Tech stack changing as prompt requests shorten

As AI increases its capabilities, there’s less of a need for users to give their chatbots lengthy directives—which is leading to some interesting changes, one company found.

Research from Elvex, an enterprise AI adoption platform, indicates that AI prompts are shortening as the underlying technology gets better at predicting what users want. That’s leading to higher positive feedback from customers, Elvex CEO Sachin Kamdar told IT Brew.

“They’re getting more back from the systems because of integrations and context, and they’re liking those responses more,” Kamdar said.

Kamdar acknowledged that Elvex is relying on internal data, but he suspects there’s a universality to the findings once you control for customer quality.

“The use cases are somewhat similar if they’re using us versus Copilot versus OpenAI versus Claude or any of those other platforms,” Kamdar said. “But the bias is really around the [Elvex] customer set, which is high mid-market to enterprise companies across verticals like technology consultancies, med tech, healthcare, media, and some others.”

Do the evolution, baby. The AI landscape is changing both human inputs and how the technology is being deployed. Developers are figuring out how to best integrate AI tools and hand over code reviews to chatbots, as IT Brew reported in July.

Darren Meyer, a security researcher with Checkmarx, told IT Brew that there are myriad factors at play in this AI-driven evolution of the tech stack.

“AI is part of software development now and will continue to be—and with that comes pressure from the market, pressure from management, pressure from investors to ship faster and faster and faster,” Meyer said. “That, to me, is why we’re seeing the reduction in human review.”

The evolution in AI prompting is also leading to a shift in hiring, as prompt engineers become less important to workflows. As tech commentator Joe Procopio wrote for Inc. on July 15, AI is getting better at managing inputs: “As both ends of the technology improve, the machine gets better at interpretation of the prompt request and the need for a middleman between requestor and machine evaporates.”

But the prompt shortening that Elvex is seeing shouldn’t be seen as reflecting reduced interest in utilizing AI, Kamdar said. In fact, Elvex data showed the opposite—enterprises are leaning into the tools even more.

“It’s not that people are just getting tired of this stuff,” Kamdar said. “They’re getting value out of it, so they can do it in shorter ways, so they’re leaning in harder in the other areas that are more behind the scenes to make sure that that persists over time.”