You renamed your company or were acquired. How do you get AI answers to use the new name?
A rename plan for AI answers: state the old and new name together, update the sources engines cite about you, redirect the old domain and re-test.

The short answer
To get AI answers to use your new company name, publish one page that states the old and new name together, then update the third-party profiles engines cite about you. Redirect every old URL to its matching page, keep the old domain, and re-run the same buyer prompts under both names for at least a quarter.
Key takeaways
- A rename reaches AI answers through two routes: engines that search the web can pick it up within weeks, while answers drawn from a model's memory change only when the model is retrained or replaced.
- Ahrefs found that 14.4% of 66.6 million free, logged-out ChatGPT responses in August and September 2026 triggered a web search, so most answers in that sample came from what the model already held.
- In an Ahrefs panel of 100 brand-name searches on US desktop from July to September 2026, the AI Overviews with visible sources cited Wikipedia in 71.9% of cases, so outside profiles decide much of what a branded answer says.
- The old name should stay on your site as a stated fact, because buyers and models will keep using it, and a page that links both names is what lets an engine treat them as one company.
- Our view: test every rename on two prompt sets, one under the old name and one under the new, and treat any answer that describes two separate companies as the first thing to fix.
In this article
Why do AI answers keep using your old company name?
AI answers keep the old name because much of what an engine says about you was learned before the change. A model that answers from memory repeats what it read in training. An engine that searches finds whatever the web says today, and most of the web still uses the name you dropped.
The split matters, because each route changes at its own speed. Ahrefs, which sells SEO and AI visibility tools, tracked 66.6 million free, logged-out ChatGPT responses in August and September 2026 and found that 14.4% triggered a web search.1 The rest drew on what the model already held. Your new name can appear in searched answers within weeks, while answers from memory wait for the next model.
Model changes do arrive, though not on your schedule. OpenAI's ChatGPT release notes record GPT-6 Sol becoming the model for paid tiers from 7 October 2026 and GPT-6 Luna reaching Free and Go users from 8 October 2026.2 A rename announced after a model's training data was gathered can stay invisible to that model for its whole life.
Where an AI answer finds your company's name
- Buyer prompt
- Model memory, from training
- Web search, if triggered
- Pages and profiles that state the name
- Answer names a company
Prior work, and what this plan adds
Company renames are well covered for search. Google's guide to site moves with URL changes covers domain changes, redirects and the Change of Address tool, and we follow it below.8 Our own notes on keeping AI citations through a redesign and on whether you need a Wikipedia page cover parts of the ground. This piece adds the name itself: how engines join an old name to a new one, and how to test that they have.
A disclosure first. Sigzen AI sells AI visibility audits and sprints at published prices, so weigh our view with that in mind. Several studies cited here come from companies that sell tools in this category.
What goes wrong in AI answers after a rename
Four failures cause most of the damage after a rename or an acquisition. Each comes from a source that still describes the old company, and each shows up in a different kind of prompt. Knowing which one you have tells you where to work first.
- The old name only. The answer names the old company and never mentions the new one. This is the usual pattern in answers drawn from memory.
- Two companies. The answer lists the old and new names as separate firms, sometimes as competitors, with different founding dates or locations.
- Mixed facts. The new name appears, but with the old company's products, prices, leadership or contact details attached to it.
- Wrong owner. After an acquisition, the answer still treats the brand as independent, or credits it to a previous parent company.
A mixed-facts answer is the most dangerous, because it reads as correct until a buyer acts on it.
Our view: most rename problems are entity problems, not wording problems. Engines need evidence that two names refer to one company. Until that evidence is on pages they read, they will keep treating the names as two things.
Which sources shape what AI says about your new name?
Your own site carries the most weight, but outside profiles fill much of every branded answer. A rename plan has to cover both, and in the order engines read them. The figures below show why neither can be skipped.
Company sites are the largest single source of citations. Profound, which sells an AI visibility platform, counted 11.84 billion citations on eight engines from 16 April to 16 July 2026, and about 57% pointed to company-owned websites, with Gemini at 69% and ChatGPT at 47%.4 Semrush's 2026 AI Visibility Index, built on 126 million US prompts from January to April 2026, found ChatGPT cited about 15 sources per response against about 3 for Gemini.5 ChatGPT reads widely, so outside pages that still carry the old name reach it more often.
Branded searches lean on outside profiles too. Ahrefs analysed 232.8 million US desktop Google result crawls from July to September 2026, and on 29 September 82.91% of branded results showed an AI Overview.3 In a separate panel of 100 brand-name searches over the same months, the AI Overviews with visible sources cited Wikipedia in 71.9% of cases, YouTube in 38.6% and LinkedIn in 22.8%.3
Branded answers in 2026
Your new name is summarised from many places
Our view: if your company has a Wikipedia article, its first line decides much of what a branded AI Overview says after the rename. Getting that line right through the proper process matters more than any page on your own site.
