Companies big and small are embracing a DIY approach to software.
With the help of AI, they’re building tools that they could otherwise purchase from large enterprise software vendors like Workday, SAP, and Salesforce.
West Monroe, a Chicago-based consultancy, wanted a program that could check payroll for errors and another that could surface useful insights for managers about its roughly 2,000 employees.
Buying those capabilities from a vendor would have cost the company more than $300,000 a year, Tanya Moore, its chief people officer, estimated. To avoid that expense, employees built the software themselves earlier this year using OpenAI’s ChatGPT and Codex, she said.
For employers seeking custom internal tools, the options used to be limited: Buy from a vendor or keep doing the work manually. Generative AI has created a third option.
While most companies aren’t replacing entire enterprise software platforms with in-house alternatives — the kind of change that prompted SaaSpocalypse fears at the start of the year — the strategy could dent the incremental revenue that enterprise software vendors earn from selling add-on features.
About a third of organizations decided against buying at least one software product or feature because they could build it using AI tools, according to a recent survey of 1,719 executives and managers from McKinsey.
The trend “is going to erode a bit of the walled garden,” said Kyle Lagunas, an HR-tech industry analyst. HR teams “can self-serve a lot more.”
Omega Venture Partners, a technology investment firm in Silicon Valley, put the potential threat in starker terms.
“The vendor’s most dangerous competitor may now be the customer,” it said in an August report on its website.
Even so, the software industry at large is expected to remain on strong footing. The research firm Gartner forecasts worldwide IT spending on software to grow nearly 16% this year, reaching $1.47 trillion.
A range of use cases
Companies are using AI to create internal tools for HR, sales, finance, and other areas of their businesses.
Last year, employees at Spotify built HR Bot, a program that serves as a personalized HR help desk. Drawing from the music company’s regularly updated employee handbook, it can tell staffers how many vacation days they have left, what Spotify’s parental leave policy is for their country, and answer other routine workplace questions.
If AI didn’t exist, Spotify might have sought external support to build it, said Anna Lundström, chief HR officer at the company, which has about 7,000 employees. She added that HR Bot, which debuted in November 2025, is widely used by staff and has freed up her team to focus on other tasks.
Courtesy of Spotify
Similarly, Twilio, a communication software provider with around 5,500 employees, built an internal AI tool last year called Jarvis for its salespeople that can listen to sales calls, analyze a rep’s performance, and serve as a personalized coach. A spokesman said it has since achieved about 80% adoption across the company’s sales teams and reduced the sales-deal lifecycle by 54% for users compared with non-users.
Twilio CEO Khozema Shipchandler told Business Insider that if a tool like Jarvis already exists or comes to market, Twilio “probably won’t use it, because we’ve already built our own.”
Risky business
The DIY movement is notable given that investors erased hundreds of billions of dollars from software companies’ market values earlier this year over concerns that AI could make it cheaper for businesses to build software than to buy it.
Yet maintaining even small homemade features over time can be challenging for companies. Someone has to fix bugs, roll out updates, respond to feature requests, and ensure the software continues working as more employees rely on it. There’s also the risk that an AI-powered tool will make costly mistakes, such as leaking staffers’ private information.
“If an agent skips a compliance step or approves pay outside your grid, you’re not dealing with a minor hallucination,” wrote Workday’s chief technology officer, Gabe Monroy, in a June blog post. “You’re dealing with an audit, a regulator, or a lawsuit.”
Enterprise vendors also offer many years of experience building HR software and the vast amounts of data they’ve collected. At a recent Goldman Sachs conference, SAP CEO Christian Klein argued that large language models aren’t enough to replace those advantages.
West Monroe’s Moore isn’t deterred. She said employees at the firm have so far built eight tools in lieu of vendor offerings. Each underwent several months of testing before launch, and they’re reviewed every few months to ensure they’re functioning properly and up to date.
Still, Moore said she recognizes that other businesses may not have the expertise or the stomach to go down the same path.
“There’s a lot riding on that accuracy,” she said.
For some companies, the potential for massive savings is worth the challenge.
Gray, a design and construction firm in Lexington, Kentucky, with about 2,300 employees, expects to avoid paying $1 million to a vendor by building a financial forecasting tool in-house starting later this year with Microsoft’s Copilot and Anthropic’s Claude.
That is how much Gray’s finance chief, Scott Parker, said the company already paid for a comparable piece of software a few years ago, which came with another $100,000 a year in service fees.
“We’re realizing the power of AI,” said Parker.
