Why Does Claude Describe Our Merged Platform as Two Separate Products?
Claude describes your merged platform as two separate products because most of what it can read still describes two. There may be two websites or product sections, and two sets of docs and pricing. Review sites and marketplaces carry separate listings, and the acquisition coverage said the products would run independently. When Claude answers from model knowledge rather than a web search, it also reflects what it learned before the merger. Anthropic publishes a training data cutoff for each model. The fix is one plain statement of what the platform is and what the former products became. Redirect the old product pages to the platform's pages, and merge docs, pricing and profiles where you can. Then update the boilerplate that still describes the old setup.
Where does the two-product picture come from?
- Acquisition announcements. "[Product B] will continue to operate as a standalone product" was true at the time, and is the most widely syndicated statement about the deal.
- Two sites or product sections still live, each describing its product as complete on its own.
- Separate docs, pricing and sign-up flows, which read as separate products even if they now share a login.
- Separate review profiles and marketplace listings, each with its own reviews, categories and comparisons.
- Two company pages on LinkedIn and business databases.
- Structured knowledge. Wikidata and Wikipedia entries that describe each product separately, with no link between them.
- A vague platform page. "One platform for everything" without naming the former products gives assistants nothing to connect.
Claude's web search draws on Brave Search, which Anthropic names as a subprocessor for that purpose, so the pages ranking there for each old product name shape the answer. If the old product sites rank well, they can keep being cited.
Why does Perplexity use our old company description after a merger?
Perplexity searches live and answers from the top results for your company name. Those are usually profile pages and press material, which keep the old description:
- Press release boilerplate. Every release published before the merger carries the old "About [company]" paragraph. Syndicated copies sit on dozens of news sites.
- Profile descriptions on LinkedIn, business databases and review sites that nobody updated.
- Your old about page, if it is still live on the former company's domain.
- Analyst and partner listings written at the time of the deal.
Fix the sources that rank. Update the about text on every profile you control. Publish new releases with the new boilerplate. Add a short editor's note to the top of recent releases on your own newsroom, such as "[Company] is now part of [platform]". Redirect the old about page. How boilerplate spreads into answers is covered in how to test whether AI assistants use your approved company boilerplate.
What should the platform page say?
State the relationship in sentences an assistant can quote without other context:
"[Platform] brings together [Product A] and [Product B], which were separate products until [year]. They are now the [module names] of one platform, with one login, one price list and one support team. Existing customers of either product have been moved to [platform] with their data."
Then say what changed and what did not: names, pricing, contracts, support, integrations and anything customers must do. If some parts still run separately during migration, say so, with the expected date. An honest "until [date], [Product B] customers log in at [URL]" is better than an answer describing a merger that has not finished.
What should happen to the old product pages?
- Redirect each old product page to its equivalent on the platform site, not to the home page.
- Keep one "formerly [Product B]" line on the module page that replaced it, so old names still resolve.
- Merge the docs into one documentation site, with redirects from old doc URLs.
- Unify pricing on one page, naming the modules.
- Retire separate sign-up flows once migration allows, or label them clearly as part of the platform.
How to handle the old names over time is covered in which website pages should keep the old product name after a product rename.
Which external profiles should you merge or update?
| Profile | Action |
|---|---|
| A Page super admin can request to merge Company Pages, provided they administer both and meet LinkedIn's criteria | |
| Review sites | Ask how they handle merged products: some combine profiles or redirect one to the other; otherwise update both descriptions |
| Cloud and app marketplaces | Update listings to name the platform, or retire the old listing with a pointer |
| Business databases | Record the acquisition and update descriptions and products |
| Wikidata | Relate the items, for example with "replaced by" for a retired product, following the site's rules on conflicts of interest |
| Wikipedia | Suggest updates through edit requests with independent sources |
| Partner directories | Ask partners to update integration names and descriptions |
In what order should you make the changes?
- Write the platform statement and put it on the platform page, the about page and the new press boilerplate.
- Redirect the old product pages and docs, once equivalent platform pages exist.
- Update the profiles that rank for each old name and the company name: LinkedIn, business databases and review sites.
- Add editor's notes to recent releases and retire old about pages.
- Brief sales, support and partners with the same statement, so what customers hear matches what assistants read.
- Retest a month after each step, so you can see which change moved the answers.
How do you test what Claude and others say?
Ask Claude, with and without web search where you can, and for comparison Perplexity, ChatGPT and Gemini, several times each:
- "What is [platform]?"
- "Is [Product B] part of [platform]?"
- "[Product A] vs [Product B]"
- "What does [company] do?"
- "Is [Product B] still available?"
Record whether each answer describes one platform, two products, or something in between. Note which sources it cites and whether the old boilerplate appears. Answers without search will change more slowly. Track them separately, so slow model updates are not mistaken for a failed fix. Other ways Claude can misdescribe a website are covered in why Claude gives vague or incorrect answers about your website.
What might an illustrative case look like?
Consider an illustrative HR software company that acquired a payroll product and merged both into one platform a year later. Claude answered "What is [platform]?" by describing the HR product and adding that "[payroll product] is a separate payroll tool owned by the same company". Perplexity described the company using its pre-merger boilerplate, citing a syndicated press release and a business database. The payroll product's old site was still live, and the platform page said "all your people operations in one place" without naming either product. The company redirected the old site page by page and rewrote the platform page's first paragraph to name both former products. It merged the LinkedIn pages, updated three databases, and added editor's notes to its recent releases. Two months later, most search-based answers described one platform.
What do people ask about merged products and AI answers?
Should we delete the old product site?
Redirect it page by page rather than deleting it. Deleted pages can make assistants conclude the product was discontinued.
Do we need to keep the old product names at all?
For a while, as "formerly" lines on the relevant module pages. Customers and assistants still use them.
How long before model knowledge catches up?
It depends on model updates. Keep search-based sources correct in the meantime, and check both kinds of answer.
Published by The Kovmere Team · updated Sep 29, 2026 · More resources