Small businesses are increasingly turning to artificial intelligence (AI)-powered coding assistants to slash their technology development expenses while maintaining competitive digital capabilities, according to industry practitioners.
AI coding assistants are tools based on large language models to generate code, fix bugs and more. They enable smaller teams to do tasks that would cost a lot more and require additional engineers. They also cut development time for businesses to bring products to market faster.
“The game has changed with the advent of AI coding tools, particularly for small firms,” Lei Gao, CTO of social engagement platform SleekFlow, told PYMNTS. “They don’t need an entire engineering team anymore because now they can save money and time too by developing apps, trying out new frameworks or automating procedures.”
That’s the experience of Mike Stone, co-founder of customer web and mobile development firm The Gnar Company, whose clients include the state of Massachusetts, Grubhub and AARP.
Stone described the old playbook as the following:
Hire five junior developers
Add two senior developers to mentor them
Hope they gel as a team.
Wait six months to a year for productivity gains
With AI coding tools, Stone said the playbook has changed to this:
Hire two exceptional senior developers
Equip them with AI tools
Watch them outperform entire teams
Ship product code from ‘Day One.’
“One senior developer with the right toolchain now delivers what used to take a small team,” Stone told PYMNTS.
Small businesses recognize the benefits of AI. According to a recent survey by the U.S. Chamber of Commerce, nearly all (98%) small businesses are using an AI-enabled tool, and for good reason.
“AI allows small businesses — who many times do not have the staff or resources of their competitors — to punch above their weight,” said Jordan Crenshaw, senior vice president of the U.S. Chamber’s Technology Engagement Center, said in a blog post.
That data dovetails with PYMNTS Intelligence research, which shows 82% of enterprise CFOs surveyed either use AI in their accounts payable (AP) functions or are considering its adoption.
Popular AI coding assistants include GitHub Copilot, Amazon Q Developer, Cursor, Windsurf, Tabnine, Replit Ghostwriter and others.
Particularly salient is the latest trend called “vibe coding,” which is a new, more intuitive way for people to write computer code using natural language — like how one would talk to a friend.
Instead of writing complex code in a particular syntax, users just need to describe what they want the software program to do, and an AI model helps turn that into working code. This speeds up the creative process and lets users focus more on ideas and less on technical details.
Gao said AI coding tools give small firms the potential to move quickly by eliminating monotonous tasks and providing code concepts.
“We utilize AI on our platform at SleekFlow to enhance customer engagement for businesses with limited advanced technical resources,” Gao said. “For small business players, it is doing more with less while maintaining quality and staying true to their vision.”
AI coding tools will get even better. Meta CEO Mark Zuckerberg said in April that the company plans to unveil AI that can do the work of a mid-level software engineer in 2025, with plans to scale the assistant next year.
See also: Apple and Anthropic Building AI-Powered Coding Platform
Understand What AI Can and Can’t Do
However, Ilia Badeev, head of data science at the Trevolution Group, told PYMNTS that non-technical folks using AI coding tools can only go so far. “It is quite difficult to write a fully functional solution that will scale and continue to develop at the level of large tech companies without engineering skills or experience in industrial software development.”
Daniel Gorlovetsky, CEO at TLVTech, agreed. While solo founders and lean teams get prototypes off the ground faster, AI coding assistants “don’t replace strong engineering,” he told PYMNTS.
However, they do “lower the barrier. You can go from idea to working MVP (minimum viable product) without hiring a full dev team,” Gorlovetsky said. “That’s massive for SMBs that can’t afford $150,000+ engineers or a full-stack agency.”
Still, there’s a trade-off. “AI tools can get you 80% of the way, but if you don’t know what ‘good’ looks like — clean architecture, security, scalability — you risk building something you’ll have to throw away later.”
The smart approach is to use AI to move fast and test ideas. Once it works, bring in “real” developers to lock it down, Gorlovetsky said.
But Santiago Nestares, co-founder of DualEntry, told PYMNTS that his company was able to build an enterprise resource planning (ERP) system like NetSuite despite being told that they would need at least $100 million, which was out of reach.
Also, “we don’t have a 30-year CTO with a track record of enterprise architecture. We’re a team of ambitious, technical people who lean heavily on AI tools to fill those gaps,” Nestares said.
The company used ChatGPT daily to pressure-test its system designs, validate architectural decisions and explore edge (or one-off) cases. It used Cursor to write and refactor code with large language models (LLMs) embedded into its development workflow. It also used AI to help run automated code reviews and maintain quality without adding management layers or quality assurance processes.
“With a team of just 11 people and nine months of focused work, we’ve built a fully capable ERP that has feature parity with NetSuite,” Nestares said. “AI made that possible.”
“If we had to hire for all the knowledge we now get from AI, we’d need a team two to three times the size,” Nestares added. “AI is the great equalizer. … For small businesses without the resources for a full tech org, AI coding tools aren’t just helpful — they’re transformational.”
Read more: Programming Proves to Be Profitable Application of GenAI Technology
Read more: Report: OpenAI in Talks to Acquire AI-Powered Coding Tool Windsurf
Read more: AI Software Tools Could Streamline and Personalize the Commerce Experience