Step 1: publish one page that joins the two names
Start with a single page on your own site that says, in plain text, that the old name is now the new name, since when, and why. Engines need one clear statement to connect the names. Buyers searching the old name need it too.
Write it as a fact, not a press release. One opening sentence should carry the whole change: "Tailspin is now Fabrikam Field Service; the company changed its name on 1 March after Fabrikam acquired it." Then state what stayed the same, such as the products, the team, the contracts and the support routes, and what changed.
Put the page where engines look for company facts, usually the About page or a dedicated "Our name" section linked from it. Keep it in the HTML the server sends. Google's guidance for its AI features asks for important content in textual form, and the same applies to every engine that fetches pages.9
What the page should state
- The old name and the new name, together, in the first sentence.
- The date of the change and, after an acquisition, the parent company's name.
- The legal entity name, if it differs from the trading name.
- What carried over: products, team, support routes and existing contracts.
- Links to the profiles you control under the new name.
Back the page with structured data. Schema.org's Organization type includes properties for a legal name, an alternate name, a parent organisation and sameAs links to profiles elsewhere.10 Set the new name as the name and the old one as an alternate name. Be honest about the evidence, though: Ahrefs added JSON-LD schema to 1,885 already heavily cited pages between August 2025 and March 2026 and found no reliable lift in citations.13 Markup helps machines join the names; it does not earn citations on its own.
Step 2: move the domain without breaking what engines cite
If the rename brings a new domain, treat it as a site move with URL changes. Every old URL should redirect in one step to the page on the new domain that carries the same facts. Never send the old site to the new homepage in one sweep.
Google's site move guide asks for server-side permanent redirects, sent straight to the final address and kept for as long as possible, generally at least one year.8 It also describes the Change of Address tool in Search Console, which applies when the domain itself changes.8 Our view: keep the old domain registered for good, not for a year. Answers drawn from memory can quote an old address long after the move, and a lapsed domain can end up with someone else.
Carry the facts over word for word in the first release, then rewrite. If the domain, the name and the copy all change on one day, any fall in AI citations afterwards cannot be traced to one cause. Our redesign guide covers the citation log, the redirect map and the re-runs in detail, and every step applies to a rename.
Two ways to handle the old name
Erase the old name
- Old URLs redirect to the homepage
- No page mentions the old name
- Profiles show only the new name
- Engines see two unrelated companies
Bridge the two names
- Each old URL redirects to its match
- One page states old and new together
- Profiles say "formerly" for a while
- Engines can join the two names
Step 3: update the profiles engines cite about you
Work through outside profiles in the order engines cite them for your company, not in the order your team happens to own them. Your prompt runs show which domains appear in answers about you. Fix those first, and leave the long tail for later.
Most companies can update a short core list themselves within a fortnight: LinkedIn, Google Business Profile, YouTube, the main review sites and trade directories in their industry. Add official registries and regulators where they apply, since many engines treat them as authoritative. Use a "formerly" line on each profile for at least a few months, so the old name still leads somewhere.
Wikipedia and Wikidata need more care. Wikipedia's conflict-of-interest guideline discourages people with a close connection from editing articles about their own company, and asks them to propose changes on the article's talk page instead.11 Post the rename there with an independent source, such as a registry filing or press coverage. Wikidata has its own notability policy; if your company already has an item, the name change can be recorded there with a reference.12
| Source | Who can change it | What to do | Priority |
|---|---|---|---|
| About page and "Our name" section | You | State old and new name together, with the date | Day one |
| LinkedIn company page | You | New name, "formerly" in the description | Day one |
| Google Business Profile | You, subject to review | New name and website, matching signage | Week one |
| Wikipedia article | Independent editors | Propose the change on the talk page, with a source | Week one |
| Trade directories, review sites | You or the site's staff | Request the change; keep the old name as former | Weeks two to four |
| Old press and partner pages | Their owners | Ask the few your prompt runs show engines cite | Month two |
Illustrative: a name-change register for a fictional company. Order sources by how often engines cite them in your own prompt runs.

How should an acquired brand appear in AI answers?
An acquired brand should appear with its new owner named, whether or not the brand name itself changes. Buyers ask who stands behind a supplier, and an answer that still calls the brand independent can mislead them. The fix depends on whether the brand keeps its name.
If the brand keeps its name, add the parent company to the About page, the legal footer and the Organization markup as a parent organisation.10 The parent's site should list the brand and link to it. That pair of pages, linking each other, is the cleanest evidence an engine can find.
If the brand is folded into the parent, the old site's pages should redirect to the matching pages on the parent's site, and the parent should keep a page that names the former brand. Products sold under the old name need their own line, because buyers ask about a product long after the company name has gone.
