What we're seeing, Autumn 26. Updated quarterly.

Trend 3: How products get built and businesses run is changing

AI is changing how businesses are run and how digital products get built, and building digital products is likely the first place it significantly transforms an organisation.

Our hypothesis

AI is changing how businesses are run and how digital products get built, and building digital products is likely the first place it significantly transforms an organisation. Agents can take on much of the coordination work of meetings, status and planning, so teams can do more with the talent they already have and spend more of their time on the work that matters most to players. Software factories are now the main focus for the frontier teams, and the time and cost of a build is falling sharply, so the question for an operator moves from "which vendor do we buy from?" to "what do we build, what do we buy, and what do we run with agents?" We expect the technology to move faster than the people. How a business brings its people and its customers with it, and what it stands for, sets the pace, which makes this the most human and slowest-moving of our trends. The pace will be set by challengers with capital and no legacy, and value moves to what is hard to copy: licences, data, brand and the decisions made on them.

In iGaming it is early. Customer support looks likely to feel it first, and we think the trigger for incumbents may come from a challenger on the new front door (Trend 2).

What this means for you

For the CEO, CHRO and COO, this is major transformation territory at a pace businesses have not experienced before, and the limit is communication more than technology. The people who do the work today hold the know-how agents learn from, so transformation works best with them involved from the start, and with a clear, early plan for how roles grow and where their time goes next. Customers, your own people and regulators will all be watching how AI is used, so saying early what the business stands for builds trust. On the product side, writing down your product principles and reopening build versus buy comes first, and being clear about your IP and how you create value protects what challengers cannot copy. The right next step differs for every business, and we help teams work out theirs.

What we are seeing that we think is important

  • -Large companies are being redesigned around agents. In a video discussed in Mark Minevich's LinkedIn post, Eric Schmidt says AI platforms can take on much of the coordination work of meetings, status, planning and chasing, leaving people free to focus on building, selling, working with regulators and leading. Minevich adds a redesign with leaders' spans of control widening from 7 to 30 and explicit ownership contracts for small teams.
  • -Capital is flowing to challengers. The Block (24 Jun 2026) reports, citing the Financial Times, that Kalshi is in talks to raise at a $40B valuation, after a $1B Series F at $22B in May 2026 led by Coatue with Sequoia, Andreessen Horowitz and Morgan Stanley, and a $5B valuation at its Series D in October 2025. It cites The Information for annualised revenue above $2B, and its CEO has said an IPO is under consideration but not this year.
  • -A licensed bookmaker and casino built in under four weeks. On 26 Aug 2026 The Straight reported Tim Heath's claim that an AI-driven team of nine, six of them software engineers, took a platform from first code to its first live real-money wager in under four weeks, against 12 to 18 months for a typical build. Heath says "If software goes to zero, a company whose moat is software dies" and "Data is worthless until it becomes decisions".
  • -More companies are choosing to build. Forkast (1 Sep 2026), citing McKinsey's State of AI 2026, reports that 32% of organisations have decided against buying off-the-shelf software and are building with agentic coding tools, and that 40% of companies with over $1B revenue are scaling agents in at least one function, up from 27% a year earlier.
  • -What counts as your IP is becoming a board-level question. Nestor Dubnevych reports on LinkedIn (about 28 Oct 2026) rumours on X that BigLaw is moving to local models as law firms ask whether their institutional knowledge could leak through cloud AI. He notes that Latham & Watkins bought GPUs in September to fine-tune its own models.

What we're finding out next

  • -What triggers an incumbent to transform, and how far challengers move ahead before it does.
  • -How to bring the people who do the work today into the design of automation, and how to talk with them about it.
  • -Where it lands first in iGaming, and how customer support gains from handling higher-value queries.
  • -What a business operating system or ontology looks like for an operator, and how the frontier teams run their software factories.
  • -How an operator writes product principles an agent can build against.
  • -How licensing and compliance sign-off keep pace with a four-week build.