Salesforce buys Loops to make Tableau explain why your numbers moved

Salesforce has signed a deal to buy Loops, a small AI analytics startup, and fold its technology directly into Tableau — betting that the next version of business intelligence doesn't wait for someone to open a dashboard and ask why a number moved.

Loops uses machine learning to watch a company's key metrics continuously, flag anomalies, and explain what likely drove a change, pushing the finding to a Slack channel or an inbox before anyone goes looking. Once the deal closes, expected in early November of Salesforce's 2027 fiscal year, Loops stops existing as a standalone product; its team and technology move into Tableau. Neither company has disclosed a purchase price.

What changed

Tableau built its business on letting people explore data themselves: drag a field onto a shelf, build a view, find the story. Loops inverts that. Instead of a person searching a dashboard for what changed, the system watches the metrics on its own and writes the explanation first. Futurum Group's Keith Kirkpatrick frames it as Salesforce shifting competition away from who has the prettiest charts and toward who explains the data best, inside tools people already use for work.

Mark Recher, Tableau's EVP and GM, put the rationale this way: Loops "accelerates this mission, using powerful agentic AI to uncover the drivers behind business performance," according to Pulse2's report on the deal. Loops co-founder and CEO Tom Laufer framed the company's premise more plainly than most vendor quotes bother to, per Futurum Group's write-up:

Who it affects

The immediate audience is narrower than "every marketer": teams that already run Tableau for revenue, funnel or retention reporting, and anyone whose marketing ops stack sits on top of Salesforce. For them, a KPI alert that already explains itself — rather than a dashboard that only shows the drop — changes who has to be the one staring at the data to catch a problem.

It also matters to marketers evaluating analytics vendors generally, because this is now the pitch across the category. Per Futurum's own 1H 2026 survey of 830 decision-makers, 75.8% rank predictive or AI-driven analytics among their top three software priorities, and 62.4% say the same about autonomous agents. Salesforce already leads analytics and BI software with an 8.9% revenue share, well ahead of Microsoft's 4.8%, which is exactly the installed base Loops gets layered onto first.

"Analytics should help businesses understand what drives their results and decide what to do next." — Tom Laufer, CEO and co-founder, Loops

What a software marketer should do differently

Marketing leaders who already live inside Tableau should treat this as a preview of what their own reporting will start doing for them, not a reason to buy anything new yet — Salesforce hasn't said when Loops' features actually ship inside Tableau, only that the acquisition is happening. The more useful move right now is diligence, not adoption.

• Ask any analytics vendor, Salesforce included, to show a causal explanation next to a real anomaly in your own data, not a demo dataset, before trusting the "why" it gives you
• Separate correlation from cause in how these tools get evaluated: a system that reliably spots that leads dropped is useful even if its explanation for the drop is wrong, but only if someone checks
• Audit where KPI alerts already land — Slack, email, a dashboard nobody opens — and decide whether an "explained" alert changes who owns the follow-up
• Watch the free tier of whatever BI tool you already pay for; Futurum's survey puts integration and time-to-value as the top purchase criteria, which is where a bundled feature usually beats a point solution

The deeper question Futurum raises is whether Microsoft and other BI vendors answer with their own acquisitions or partnerships before Salesforce's installed-base advantage compounds. For a marketer, that competitive pressure is good news regardless of which vendor you use: "why did this number change" is becoming a question software is expected to answer unprompted, not a question analysts get paid to chase down.

The build-versus-buy question gets smaller, not bigger

Acquiring Loops rather than building root-cause explanation in-house is itself a signal worth reading. Tableau has had years to ship this kind of feature organically and instead bought a small outside team for its models and its talent. That's a tell for anyone budgeting an internal "AI analytics" project this cycle: the vendors with the deepest data and the most engineers still decided buying specialized talent was faster than building it. A marketing team weighing the same build-versus-buy call on a smaller scale — an internal dashboard that flags anomalies, say — should treat that as a reason to shop before committing a quarter of engineering time to the same problem.

Disclosure: MarTech, one of the sources below, is owned by Semrush.