Buyers do use these answers to choose. In G2's survey of 1,076 B2B decision makers in North America, EMEA and APAC in March 2026, 51% said they start software research in an AI chatbot more often than in Google.6 In the same survey, 69% said an AI chatbot had surfaced information that led them to choose a different vendor than they expected.7 An answer that splits your brand from its owner can send that choice elsewhere.
Step 4: test the rename on two prompt sets
Run two prompt sets: one that uses the old name and one that uses the new name, plus your usual unbranded buyer prompts. The old-name set shows whether engines redirect buyers who still use it. The new-name set shows whether engines know who you are.
Take the baseline before the announcement, if you can. Run each prompt several times on each engine you track, because answers vary between runs, and record the name used, the facts stated and the URLs cited. Our guide to building a prompt set covers where the unbranded questions come from.
What to score in each answer
- Which name the answer uses, or whether it uses both.
- Whether it treats the names as one company or two.
- Whether the facts attached to the new name are current.
- Which of your URLs it cites, old domain or new.
- After an acquisition, whether it names the right parent company.
Separate searched answers from answers drawn from memory where the engine shows it. A searched answer that still uses the old name points to a page you can find and fix. An answer from memory may simply be waiting for the next model. When a model changes between two re-runs, our note on measuring around a model release explains how to tell its effect from yours.
An AI answer plan around a rename
Baseline under the old name; list the domains answers cite about you.
Name page live, structured data, redirects, LinkedIn and core profiles.
Wikipedia talk page request, directories and review sites.
First re-run on both prompt sets; fix any two-company answers.
Re-run again; ask owners of cited old pages to update them.
What will not change quickly?
Answers drawn from a model's memory will not change until the model does. You cannot speed that up, and no form lets a company rewrite what a model learned in training. What you can do is make sure the next model, and every searched answer before it, finds the new name joined to the old one.
Old coverage moves slowest. News articles, conference pages and partner listings keep the old name for years, and most will never be edited. That is fine as long as your own name page explains the change, because an engine that finds both names can join them through it. Ask only the few owners whose pages engines cite most often, which your prompt runs will show.
Wrong facts can also outlive a rename. An answer may attach last year's prices or a former office to the new name. Our guide to fixing wrong AI answers shows how to trace a wrong fact back to the page it came from, and the method applies unchanged.
Our view: plan for a long tail and report it honestly. A rename is done for buyers when your site, your core profiles and searched answers all use the new name. It is done for AI answers only when a new model has been trained after the change.
A rename checklist for AI answers
Add an AI answer stream to the rename plan your brand and legal teams already run. It needs one owner, two prompt sets and a list of the sources engines cite about you. Most of the work happens in the first month.
- Run a baseline under the old name before the announcement, on every engine you track.
- Publish one page that states the old and new name together, with the date and any new parent company.
- Add Organization markup with the new name, the old one as an alternate name and links to your profiles.
- Redirect every old URL to its matching page, keep the old domain, and use the Change of Address tool.
- Update LinkedIn, Google Business Profile and the directories engines cite, each with a "formerly" line.
- Propose the Wikipedia change on the talk page with an independent source; never edit the article yourself.
- Re-run both prompt sets at four and eight weeks, then each quarter, and fix two-company answers first.
Our view: the name page and the two prompt sets cost a day and save months of guessing. If you want an outside baseline, Sigzen AI's AI visibility audit runs one across five engines and names the three pages to fix first.
Sources
- Ahrefs, ChatGPT query fan-out study: 66.6M free, logged-out GPT-5.6 Luna responses, Aug–Sep 2026, all countries; factual subset 20,000 US prompts; citation ranks US, Mar–Sep 2026 (Oct 2026)
- OpenAI, ChatGPT release notes: GPT-6 Sol for paid tiers from 7 Oct 2026 and GPT-6 Luna for Free and Go from 8 Oct 2026 (Oct 2026)
- Ahrefs, AI Overviews on branded searches: US desktop, Jul–Sep 2026; source shares and position from a separate panel of 100 brand-name queries, 14 weekly checkpoints (Oct 2026)
- Profound, where AI citations come from: 11.84B citations, 8 engines, 16 Apr–16 Jul 2026 (Jul 2026)
- Semrush, 2026 AI Visibility Index: 126M US prompts, Jan–Apr 2026 (Jun 2026)
- G2, 2026 AI Search Insight Report: 1,076 B2B decision makers, North America, EMEA and APAC, Mar 2026
- G2, 2026 AI Search Insight Report: 1,076 B2B decision makers, Mar 2026
- Google Search Central, Site moves with URL changes (updated Aug 2026)
- Google Search Central, AI features and your website (updated Dec 2025)
- Schema.org, Organization type
- Wikipedia, Conflict of interest (community guideline)
- Wikidata, Notability (policy)
- Ahrefs, schema and AI citations test: 1,885 already heavily cited pages, Aug 2025–Mar 2026 (May 2026)


