Klaviyo Co‑CEO Andrew Bialecki Reveals AI Build System and Workforce Upskilling at SaaStr AI Conference

The Dark Factory: An Internal AI Assembly Line

At the SaaStr AI summit in San Francisco, Klaviyo co‑founder and co‑CEO Andrew Bialecki outlined how the 2,300‑person public company creates its AI products. The session focused on the internal „Dark Factory,” a tool called Composer, and the mandate that every employee reach an L3 competency level.

Bialecki explained that Klaviyo’s growth to a $1.5 billion valuation hinged on a single insight: linking marketing messages directly to shoppers’ carts and Shopify data, rather than merely tracking email opens. That data‑first approach now drives the company’s AI roadmap. To scale this vision, Klaviyo built a dedicated AI production line—nicknamed the Dark Factory—where engineers, data scientists, and product managers collaborate under a unified workflow. Composer, the in‑house orchestration platform, automates model training, testing, and deployment, reducing time‑to‑market from months to weeks.

The Dark Factory operates like a „factory floor” for machine‑learning models. Teams submit feature requests, and Composer routes them through standardized pipelines that include data ingestion, feature engineering, and continuous validation. Bialecki said the system eliminates siloed experiments and enforces consistent quality checks. By centralizing resources, Klaviyo can iterate on personalization algorithms that predict which products a shopper is most likely to buy, boosting conversion rates across its merchant base.

Why Did Every Employee Need to Hit L3?

Klaviyo required every staff member, from sales to engineering, to achieve an L3 level—a competency benchmark for AI literacy. „If the whole organization understands the basics of model behavior, we can surface insights faster and avoid costly missteps,” Bialecki noted. Training programs paired with hands‑on projects ensured that non‑technical roles could interpret model outputs and provide feedback. This cultural shift helped align product development with real‑world merchant needs, reinforcing the company’s data‑driven ethos.

The company’s AI push has already shown measurable impact. Early adopters report a 12 percent lift in average order value after integrating Klaviyo’s predictive recommendations. Bialecki expects the Dark Factory to double the number of production‑ready models within the next year, further cementing Klaviyo’s position as a go‑to platform for e‑commerce marketers.

Frequently Asked Questions

What is Composer and how does it differ from typical ML pipelines? Composer is Klaviyo’s proprietary orchestration layer that automates each stage of model development, from data collection to deployment, ensuring uniform standards across teams.

How does the L3 requirement affect day‑to‑day operations? Employees undergo a structured AI literacy program, enabling them to interpret model results, ask informed questions, and contribute to feature ideation, thereby accelerating product cycles.

Will Klaviyo’s AI tools be available to all merchants? Yes. The company plans to roll out its AI‑enhanced recommendation engine and predictive analytics to its entire merchant base by the end of the fiscal year.