AI changes more than how quickly software can be built. Cheaper production changes who can compete, tailored production changes what customers can make for themselves, and agents may change how people obtain insurance.
This forecast brings together four related developments: repricing and specification-led software, bespoke and just-in-time production, agent-first insurance, and a longer-term transformation toward social care. They can overlap; they are not an inevitable sequence with fixed dates.
Production economics
Repricing & Spec Software
Our working estimate is that software-production costs have fallen roughly tenfold and will continue falling. This is a Growth Lab estimate, rather than an independently measured industry-wide statistic. The competitive implication is that customers can build for themselves and new providers can enter with cheaper, faster offerings.
Specifications preserve intended behavior while implementations become cheaper and more replaceable. A useful specification captures the problem, behavior, constraints, and acceptance evidence. For a submission tool, that could mean saving and resuming a draft across sessions, keeping every field intact, and checking the reopened draft against the original.
Customers and disruptors do not need perfect feature parity to offer a viable alternative. ArtCraft’s open-source alternatives across Adobe workflows illustrate this pressure. Buddy reports verifying that enough functionality was available to make migration viable. That is a sufficiency threshold, not a completed migration or a claim of complete Adobe-suite parity.
ArtCraft and rea are separate examples. No primary evidence in the retained research links their development. Their repositories publish licensing and separate asset notices; the examples do not independently establish the tenfold estimate.
Customer fit and timing
Bespoke & JIT Software
When software can be produced for one customer when it is needed, tailored fit becomes less dependent on a large installed base. Customers can create software around their own circumstances. Disruptors can offer cheaper, faster, on-demand alternatives that challenge incumbent providers.
Buddy reports already seeing customers become competitors in the marketplace. That observation gives the forecast a practical question: which customer needs can now be met directly, and what value must an insurance-software provider continue to add?
Bespoke fit and just-in-time production reinforce the repricing pressure. The form of the product can change even when the underlying customer need remains. Useful responses begin with that need and test the full outcome, including quality and the cost of delivery.
An emerging insurance channel
Agent-first Software
Our strategic aspiration is to be the first software provider to support commercial P&C insurance on an AI-native platform like OpenAI. The intended experience is direct: a person brings their need and circumstances to an agent, and obtains the insurance outcome through that agent.
That experience requires insurance knowledge, authorized fulfillment, and dependable connections to insurance capabilities. The channel changes from navigating a provider’s application to expressing an intent that an agent carries through to a result.
The former Compass course cited OpenAI dots and SpaceXSI Grok Bot / Team Bots as examples of frontier-platform agents. Its October 12, 2026 source review described primary dots as rolling out, specialist organizational dots as preview or pilot, and Team Bots as a public beta. Those dated examples support an investigation of the channel; they do not establish that Growth Lab has delivered direct consumer insurance through it.
A longer-term thesis
Industry Transformation
In our longer-term forecast, the superintelligence (SI) economy expands reasoning, experimentation, and production enough to change the economics of scarcity. Outputs of that economy may handle insurance implicitly, changing the role of risk pooling and separate financial protection.
Insurance could dissolve or transform into social care: a system in which people help people. The purpose of care remains even if the industry’s present form changes.
This is a forecast of post-scarcity economics, not a deployed condition or a promised timetable. The useful question is what social care could become when more of the protection and recovery people need can be supplied directly.
Evidence and revision
Testing the forecast
Repricing and bespoke production describe competitive pressure from customers and disruptors. Agent-first insurance describes a strategic opportunity. Social care describes a longer-term forecast. Keeping those distinctions clear makes the ideas easier to test.
Use real work to inspect the thesis: identify the customer need, state the assumption being tested, and look for evidence that would change the explanation. The Growth Lab Compass guides today’s work through Train, Target, and Execute; this forecast supplies a wider view of what may change.
Adapted from the former homepage forecast and the six-slide Compass course. External examples retain the context of the October 12, 2026 source review.
Return to the Growth Lab Compass →