Temporal, a workflow‑automation firm backed by a recent financing round, says its AI‑driven tooling has accelerated feature delivery by roughly a quarter while revenue has roughly doubled this year.
AI‑focused “reading period” spurs faster development
For the final two weeks of 2023, Temporal cleared its calendars for a “reading period” in which staff held no meetings and focused on learning and building.
Co‑founder and CTO Maxim Fateev spent that time coding with AI agents and returned reporting that tasks once taking six months were now finished in under thirty days. CEO Samar Abbas credits the experiment with setting the company’s AI posture for 2024, noting that such internal results carry more weight than any formal directive.
The experiment set a new pace.
Related: Nscale acquires Anyscale to advance multi-cloud neutrality
Temporal’s core product provides “durable execution,” a reliability layer that keeps long‑running software alive through crashes and outages. Clients such as Nvidia and Netflix rely on this infrastructure to avoid costly downtime.
Company‑wide push for AI adoption
In February, Temporal raised a large sum at a $5 billion valuation, pledging to make agentic AI reliable for broader use.
Abbas, who swapped roles with Fateev earlier this year, says the firm must first demonstrate the technology internally. “We are a company of builders, which means it’s not just engineering,” he told the outlet. “The entire organization needs to re‑evaluate how business gets done and how it uses these tools.”
With more than 500 employees, including 200 engineers, the expectation is that every staff member examines their workflow, experiments with AI tools, and adapts accordingly. The mandate is not to adopt a specific product; rather, employees must adopt the practice of improving their work with AI, or risk losing their role at Temporal.
Related: The Engagement Ring Proposal: Planning the Perfect Moment
Temporal tracks vendor costs in an internal system called Vantage, imposing hard spending limits and reviewing outliers bi‑weekly. The firm avoids public leaderboards for token usage, believing they create counterproductive competition.
Among the tools in use, the company has adopted Claude Code as the default coding assistant after the holidays, while also permitting Cursor and foundation models from OpenAI and Anthropic. Open‑weight models remain unapproved pending security and privacy assessments.
Financially, AI‑related spend has risen about fivefold this year, driven largely by coding agents. The firm keeps detailed usage data but does not expose it as a performance metric for employees.
Temporal’s internal security workflow illustrates the platform’s flexibility. Previously, the security team filed vulnerability tickets that competed with feature work for engineering attention. Now, using a Durable workflow called Deputy, the team automates the entire remediation cycle—from detection to deployment—within Temporal’s own infrastructure.
Related: Small firms turn to AI for customer help
Abbas believes the broader market is moving toward “long‑lived, asynchronous, mission‑critical” agents, a problem his firm has been solving for years with durable execution. He describes the platform as model‑, language‑, and cloud‑agnostic, offering an open‑source server that customers can run on premises.
The company’s recent Replay conference showcased serverless workers and durable streaming aimed at agent workloads, reinforcing its commercial narrative.
Looking ahead, the rapid growth in AI spend and the shift in internal bottlenecks suggest that Temporal will need to refine its cultural and operational safeguards. As the firm continues to scale its AI initiatives, maintaining the balance between speed and reliability will be essential to turning accelerated development into tangible customer value.
Abbas remains cautious about drawing a direct line between the AI‑driven speed gains and the revenue increase. He acknowledges that while engineering is delivering features 20 % to 30 % faster with the same team size, and revenue has roughly doubled, the firm has not yet proven a causal relationship. “It’s not about shipping features faster. It’s about shipping value customers care about,” he repeats, emphasizing the need for measurable outcomes.
