What did reporting season teach us about software and AI?

Can software businesses survive the threat of AI and perhaps even benefit from it? That is the central question facing equity investors in the sector today, and one that the February reporting season began to meaningfully answer.

Monday, March 16th 2026, 10:54AM

by Mint Asset Management

By John Middleton

Nvidia, DeepSeek, ChatGPT and Anthropic (Claude) have dominated headlines over the last year, creating heightened volatility in equity markets. This has raised pointed questions about how artificial intelligence is set to affect the world of IT software.

Two holdings in the fund, Xero and WiseTech, sit squarely in the crosshairs of this debate, and we believe the February results season supports our conviction in both.

Markets don’t like uncertainty, and volatility is at an extreme

It is very difficult to address concerns about AI that are not yet affecting current earnings, and the February reporting season once again drove more beats than misses. The potential impact manifests outside the current forecasting horizon, in the terminal stage of valuation analysis. It is equally hard to refute a risk that cannot be seen or quantified with any confidence.

Markets do not like uncertainty, and current volatility means they tend to price in a worst case until evidence emerges to change their view. February saw the highest number of greater-than-10% moves on results day on the ASX200 in the last 20 years, many in the software industry.

To illustrate current market conditions, consider the week of 10 February. This single week saw 10% sector-wide declines across Insurance, Wealth Management, and Logistics on successive days — not individual stocks, but entire sectors including most of their constituents. While highly unusual and painful for investors, we believe this may have marked peak AI hysteria.

The launch of an AI tool from Insurify drove a sell-off in insurance names. Altruist launching a tax planning tool hit Wealth Management. Concerns about an autonomous logistics agent drove US logistics stocks down 10%. Negative AI sentiment arguably peaked with the Citrini Research “2028 Global Intelligence Crisis” report published on 22 February 2026, a provocative piece from an independent research firm arguing that AI would displace broad swathes of knowledge work within three years.

What does a software company need to survive the AI challenge?

The key question is whether anything can survive the AI threat. We believe the answer is yes, and that three characteristics determine which software businesses will endure and indeed thrive:

Control of / access to proprietary data. AI needs data to learn from. It is very good at ingesting vast quantities of information and structuring it into useful datasets, but it needs the data in the first place. While individuals and small businesses can open their books to AI, Xero and WiseTech hold data accrued over decades within closed systems and have no obligation to share it.
Trusted entity status and regulatory accountability. Accounting software works with banks and government bodies. CargoWise connects with these institutions as well as customs and biosecurity agencies in every country where freight forwarders operate. This adds complexity, but more importantly the software must be vetted and tested by the numerous institutions that use it before gaining API access. In the worst case this means AI faces years of bureaucracy before it could disrupt this mission-critical software. More likely, it will need to work with partners, meaningfully reducing the disruption threat and opening up the opportunity for pricing and/or material cost reduction.
Ecosystem dominance and ingrained complexity. As Xero has demonstrated in Australasia (and is still working toward in the US), when a single player dominates a market and becomes the glue holding it together, it becomes very hard to displace, particularly if it continues to improve its product and deliver enhanced customer value. WiseTech has achieved exactly this position, used by 24 of the top 25 largest global freight forwarders. Its acquisition of E2Open extends that reach further up the distribution curve. The software is so ingrained in day-to-day operations that the cost and risk of switching arguably outweighs the potential benefit.

We believe that software companies exhibiting these characteristics are well-insulated from AI disruption, at least in the near term. Both Xero and WiseTech share all three and new information from the February reporting season reinforces this view.

What Intuit’s Anthropic partnership reveals

Intuit’s decision to partner with Anthropic, offering access to Claude at currently no extra cost raises questions about the economics and pricing of TurboTax and its broader software suite. But we believe the more revealing aspect of this deal is what it tells us about the limits of AI disruption. The tie-up suggests that data ownership, trusted entity status, and ecosystem position have blocked Anthropic’s direct market entry and forced it to join forces with the incumbent instead.

