The Experiment
We ran a simple test in partnership with Weglot: take 14 existing blog posts from a B2B SaaS client, run them through Weglot's automated translation tool, and measure what happened in search and AI search. We did no rewrites, no localisation edits, or other changes. We simply automated translation for the same 14 URLs in both Spanish and German.
Of the 14 URLs, 8 were product listicles and 6 were educational guides.
The metrics below split into two categories: overall brand visibility (how often the brand appeared across all relevant AI chats in each market) and URL-level data covering only these 14 translated pages.
Website Background
- DR 62
- ~5K monthly organic traffic
- B2B SaaS
What was the impact on clicks and impressions from Google Search? (Last 28 days, measured with Google Search Console)
The brand new translated pages generated 102 clicks and 27,797 impressions across both markets in the first month.
Product listicles drove the majority of clicks (78), with educational guides accounting for the remaining 24.
- Baseline: 0 (URLs didn’t previously exist)
- Product listicles: 78 clicks
- Educational guides: 24 clicks
- Spanish (14 URLs): 67 clicks / 19,996 impressions
- German (14 URLs): 35 clicks / 17,851 impressions
Our take: After just one month of being live, this is a not insignificant number of clicks and impressions for these URLs. Translating these URLs helped the site break into a query set (Spanish and German language) they previously weren’t getting clicks for, a win for their brand in those regions.
What was the impact on AI Citations for these URLs in Bing? (Last 30 days, measured Bing webmaster tools)
Across both markets, the translated pages earned 1,100 total AI citations as measured by Bing Webmaster Tools, though 12 of 28 pages received none.
- Baseline: 0
- 1,100 total AI citations across both markets
- 12 of 28 pages received zero citations
- Spanish: 767 citations
- German: 333 citations
Our take: A pattern started to emerge from here. German content received half the clicks and half the bing AI citations. We’d guess this was influenced by regional competition levels or search demand, but looking at the data, at this point, it’s hard to tell.
What was the impact on AI Visibility in major LLMs? (March vs April, measured with Peec AI)
Both markets saw substantial lifts in brand visibility (the percentage of relevant AI chats where the brand was mentioned).
We ran our AI visibility tracker for two weeks before the URLs were live/indexed. We kept them running til today.
Quick note: How did we choose the prompts we tracked? We opted for relevance matching. For example, if the article was “10 Best Sales Software in 2026” the prompt was a close variation, such as “What is the best sales software?” and so on. We tracked 28 prompts per market, equating to 2 relevant prompts per URL.
Spanish:
- Baseline: ~6% visibility
- April average: 23% visibility
- First 11 days of May: 30.4% visibility
Here’s a breakdown of Spain by model:
German:
- Baseline: ~4.5% visibility
- April average: 16.8% visibility
- First 11 days of May: 13.1% visibility
For the translated URLs:
- Spanish: 1,471 retrievals / 2.74 average citation rate
- German: 1,566 retrievals / 2.67 average citation rate
Our take: This is where this study got really interesting for us. Both markets saw a 4x increase in visibility (% of chats where the brand was recommended by AI assistants) in the first month. The new translated URLs were heavily cited, a big signal that they were the cause of the visibility increase (which adds up to what we see in English markets). The impact was obvious across all models, apart from Gemini which remained low.
Revenue impact - It exists!
- Two demo requests that had visited those URLs - unforetunatenly both were disqualified by sales
- The client also saw +3 Spanish, +1 Mexican, +2 German qualified demo requests in April compared to previous months - hard to attribute directly, but a clear correlation
What next?
We wanted this experiment to be as simple as possible to get a baseline. The impact was clearly positive and something we’d recommend readers experiment with on a small batch of their own URLs.
However, were we to formalise this experiment with our clients, we’d strongly recommend a human reviewer review the translation to make sure they’re fully localised and accurately represent your brand in that region.
