"Across Big Tech, Wall Street, and other major industries, employers have pushed their workforce to embrace artificial intelligence through a mix of tactics that go beyond mere nudging, including tracking how often they are logging into these workplace tools and linking annual performance reviews to AI usage."
"But at software-as-a-service company HubSpot, 100% of engineers are using AI in their workflows, which the company says has resulted in a 73% increase in the lines of code updated by its engineers. Duncan Lennox, chief product and technology officer at HubSpot, says the milestone was achieved following a three-part phased rollout that began in 2023 and didn't depend on a mandate."
""We found mandates not to be effective," says Lennox, who has steered the engineering, product, IT, and security teams at HubSpot since September, after holding multiple leadership roles at Google and Amazon Web Services."
"But Lennox says that the biggest barrier he had to overcome when initially seeking to leverage large language models in coding was less about fear of job replacement and more about concerns around introducing errors, quality issues, and reliability. After achieving 30% adoption for a coding copilot that was rolled out by HubSpot, Lennox and his team shared incident data that showed engineers that the work being do"
Employers across major industries have promoted artificial intelligence adoption using measures beyond simple encouragement, including monitoring tool logins and tying annual performance reviews to AI usage. HubSpot reports that all engineers use AI in their workflows, producing a 73% increase in the lines of code updated. HubSpot attributes the milestone to a three-part phased rollout starting in 2023 that did not rely on a mandate. HubSpot leadership says mandates were not effective. Agentic coding tools have raised concerns about the future of work, alongside rising tech layoffs and weaker employment trends for programming roles. HubSpot’s chief product and technology officer says the main early barrier was preventing errors and ensuring quality and reliability. After reaching 30% adoption for a coding copilot, incident data was shared to demonstrate safe outcomes.
Read at Fortune
Unable to calculate read time
Collection
[
|
...
]