Shopify ships three operator changes, Google names fake author bios, and brand SERPs quietly flip to AI Overviews
Verified changes landed this week. Most of them touch work you were already planning to do before Black Friday.
Shopify now feeds compare-at prices to AI shopping agents
Shopify's Catalog now shares your compare-at price with AI shopping agents, so an agent reading your catalog can see when a product is marked down and may advertise strikethrough pricing. Previously, agents only saw the current price. The toggle lives in Shopify admin under Sales channels, Agentic, Sources, Shopify Catalog, and each store starts with a default selection that does not change automatically when you change how you use compare-at prices elsewhere.
This hits every Shopify merchant selling through ChatGPT, Google's AI surfaces, Copilot, or Meta's agentic channels. It is especially sharp for merchants with MAP agreements or stores where compare-at was used as an internal reference rather than a true was-price, because an AI agent will now surface that difference as a discount claim.
The move this week: open admin, go to Sales channels, Agentic, Sources, Shopify Catalog, review the product data mapping, and set the compare-at toggle deliberately before BFCM promo pricing goes live and the agent surfaces repeat whatever is in the field.
EU and UK stores can finally auto-sync translations across up to eight languages
For merchants in Austria, Belgium, Denmark, France, Germany, Ireland, Italy, Netherlands, Spain, Sweden, Switzerland, and the United Kingdom, Shopify's Translate and Adapt now supports weekly auto-sync and raises the automatic language cap from two to eight. Manually written translations are never overwritten. Auto-sync is off by default, runs weekly at the same day and time, and the first run can be up to a week after you enable it. Stores outside this country list are still capped at two automatic languages.
This hits any EU or UK merchant whose product descriptions, policies, or landing copy change frequently, where translated storefronts have been drifting behind the source language and buyers have been reading stale content.
The move: open Translate and Adapt, pick each published language, flip auto-sync on, and do it this week so the first weekly run lands before your BFCM copy changes. Review the first auto-generated pass against anything you had hand-translated for voice.
Auto-generated store policies are live in Shopify
Return and refund, contact information, terms of service, shipping, and legal notice policies can now be generated and kept in sync with your store settings automatically. The system pulls from your shipping config, tax settings, fulfillment methods, and contact details so the policy text stays consistent with how the store actually operates. Shopify is explicit that the merchant remains responsible for compliance with applicable laws.
This hits new launches, merchants expanding into new markets, and existing stores whose policy pages have not been touched in a year and no longer match current shipping or returns practice. Mismatches between stated and actual policy are a documented driver of chargebacks and disputes, which matters more in the next six weeks than at any other point in the year.
The move: go to Settings, Policies, pick each policy, enable Use automated policy, and save. Then read what it produced against your local legal requirements, especially if you sell into the EU or UK.
Google now calls fake author profiles deception
Google updated its helpful, reliable, people-first content guidance with a new paragraph telling site owners to avoid deceptive authorship information. It specifically names AI-generated headshots, made-up names, and false credentials, and says any form of deception makes a page untrustworthy to both users and automated quality systems. Separately, Google's September spam update finished rolling out on October 8 after nearly 14 days, the longest spam update of the year.
This hits any site, agency client, or content program running placeholder bylines, stock-photo author headshots, or AI-generated personas on blog posts and resource pages. Google has not described a specific penalty tied to this guidance, but it has reframed the behaviour as a trust signal that both humans and ranking systems act on.
The move: pull a list of every author profile on the site, and this week replace any fabricated headshot, invented name, or hollow credential with a real person, a real photo, and a byline link that resolves to something verifiable. If a page cannot be assigned to a real author, either reassign it or remove the author block entirely.
SEJ SEO Pulse: spam update ends, crawl timing, fake author guidanceSEJ: September 2026 spam update complete
Branded search is quietly becoming an AI Overview surface
Three separate trackers now show AI Overviews on the majority of branded queries. Ahrefs tracked 100 global brand names weekly from July through September and found Overviews on five of 98 results on July 1, rising to 63 of 100 by September 30. Nectiv's Chris Long reported a one-off scrape on September 30 finding Overviews on 93 of 100 enterprise brands. DemandSphere's tracked branded keywords rose from 26.12 percent AI Overview presence on September 1 to a peak of 90.48 percent on September 27. Google has not announced the change.
This hits any brand that assumed its own name SERP was safe owned traffic. The data does not yet show click effects or how often brands' own sites are cited inside the Overview, so you cannot assume the Overview is pulling from you.
The move: this week, search your own brand name in Google in an incognito window, screenshot the Overview if one appears, and check which sources it cites. If your own site is not cited, or if the Overview summarises your brand from a stale third-party page, prioritise refreshing your About, pricing, and product-overview pages so the next crawl has current, accurate text to pull from.
Outlook.com is the Q4 deliverability pain point to fix first
Al Iverson's BFCM 2026 preview on Spam Resource, published October 10, reports that last year's reader survey put Microsoft Outlook.com at the top of Q4 mailbox provider pain points, with Gmail expected a close second this year. His core infrastructure tip is unchanged and specific: lock in warmed IPs and configured domains well before November. Mid-Q4 is the worst time to ramp new sending IPs, launch new subdomains, or run a brand migration. Black Friday falls on November 27 and Cyber Monday on November 30.
This hits any team sending high-volume promotional email through November, especially those who were planning to spin up a new subdomain or move ESPs in the next six weeks.
The move: this week, do not start any new IP or subdomain warm-up you have not already begun. Instead, pull your SPF, DKIM, DMARC, and PTR records, verify alignment, and check your Gmail Postmaster and Microsoft SNDS data for the sending domains you will actually use in November. Freeze infrastructure changes until December.
Checked, and quiet
Regulated paid media is represented in this week's input through a Search Engine Journal piece on AI default settings in Google and Meta campaigns, but there was no new platform policy change behind it this week. The underlying compliance risks for finance, healthcare, and legal advertisers are real and ongoing, and we will cover them when a discrete platform change gives you a specific setting to flip.
Signals checked weekly, quiet weeks admitted. If you want these signals checked against your own site specifically, run the free read at app.signalandscience.com.
Get the next Signal & Science brief when there is something worth acting on.