getpliant.com — AI Site Grade

Analysis

Pliant (getpliant.com) — AI-Visibility Audit

Cold LLM knowledge describes Pliant as a US-focused SMB rewards card earning 3x points on software, advertising, and shipping — but the actual site is a Berlin-based B2B card infrastructure platform with four distinct product lines, zero mention of that rewards structure, and no US consumer-style points program anywhere.

Crawler Access

All 11 AI crawlers tested (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User, anthropic-ai, Bytespider, Applebot-Extended, Perplexity-User) receive 200 with full HTML content identical to browser baseline — 685,961 bytes served from Vercel with no UA-based blocking. The robots.txt is minimal: a single wildcard rule allowing / and disallowing /api/ and /internal, with zero AI-bot-specific directives. No llms.txt exists (404). The sitemap contains 1,928 URLs across 10+ locale variants, indicating heavy i18n duplication that crawlers must navigate.

Cold-Knowledge Gap

The LLM's prior knowledge is fundamentally misaligned with the live site. The model describes a consumer-style rewards card ("3x points on software, advertising, and shipping, 2x on other business services, 1x on everything else") with no annual fee and QuickBooks/Xero integration. The actual Pliant offers no such points structure. The site promotes cashback (unspecified rates), Visa Infinite premium perks (lounge access, travel insurance), and a 2% FX fee. The model also claims Pliant was "founded by former Brex employees" — the site shows founders Malte Rau and Fabian Terner with backgrounds in "Fintech and banking" but never mentions Brex. The core business — Cards-as-a-Service (CaaS), Pro API, and Card & Spend OS for banks and enterprises — is entirely absent from the model's prior. An AI engine retrieving this site cold would contradict itself.

Schema Posture

The homepage carries a single OnlineBusiness JSON-LD block with address, founders, founding date (2020-05-06), employee count (120), VAT ID, and contact points in 9 languages. However, every sub-page — /payment-apps, /cards-as-a-service, /about-pliant, /blog, feature pages — has zero JSON-LD schema. No Product, FAQPage, SoftwareApplication, Organization (beyond homepage), or Article schema exists anywhere on the deeper site. The FAQ sections on /payment-apps and /cards-as-a-service are plain HTML with no FAQPage markup, meaning AI engines cannot extract structured Q&A.

Content & Structure

The homepage is a Next.js prerendered page (server-rendered, not a JS shell) with ~920 words of visible text. Heading structure is flat: a single H1 ("Payments optimised" or "Payments optimized" per locale) followed by repeated H2/H3 groups. The four product lines (Payment Apps, Pro API, CaaS & BaaS, Card & Spend OS) are clearly delineated. The blog contains 67 posts covering travel, marketing agencies, embedded finance, and tech — strong topical depth. The site redirects getpliant.com to www.getpliant.com/en-gb?set-country=gb via geolocation, creating a canonical chain that may confuse crawlers expecting a locale-neutral root.

External Signals

DNS records reveal OpenAI domain verification (openai-domain-verification=dv-...) and Anthropic domain verification (anthropic-domain-verification=...), confirming the site has proactively registered with both major AI providers for retrieval access. Additional verification tokens for HubSpot, Stripe, Google (7 keys), Apple, Facebook, LinkedIn, Atlassian, Notion, Figma, Miro, Cursor, and Zoho indicate a broad SaaS integration stack. The site runs on AWS Route53 (NS records) with Google Workspace mail. No negative press or Reddit threads surfaced in search.

Findings

1. LLM cold knowledge describes a US consumer rewards card, not the actual B2B platform High

   The LLM prior knowledge describes a 3x points rewards structure, no annual fee, and Brex-founder background, but the live site offers cashback, Visa Infinite perks, and founders with no Brex mention. The core CaaS, Pro API, and Card & Spend OS products are absent from the model's knowledge.

   What to change: Update the site's content and structured data to explicitly describe the B2B platform, product lines, and cashback model. Consider publishing an llms.txt file to provide authoritative context for AI crawlers.

