Salesforce wants to acquire Loops to help Tableau explain KPI changes

Nov 1, 2026 | 360° IT News, Data

Salesforce has announced a definitive agreement to acquire Loops, an AI-based analytics platform, with the intention of integrating its technology into Tableau after closing.[1] The aim is to monitor metrics and bring out the factors associated with their changes. For businesses, the promise is to spot earlier what deserves analysis, provided they have reliable data and do not mistake a suggested explanation for a proven cause.

01

A deal announced before an integration that has yet to be built

The transaction is expected to close in early November, during Salesforce’s fiscal year 2027, subject to customary conditions. The press release gives no precise date for when Loops’ capabilities will be available in Tableau.[1] This is therefore an announced product direction, not a new feature already available to all users.

This distinction matters for a team preparing its budget or its analytics architecture. It can follow the project and identify its needs, but it should not base an operational deadline on an integration whose terms have yet to be specified.

02

Metrics become the starting point for an investigation

According to Salesforce, Loops monitors performance, detects trends and anomalies, and then brings out what contributes to changes. Its functions can be triggered from Slack or by email. The vendor wants to use them to strengthen Tableau’s agentic analytics and the prioritisation of actions.[1]

A key performance indicator, or KPI, summarises one dimension of the business: conversion, processing time or retention, for example. When it changes, the first need is often to work out where to look. Does the change concern all customers, or one particular region, product or channel? Proactive analysis can help organise this search.

03

Detecting a change is not enough to make a decision

This grid distinguishes three stages of analytical reasoning and the checks they require. It is an editorial reading of the business need.

Stage Question to address Useful check
Detection Which metric has changed and over what period? Check data freshness and the comparison baseline.
Explanation Which segments contribute to the change? Examine volumes, changes in measurement and other factors.
Action What intervention could improve the result? Validate the hypothesis and measure the effects before rolling it out.

04

A factor associated with a decline is not always its cause

A channel may account for a large share of a drop in sales because its traffic has changed, because a product is unavailable there or because data collection is incomplete. Breaking down a result helps locate the phenomenon; it does not automatically decide between all these explanations.

Before changing a campaign or a process, the analysis must be compared with business events. A controlled comparison or an experiment can then help test a hypothesis. The decision depends as much on this verification as on the clarity of the generated commentary.

05

Defining KPIs remains essential groundwork

Two departments may use the same word for different measures. A conversion rate calculated per visit does not match a rate calculated per user. Likewise, a cancelled order or a late-arriving event can change the result depending on the rules applied.

Continuous monitoring therefore requires a shared definition: formula, population, period, exclusions and reference source. The history of changes must also be kept. An alert triggered after a formula change does not necessarily describe a real change in activity.

06

An SME can test the value on a single metric

Take an illustrative scenario: an online shop notices a drop in its conversion rate. The team compares devices, channels and product categories. If the drop is concentrated on mobile, it then checks for payment incidents, stock shortages and recent changes to the customer journey. This scenario is not a customer result published by Salesforce.

The pilot should check that the tool surfaces useful observations without multiplying secondary alerts. The team can measure the time to a confirmed anomaly, the time needed to qualify the explanation and the share of alerts that are genuinely actionable. Nor does the absence of an alert prove the absence of a problem.

To assess a future integration, data access, costs and the scope of sources will need to be clarified. The press release provides no independent measure of accuracy or of generalisable time savings. These points remain to be verified on the company’s own data.

07

The IT skill to remember: Data Management

This development draws directly on Data Management. Professionals must know where information comes from, check its quality and ensure that transformations preserve the meaning of measures. A convincing explanation can still be misleading if the starting data is incomplete or poorly reconciled.

Their role includes documenting metrics and managing access rights. Users must understand what a KPI measures and what limits frame its interpretation. The details surfaced in an alert must also respect authorised access, particularly when they concern customers or employees.

The work continues in dialogue with business teams. Analysts must be able to verify the observations put forward, spell out uncertainties and track the effect of decisions taken. Automation can speed up investigation; the responsibility for giving results a reliable meaning remains.

08

Proactive analysis must steer a useful investigation

The value of proactive analysis is measured by its ability to steer a useful investigation. A dashboard that comments on its own changes does not thereby become an autonomous decision-maker: data quality, hypothesis testing and monitoring of results remain decisive.

Further reading

Four CCCLX reads to explore data quality, sharing and use in more depth.

Storage and monitoring

Google Cloud Storage Intelligence Advisor detects storage anomalies

Detecting anomalies in storage environments.
Interfaces and data

Talk to your database: Oracle improves its NL2SQL AI

Conversational access to company data.
Data and sharing

Databricks extends Delta Sharing to AI and aims to break down data silos

Data exchange between tools and environments.
Cloud and agentic AI

Merck and Google Cloud commit one billion dollars to agentic AI

Cloud investment supporting new uses of AI.

Sources

  1. 01

    Salesforce (12 October 2026). “Salesforce Signs Definitive Agreement to Acquire Loops”. Announcement of the acquisition agreement and plans to integrate Loops technology into Tableau.

    View source →

360° IT News, decoding technology to understand the jobs of tomorrow.

Discover CCCLX →