Good morning. Twilio’s turnaround didn’t hinge on a single quarter. It was a multi-year overhaul of its strategy, cost base, and culture that is now clearly showing up in the numbers.
Bank of America has turned increasingly bullish on Twilio. Analyst Koji Ikeda has reiterated a Buy rating and added the stock to his “Fab Five” basket (alongside Datadog, JFrog, MongoDB, and Snowflake), a group the firm expects to keep outperforming as companies step up spending on AI infrastructure.
“Our transformation is really the result of the disciplined work we’ve been doing for many years,” Twilio CFO Aidan Viggiano told me. “We didn’t just wake up here. We built our way here one decision at a time.”
Twilio lets companies embed voice, video, text, email, and other communications into their products. As AI adoption accelerates, Viggiano describes the company as the “connective tissue” for AI agents: LLMs provide the intelligence, data platforms provide the context, and Twilio manages the customer interactions.
During the pandemic boom, Twilio saw demand pulled forward. Afterward, usage slowed as customers cut back, and revenue decelerated faster than at subscription-based peers, with costs and ambitions built for a very different growth environment.
Phase one: getting disciplinedViggiano described that period as a turning point, when leadership had to abandon the “do everything” mindset of the growth heyday and accept that it could no longer fund every experiment. The first phase was about discipline.
The leadership team aimed to protect the product roadmap and innovation capacity, while directing most reductions to G&A, sales and marketing, and corporate functions. Twilio cut roughly 40% of its workforce over 2022–2023 but avoided blunt, across-the-board reductions, Viggiano said. At the same time, she and CEO Khozema Shipchandler closely examined capital allocation, efficiency, and organizational sprawl. Profitability arrived faster than expected, but growth stayed sluggish.
The Segment call that changed the storyThe tougher calls—and arguably the ones that define Twilio’s new story—came on strategy and focus. A centerpiece is Segment, the customer data platform Twilio acquired in 2021. By early 2024, Segment wasn’t growing well, it was losing money, and some investors wanted it sold.
Selling Segment would have been the easier short-term answer. Instead, Viggiano, Shipchandler, and the leadership team made it profitable and integrated its data capabilities into Twilio’s communications platform, betting its data layer would be critical for AI-powered customer engagement. Many of the contextual data and AI products unveiled at SIGNAL, Viggiano noted, trace back to that decision.
Phase two: earning growth backThe second phase of the turnaround—reaccelerating growth without backsliding on discipline—has been more complex. Viggiano said she began to feel confident and saw signs that the growth side of the strategy was working by mid-2024, after months of targeted work on self-serve tools for developers, partnerships with software vendors, more granular use-case analytics, and efforts to upsell and cross-sell customers.
Because Twilio charges based on usage, her team focused on how customer volumes were trending. Just as important was shifting the culture from “growth at all costs” to balancing growth with profitability and cash flow—a change that required constant communication and clear priorities across the company.
In the first quarter of 2026, one of Twilio’s key metrics—gross profit growth in dollar terms—accelerated to 16% year over year, up from 10% in Q4 2025. Bank of America highlighted Voice AI products as a contributor to that acceleration, seeing Twilio as a beneficiary of rising AI adoption and deeper integration into the AI ecosystem.
“It took multiple years to get there,” Viggiano said, “but we knew we had to stick to the strategy.”
Twilio’s new AI-era moatMore than a cost-cutting story, the turnaround has repositioned Twilio in the AI stack. Viggiano argues that the company’s moat lies in its more than 4,800 carrier connections, regulatory expertise, and fraud-prevention models trained over more than a decade. In her telling, that’s not something you can quickly replicate with AI alone.
Sheryl Estradasheryl.estrada@fortune.com