Klaviyo Acquires Elias Torres' Agency

Klaviyo (NYSE: KVYO) announced on August 5, 2026 that it has agreed to acquire Agency, a three-year-old AI customer success startup that had raised $32 million from Sequoia, Menlo Ventures, and Felicis — and the deal comes with something rarely attached to an acquisition price: the founder himself, Elias Torres, who joins the publicly traded e-commerce platform as its new Chief Product Officer.

What Changed

Agency was founded in 2023 by Torres, a serial entrepreneur with two clean exits already on his record. He co-founded Performable, which HubSpot acquired in 2011, then spent eight years as CTO at conversational-commerce startup Drift before its $1.2 billion sale to Vista Equity in 2021. Agency was his third act — a small but well-capitalized AI shop focused on autonomous customer success tooling, built on the same conversational premise he explored at Drift but aimed squarely at AI agents rather than chatbots.

The acquisition price was not disclosed. What was disclosed is the scope: Klaviyo is absorbing Agency's full 25-person team, and Torres will report directly to co-founder and co-CEO Andrew Bialecki. His remit as CPO is to lead the development and expansion of Klaviyo's two existing AI agent products — Composer, which automates the construction of marketing campaigns, and Customer Agent, which handles post-sale support tasks like returns and order tracking. These are domain-specific agents, not general-purpose assistants; both names and scopes were already in place before the acquisition.

Klaviyo's IPO in September 2023 valued the company at $9.2 billion, though multiple reports note its stock has since declined alongside SaaS peers in a rougher rate environment. The acquisition is being framed less as a financial play and more as a strategic acceleration — Klaviyo is betting that putting an experienced AI-agent builder in charge of its product roadmap, backed by its existing data infrastructure, is how it wins a market that pure-play rivals are also actively pursuing.

The deal is subject to customary closing conditions; no close date was specified in the announcement.

How It Works

The relationship between Torres and Bialecki predates Klaviyo entirely. Around 2010, Torres hired Bialecki — fresh out of Harvard — as one of Performable's first engineers, and by Torres' own account, Bialecki absorbed early-stage startup dynamics quickly. When Bialecki left to bootstrap Klaviyo, and when Klaviyo raised its first outside capital in 2015, he invited Torres back as an angel investor. This acquisition is, in that sense, the third chapter of a professional relationship that spans fifteen years.

That history matters to how the deal is structured. This is not Klaviyo buying a technology asset and absorbing it into an existing engineering org. It is Klaviyo buying a CPO. Torres comes in with direct authority over the agent product line, and the 25-person Agency team travels with him. The structure follows a pattern that has accelerated in enterprise software over the past two years: an established SaaS platform with distribution and customer data acquires a small AI-specialist startup not primarily for the IP but for the builders — and for a leader who already knows how to ship agent products.

Klaviyo's official press release describes the company as an "autonomous B2C CRM," a framing that positions Composer and Customer Agent as the core of the platform's future rather than optional add-ons. Bialecki stated the goal plainly: bring these agents to Klaviyo's current 200,000 business customers, and eventually to millions more.

What It Means for Developers

The explicit competitive thesis here is worth attention. Bialecki told TechCrunch directly that Klaviyo's years of accumulated customer behavioral data give its agents an edge over dedicated AI agent competitors like Decagon and Sierra. The argument is not that Klaviyo's models are better — it is that the data already living inside the platform (purchase history, engagement signals, customer profiles built across years of CRM usage) makes its agents more accurate and more useful than anything a newer entrant can deploy without that history.

This is the argument every platform-native AI product will eventually have to make, and it has real weight in the agent context. General-purpose agents operate on whatever context you put in front of them. An agent trained and tuned against years of real transactional data from 200,000 businesses in a specific vertical starts from a structurally different position. Our read is that this framing — "the model is not the moat; the data is" — is one of the cleaner distillations of platform AI strategy to emerge from a real acquisition announcement in 2026. It is explicit, it is falsifiable, and it names the competitors it is trying to beat.

For developers building in or around e-commerce, the immediate practical question is what the Torres-led product roadmap does to Klaviyo's API surface and agent extensibility. Klaviyo has not announced new developer-facing tooling alongside this deal, and the Composer and Customer Agent products are consumer-facing rather than developer-extensible in their current form. That may change under a CPO whose entire recent career has been in agent product design. The overcentral.com analysis points to the 25-person team as a signal that this is genuine product investment, not a talent holding pattern — but the actual developer experience implications will depend on what Torres ships in the quarters ahead, which remains to be seen.

What is already clear is the direction: Klaviyo is committing to autonomous agents as the primary product surface for B2C commerce, with a CPO brought in specifically to build that out. Whether Composer and Customer Agent stay narrowly scoped or expand into something more composable is the question worth watching.

Sources