Salesforce has signed a definitive agreement to acquire Loops, an Israeli startup that builds AI-powered analytics software. The company announced the deal on October 12, 2026.
The goal is to give Tableau, Salesforce’s data visualization platform, a brain that never clocks out.
The price tag is modest by Salesforce standards. The deal is estimated at around $30 million, though official terms were not disclosed. For a company that has written multi-billion-dollar checks for AI in recent years, this is closer to a rounding error than a bet-the-company move.
Advertisement
What Loops actually does
Loops was founded in 2020 by CEO Tom Laufer and CTO Ido Adiv. Its core product falls under a category called agentic analytics.
More specifically, Loops continuously tracks key performance indicators, or KPIs. These are the core metrics a business cares about, such as revenue, conversion rates, or customer churn.
When one of those numbers moves, the platform uses machine learning to explain why it changed. It then surfaces insights on its own and recommends actions.
The software connects with enterprise data sources, so it can pull from the systems companies already run. It also has what the industry calls “headless” features, meaning users can trigger analytics without opening a dashboard at all. Insights can arrive through collaboration tools like Slack and email, where people already spend most of their workday.
A small check from a big spender
The acquisition is expected to close in early November 2026, which falls within Salesforce’s fiscal 2027. Once complete, Loops’ technology will be folded into Tableau rather than run as a standalone product.
Salesforce has been steadily picking up AI-focused capabilities from Israeli startups. Loops is the latest addition to that list.
What this means for enterprise analytics
For Tableau users, the shift could be meaningful. The product has been built around visualization: charts, graphs, and dashboards that humans interpret. Loops pushes it toward “always-on” insights, where the software flags problems and proposes next steps without being asked.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.