Skip to content
AdsChatGPT now shows ads.Get early access

November 16, 2026

How Payr Grew Its Gemini Visibility by 13 Points in One Week

Robin Pautigny

Robin Pautigny

Co-founder, Refine

Summary

Payr lets UK tenants pay their rent by credit card, to earn rewards, spread the cost or pay from abroad, while landlords get paid as usual. More and more of those tenants now start with a question to ChatGPT or Gemini rather than a Google search. Within three weeks of starting with Refine, Payr knew where it stood in those answers, who it was really competing against and which questions it was losing. One week after publishing its first targeted pages, Gemini mentioned Payr in 41% of its answers, up from 28%.

Key Figures

  • +13 points of visibility on Gemini in one week (28% → 41% of answers).
  • 1 in 3 AI answers about paying UK rent by card now mention Payr, up from 1 in 4.
  • 3 targeted answer pages live within 72 hours.

Industry: fintech, rent payments. Market: United Kingdom. Buyers: tenants and international students. Started with Refine: October 2026.

Why this matters

“Can I pay my rent with a credit card in the UK?” is exactly the kind of question people now ask an assistant. The assistant answers with two or three names. If yours is not one of them, the tenant never reaches your site, and the lost visit never shows up in your analytics.

The Challenge

Payr had no way to see what AI assistants said about rent payments. Three blind spots stood out:

  • No measurement. Nobody knew how often Payr was recommended, on which engine, or for which questions.
  • The wrong competitors. The benchmark list was built around US players (RentMoola, Plastiq, RadPad, Zego, PayLease). In UK answers, four of them were never mentioned once.
  • Unknown gaps. High-intent questions (international students, paying from abroad, large upfront deposits) were answered without Payr, and the team had no list of them.

The part most teams miss: AI answers are not stable

Ask the same question twice and you can get a different shortlist. A single manual check says almost nothing. Only daily, repeated measurement across engines shows a real trend.

The Approach

1. Build the question set from real tenant situations

We wrote 30 questions from the tenant’s point of view: paying by card when the landlord does not accept cards, earning Avios or Amex points on rent, hitting a credit card bonus, paying UK rent from China or with a foreign card, paying 6 or 12 months upfront, and the B2B side (letting agents, build-to-rent, serviced accommodation).

2. Measure every day, on every engine

Each question runs daily on ChatGPT, Gemini and Claude. Hundreds of answers a week gave a reliable baseline: Payr appeared in about 1 in 4 answers on generic questions, strong on ChatGPT, weaker on Gemini, absent on Claude.

3. Find who the engines actually recommend

Refine extracts every brand named in the answers. The real competitive set turned out to be UK players (Rentd, Wise, Paymentflux, Ribbon Rewards, Rentora), not the US list. The competitor dashboard was rebuilt around them.

4. Publish answers where Payr was missing

Refine flagged the questions where Payr was rarely mentioned and generated answer pages for them, written to be quoted by AI engines, with clear facts and sources. Three went live within 72 hours: paying UK rent as an international student with a foreign card, the cheapest way to pay UK rent from abroad, and paying a large deposit upfront.

What Refine did

Daily tracking of 30 questions on 3 engines, automatic competitor discovery, gap detection, and answer pages written and published for the team, with no extra headcount on Payr’s side.

Results

  • Gemini answers mentioning Payr: 28% → 41%.
  • Gemini answers citing payr.com as a source: 16% → 22%.
  • All engines, Payr’s original questions: 25% → 34%.
  • All engines, full question set: 30% → 36%.

Before: 6 to 10 November 2026 (25 tenant questions; 27 October to 5 November for the original questions). After: 11 to 16 November 2026. Payr is counted when it is named in the answer text.

  • Visibility up where it was weakest. Gemini, Payr’s weakest engine at the start, moved the most.
  • More often a source, not just a name. Gemini increasingly links to payr.com itself.
  • A competitor view that matches reality. The team now tracks the brands tenants actually hear about.
  • A clear next target. Claude does not mention Payr yet. The next round of content focuses on it, and on the questions where Rentd and Wise still lead.

Why It Worked

  • Measure before producing. Content went to the exact questions where Payr was missing, not to generic topics.
  • Fix the competitive picture first. Knowing that Rentd and Wise were the real rivals changed what the pages had to say.
  • Speed. From first measurement to published pages in three days, then a daily read on whether engines pick them up.

The takeaway

For a product people discover by asking a question, AI answers are the new shelf. You cannot optimise what you do not measure. Once you do, a week is enough to see the needle move.

Short on time? Have an assistant summarise this page for you.