- Credit is well behind, and the reason is not appetite. Every fund we speak to has the same story underneath: systems bought over 10 to 15 years that do not talk to each other, held together by people who have quietly become the manual integration layer. Three things sit in the way, all solvable.
- IT and security, which resolves once you write down where the data sits, who sees it, and what the model can touch.
- Change management, which is easier here than almost anywhere, because nobody likes what they have got.
- Sunk cost, which softens once you accept you are adding a layer, not ripping anything out.
Levercon builds the AI operating system for credit funds: we connect your data, deploy agents across origination and monitoring, and run the repetitive work end to end.
Book a strategy call.
See where AI fits in your fund.
Walk through Levercon's custom AI operating system to see how it can benefit your fund.