AI coding tools are making some businesses reconsider whether to buy software or build a tailored alternative. A Forbes article reports that 32% of organisations in a McKinsey survey had passed on at least one software product or feature because they could build it in-house with agentic coding tools.
Watch Desk analysis
What happened
The survey covered 1,719 respondents across 97 countries. The article also describes Atlanta business owner Lamar Tyler’s team building an application to handle conference-ticket purchases and assignments that its existing CRM could not manage. Tyler says his business saves more than six figures a year through tools it no longer subscribes to and vendors it no longer hires. That is his account, not an independently audited saving.
The article cites Gartner’s forecast that up to $234 billion in enterprise application spending could be exposed through 2030, and Census Bureau research finding that 16% of AI-using firms replace equipment and software with AI. Those figures suggest a shift worth watching, not proof that every subscription is ripe for the chop. Read TerDawn DeBoe’s article in Forbes.
Why it matters
Building a small tool can mean less subscription spend and a closer fit with how a business actually works. But a subscription also comes with maintenance, security patches and someone else’s on-call rota. Custom software may save money at first and quietly hand the business a new, ongoing engineering job.
That trade-off matters especially for smaller firms. The article says AI-agent use is scaling faster among organisations with at least $1 billion in annual revenue than among smaller organisations, where the share remained flat in the cited survey. The tools may lower the barrier to building, but they do not automatically supply the time, skills or governance to keep what gets built working safely.
Our read
The useful question is not whether AI can produce a working first version. It is whether that version costs less to own over time than the software it replaces. Before cancelling a subscription, count the build, maintenance and checking time, and name the person responsible. If nobody owns the upkeep, the “cheap” option has merely hidden its invoice.
What to watch
- Whether small businesses can maintain AI-built tools without adding more cost than they remove.
- How adoption and software-replacement patterns differ between smaller and larger firms.
- Whether businesses set clear rules for approved tools, data use and responsibility before building.
Discussion spark: When a small business weighs building software with AI against paying for a subscription, should the deciding factor be the upfront saving or the long-term cost of maintaining and securing its own tool?
Sources and evidence
- Small Business Owners Are Renting Software That AI Could Build In Hours – Forbes (5 October 2026, 13:43 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.