Brandon Lazovic

Pulse · September 15, 2026

Confirmed

Google adds a Collected Info tab showing what its automated outreach gathered about a business

Google Business Profile now rolls out a Collected Info tab inside Edit profile, surfacing everything Google's automated outreach, calls, texts and WhatsApp messages, has gathered about a business from its verified phone number, including each item's source and the date it was collected, with a Delete option to remove any entry from Google's records. The feature is rolling out for select regions, languages and business categories rather than universally.

In plain termsAutomated outreach means Google's own phone, text or WhatsApp system sometimes contacts a business to double check details, and this new tab shows exactly what it wrote down from those contacts, not just what the owner typed in.

Why it mattersIt gives multi-location and single-storefront owners their first look at, and control over, data Google collected about them without asking permission first.

Our takeThe delete button only touches Google's collected-info record, not any field it already promoted onto the live profile, so an owner correcting a bad automated write-in still has to fix the public listing separately. Watch whether this format expands past select regions and categories, since a first look at an audit trail is only useful once it covers every profile a multi-location account manages.

Confirmed

BigQuery previews six SQL functions that automate causal analysis, trend detection and change-point diagnosis

Google Cloud's September 14 blog post introduces six BigQuery table-valued functions, all listed as Preview in Google's own release notes rather than generally available. AI.KEY_DRIVERS ranks what drove a metric's change, and AI.CAUSAL_EFFECT estimates a decision's causal impact via ARIMA_PLUS counterfactual modeling. Four more machine learning (ML) functions, ML.CORRELATION, ML.DETECT_CHANGE_POINTS, ML.TREND and ML.SEASONALITY, run directly against warehouse tables with no data export, chaining into Conversational Analytics for natural-language queries.

In plain termsCausal effect here means the function tries to isolate how much of a metric's change was actually caused by one specific action, not just correlated with it, by modeling a comparison against what would have happened without that action.

Why it mattersIt moves diagnostic and causal-inference work that used to need a data scientist and a separate modeling step into a single SQL call any analyst can run against the warehouse directly, once the functions clear preview.

Our takeI wrote in 'Google's TabFM brings zero-shot prediction to BigQuery, but the benchmark that sells it is a chart in a blog post' that a Google Cloud launch post is not the same thing as an independently checked benchmark. AI.CAUSAL_EFFECT's counterfactual estimate deserves the same scrutiny before a client report cites it as the causal impact, since modeling a counterfactual is still a model's guess at what did not happen, not a measured fact.