Archive · 30 Aug 2026
AI Coding Agents Drive In-House Software Building Over Purchases
A McKinsey report says many companies are using AI tools to help their own employees build software.
About one in three organizations decided not to buy at least one software product because they thought they could make it themselves.
Technology companies were most likely to do this.
Healthcare organizations and professional services companies also reported high rates.
Companies are using AI in more parts of their businesses than before.
The number of organizations scaling AI across the whole company rose from 38% to 44%.
Larger companies were more likely than smaller ones to scale AI widely.
McKinsey said companies do not need to build everything themselves.
Instead, they should carefully decide what to build, what to buy, and where to work with partners.
Nearly 32% of surveyed organizations said they had forgone software purchases because AI coding agents could build products or features internally.
Technology companies reported the highest rate of in-house development at 41%, followed by healthcare at 39%.
AI adoption at enterprise scale rose to 44% of respondents, up from 38% previously.
The share of organizations using AI across at least three functions increased from 51% to 56%.
Organizations earning at least USD 1 billion annually were more likely to scale AI enterprise-wide than smaller organizations, at 54% versus one-third.
- Who
- Organizations surveyed in McKinsey’s report, including technology, healthcare, professional services, energy, financial, media, insurance, and public-sector organizations.
- What
- Organizations are increasingly using AI coding agents to build software internally instead of purchasing some software products or features.
- When
- August 30; the report compares AI adoption with the previous year.
- Where
- Across organizations and industries surveyed globally; the article is datelined New Delhi, India.
- Why
- AI coding agents are giving companies more ability to develop software internally, while organizations assess where to build, buy, or use external partnerships.
Build In-House
Buy or Partner
Software strategy
Build In-House
Organizations can use agentic coding tools and internal teams to develop software products or features instead of purchasing them.
Buy or Partner
McKinsey cautions that companies should not build everything themselves and should continue using partnerships where they are appropriate.
Control and capability
Build In-House
Building internally can give organizations greater ownership of their technology and change agendas.
Buy or Partner
External partners can remain important, with leading organizations engaging them while developing enough internal capability to integrate and scale solutions.
- Report
- McKinsey report titled “The state of AI in 2026”
- Organizations foregoing purchases
- 32% said they had declined to buy at least one software product or feature because it could be built internally with agentic coding tools.
- Leading sector
- Technology, at 41% reporting this shift.
- Enterprise-wide AI scaling
- 44% of respondents reported that AI was scaling across their organization, compared with 38% previously.
- AI across multiple functions
- Use across three or more functions rose from 51% to 56%.
- Large organizations
- 54% of organizations with at least USD 1 billion in annual revenue reported enterprise-wide AI scaling.
- Smaller organizations
- One-third of smaller organizations reported enterprise-wide AI scaling.