Narrative to Numbers: AI’s Next Phase

• AI’s impact on software and information-services companies will be uneven.
• Companies with “systems of record” and trusted data may benefit as clients demand controls, audit trails, and accuracy.
• If major AI firms list publicly, better disclosure requirements could shift the debate from stories to hard numbers, creating more winners — and losers.

In February, we published two Strategist’s Corner pieces examining the impact of artificial intelligence on software and information services.

First, we highlighted how AI’s disruptive potential could erode pricing power and challenge the durability of returns in software. We then argued that these risks are not confined to software and are beginning to extend across industries where business models rely on information asymmetry, human expertise, or process friction.

In both pieces, our central message was the same: though markets have begun to discount these risks broadly, the ultimate impact of AI will likely be uneven. Disruption is not synonymous with destruction, and disruption is not equally distributed. While commoditization is a real risk and ranges of potential outcomes have widened, we believe that certain infrastructure and data-layer businesses are likely to experience greater demand for reasons that rhyme with Jevons Paradox — the idea that making a resource more efficient leads to more use, not less. 

We think suppliers of mission-critical systems of record offer a good example. These are not simply applications, but platforms embedded at the center of workflows, decision-making, compliance, and accountability. They often integrate proprietary, longitudinal datasets and serve as the trusted source of truth within an organization. As a result, the cost of failure — not just the cost of the software itself — becomes the binding constraint in customer decisions.

In this context, AI may reinforce rather than diminish an application’s value. As AI-generated outputs become embedded within workflows, the importance of data integrity, auditability and control frameworks rises. In other words, as AI lowers the cost of generating answers, it raises the value of knowing which answers can be trusted. This dynamic can deepen switching costs and further entrench systems of record at the core of enterprise operations.

Economic value may therefore accrue away from generic AI models and toward applications that integrate AI into workflows, leverage proprietary data, and deliver measurable productivity gains, creating both winners and losers across software and business services, and, in turn, both opportunity and risk.

Despite our persistent belief in the opportunities within AI, what we have not yet had is clarity. That may soon change.

A potential turning point: AI enters public markets

The prospective public listings of companies such as Anthropic and OpenAI could mark an important turning point, because investors may get more transparency into how these companies are creating value.

Today, as private companies, it is hard to see where revenue is really coming from, which use cases are truly making money, and how sustainable these profits are. Narratives are often shaped through selective disclosures, anecdotes, or early adoption stories, leaving investors to guess more than they can confirm.

Public listings would change that. Greater disclosure around customer segments, acquisition trends, and specific workflows could highlight where AI is replacing older tools versus simply helping people work faster. Just as importantly, public markets impose a level of accountability and consistency that private markets do not.

Said plainly: what is discussed in interviews or in podcasts will need to match what is reported in financial statements. If not, there will be consequences.

Why this matters: Moving from stories to proof

This transition is important because we believe it will move the conversation from speculation to verification. To date, markets have responded to AI-driven uncertainty with broad-based repricing, particularly across software, data, and information services. Implicit in this is a “guilty until proven innocent” framework — one that assumes competitive erosion before it is fully evidenced. 

AI does not affect all business models equally. Some may face real disruption, particularly where offerings are easily replicated or where switching costs are low. Others, especially those built on proprietary, longitudinal datasets or deeply-embedded workflows, may prove far more resilient, or in some cases, even strengthened.

As disclosure improves, investors should get a better view of market size, competition, and how durable profits really are. This may show that some risks are overstated, while other risks are bigger than the market currently assumes.

Companies with lasting advantages — through data, distribution, integration, or trust — may start to separate from firms whose value is easier to replicate in an AI-driven ecosystem. In that sense, this may be less about a whole sector being permanently damaged and more about profits shifting within that sector.

For investors, this distinction is critical. Periods of structural change often lead markets to extrapolate recent trends and apply broad discounts. But as uncertainty gives way to information, dispersion tends to follow.

Conclusion

If the first phase of AI has been characterized by market uncertainty and broad repricing, the next phase may be defined by transparency and differentiation.

Public listings by leading AI companies could accelerate that shift by giving investors a clearer view of where AI is creating value — and where it is not. In that future, the key question is not “Will AI disrupt industries?” but rather “How, and which companies benefit?”

We believe this is where active management can help most: using evidence to separate durable business models from those facing true AI-driven erosion.

Companies mentioned are for illustrative purposes only and may not be relied upon as investment advice or as an indication of trading intent on behalf of any MFS product. 

Keep in mind that all investments carry a certain amount of risk, including the possible loss of the principal amount invested.

The views expressed are those of the author(s) and are subject to change at any time. These views are for informational purposes only and should not be relied upon as a recommendation to purchase any security or as a solicitation or investment advice. No forecasts can be guaranteed. Past performance is no guarantee of future results.