2. Sub-pages lack any JSON-LD schema High

   Only the homepage has an OnlineBusiness JSON-LD block. All other pages — including product pages, about page, and blog — have zero structured data. FAQ sections are plain HTML without FAQPage markup, preventing AI engines from extracting structured Q&A.

   What to change: Add Product, SoftwareApplication, FAQPage, and Article schema to relevant sub-pages. Mark up FAQ sections with FAQPage markup.

3. Geolocation redirect creates canonical chain that may confuse crawlers Medium

   The root domain redirects to a locale-specific path (e.g., /en-gb?set-country=gb) based on geolocation. This creates a chain that may cause crawlers to index locale-specific URLs instead of a canonical neutral version.

   What to change: Implement a canonical URL strategy that points to a locale-neutral version (e.g., www.getpliant.com) or use hreflang annotations to clarify locale variants.

4. No llms.txt file published Medium

   The site returns a 404 for /llms.txt, missing an opportunity to provide AI crawlers with a curated summary of the site's content and structure.

   What to change: Create an llms.txt file that describes the site's purpose, product lines, and key pages to guide AI crawlers.

5. Robots.txt lacks AI-bot-specific directives Low

   The robots.txt has a single wildcard rule allowing / and disallowing /api/ and /internal, with no explicit rules for AI crawlers like GPTBot or ClaudeBot. While all tested bots currently get full access, the lack of explicit directives may lead to uncertainty as AI crawler behavior evolves.

   What to change: Add explicit directives for known AI crawlers (e.g., GPTBot, ClaudeBot) to clarify allowed paths.

6. Sitemap contains heavy locale duplication with 1,928 URLs Medium

   The sitemap lists 1,928 URLs across 10+ locale variants, which may dilute crawl budget and cause indexing of duplicate content.

   What to change: Use hreflang annotations and consider consolidating locale variants or using a single canonical URL per page.

7. FAQ sections lack FAQPage schema Medium

   Pages like /payment-apps and /cards-as-a-service contain FAQ sections in plain HTML, but no FAQPage JSON-LD markup is present, preventing AI engines from extracting structured Q&A.

   What to change: Add FAQPage schema to all pages with FAQ content.

8. Blog posts lack Article schema Medium

   The blog page lists 67 posts, but no Article or BlogPosting schema is present, reducing the chance of appearing in AI-generated summaries or rich results.

   What to change: Add Article or BlogPosting schema to each blog post page.

9. Flat heading structure with single H1 and repeated H2/H3 groups Low

   The homepage uses a single H1 followed by multiple H2 and H3 headings, but the structure is flat and may not clearly convey content hierarchy to crawlers.

   What to change: Improve heading hierarchy by using nested H2/H3 to reflect content sections more clearly.

What's working

• All 11 AI crawlers receive full HTML content with no blocking — Every tested AI crawler (GPTBot, ClaudeBot, PerplexityBot, etc.) receives a 200 response with full HTML content identical to the browser baseline. No UA-based blocking is in place.
• OpenAI and Anthropic domain verification tokens present — DNS TXT records include verification tokens for both OpenAI and Anthropic, indicating proactive registration for AI retrieval access.
• Homepage has OnlineBusiness JSON-LD with detailed business info — The homepage includes a comprehensive OnlineBusiness schema block with address, founders, founding date, employee count, VAT ID, and multilingual contact points.
• Pages are server-rendered with full HTML content, not JS shells — All tested pages return substantial HTML content (324–4,311 words) with no reliance on client-side JavaScript for rendering, ensuring crawlers can index text.
• Blog covers diverse topics relevant to target audience — The blog contains 67 posts covering travel, marketing agencies, embedded finance, and tech, providing topical depth that can attract AI citations.
• DNS verification tokens for multiple SaaS platforms indicate broad integration — TXT records include verification tokens for HubSpot, Stripe, Google, Apple, Facebook, LinkedIn, Atlassian, Notion, Figma, Miro, Cursor, and Zoho, suggesting a robust integration ecosystem.