The structured data of tax returns and accounting makes these areas theoretically attractive targets for AI, hence the near-halving of some accounting software valuations over the past year. Partnering with Anthropic should enable Intuit to improve its product and reduce development costs going forward. The benefit to Intuit is clear.

The more interesting question is why Anthropic chose to partner with Intuit at all. If filing a tax return were straightforward for AI, the rational move would be to disrupt the market directly. The fact that it did not points to the genuine complexity involved. Intuit has spent decades and billions of dollars ensuring its submitted returns are correct.

Who bears the liability for an incorrect return? Ultimately the individual or business filing it. Most taxpayers, outside of those with simple returns, are unlikely to accept that risk and it is this population that represents the vast majority of Intuit’s revenue.

Furthermore, Intuit’s platform holds decades of filed returns and small business accounts. While Anthropic can learn from individual filings, it does not have access to the accumulated learning that both Intuit and Xero have built over time. Neither is willing to share this data, limiting the depth of training available to any AI challenger.

Finally, tax and accounting software interfaces with multiple government and financial institutions through heavily vetted integration points. Open banking is gradually reducing some of these barriers, but the security implications of AI access remain unresolved.

It is also worth noting that on 5 March, ChatGPT announced it would no longer pursue its in-chat checkout tool for travel bookings. The explanation given was that while users were happy to research trips through ChatGPT, few were willing to book through the app, preferring established third-party portals. Operational difficulties around payment processing, tax collection, and inventory management were also cited. We believe this supports our broader thesis about the limits of AI displacement in transactions requiring trust, accountability, and institutional integration.

What WiseTech’s results revealed: significant scope for cost reduction

When WiseTech’s new CEO Zubin Appo presented results on 25 February, the standout was not the headline numbers (which were a little soft) but his commentary on AI-driven cost reduction in product development. So, while AI represents a disruption threat to software businesses broadly, for WiseTech’s it appears to be a meaningful internal opportunity.

Appoo announced plans to cut 2,000 jobs, approximately 30% of the workforce, or around 50% of the product development team over the next two years. He cited examples of development timelines compressed from six or seven months to a single day and noted that rolling out custom capabilities to a new country is now six to seven times faster than before. He subsequently clarified that these savings relate to product development and customer services, with further potential for savings in sales and other functions.

Equally important is the shift in WiseTech’s business model from per-user pricing to a transaction-based structure (CargoWise Value Packs). This model suits the AI era well. Rather than upselling new modules or seats, WiseTech can monetise its AI-enhanced capabilities through increased transaction volume. Customers, in turn, benefit from the productivity gains AI delivers to their own teams.

AI is a threat - but also an opportunity

Overall, we believe the market has been quick to price in the threat of artificial intelligence to software businesses. From our own experience it is clear that AI is having a material impact on how work is done, and product capability is advancing rapidly. That said, we believe the threat to pricing has now been largely reflected in software valuations, while recent news flow is beginning to show real limits to AI’s disruption particularly where a company controls its data and operates as a trusted, regulated entity.

What the market appears to be ascribing little value to is the cost and development benefit that AI can deliver to the incumbents themselves. WiseTech’s results are a clear example. Elevated volatility means markets are currently inclined to price in risk before evidence emerges to the contrary. We see their reluctance to ascribe value to the ongoing growth and cost benefits available to software businesses with high barriers to entry as an opportunity and Xero and WiseTech as well-placed to capture it.

John Middleton is the Portfolio Manager for the Mint Australasian Equity Fund and joint Portfolio Manager for the Mint Australasian Property Fund. The above article is intended to provide information and does not purport to give investment advice. Mint Asset Management is the issuer of the Mint Asset Management Funds. Download a copy of the Product Disclosure Statement here

Mint Asset Management is an independent investment management business based in Auckland, New Zealand. Mint Asset Management is the issuer of the Mint Asset Management Funds. Download a copy of the product disclosure statement at mintasset.co.nz

Tags: Mint Asset Management