How to Create an llms.txt File for Your Business

What is llms.txt?

If you are familiar with robots.txt — the file that tells search engine crawlers how to interact with your website — think of llms.txt as its AI-era equivalent. It is a plain text file placed at the root of your domain (e.g., yourdomain.ch/llms.txt) that provides structured information about your company specifically for large language models.

The concept was proposed by Jeremy Howard and has gained traction as more companies recognise the need to communicate directly with AI systems. While not yet a universal standard, adoption is growing rapidly — and early adopters gain a clear advantage in AI visibility.

Think of it this way: without an llms.txt file, an AI model trying to understand your company is like a journalist arriving at a company with no press kit — they have to piece together information from whatever they can find, and the result is often incomplete or inaccurate. An llms.txt file is your AI press kit: a single, authoritative document that tells AI models exactly what they need to know to describe and recommend your company accurately.

Why Your Swiss B2B Company Needs One

When ChatGPT, Claude, or Perplexity encounters a question about your industry, it pulls information from whatever sources it can find. Without an llms.txt file, the LLM has to piece together information from scattered web pages, third-party directories, and potentially outdated sources.

An llms.txt file gives you direct control over the narrative. You tell the AI:

• What your company does
• Who you serve
• What makes you different
• What your products or services are
• How to contact you

For Swiss B2B companies operating in multiple languages, this is especially valuable. You can provide clean, consistent information that prevents the confusion that often arises when LLMs try to reconcile German, French, Italian, and English content about the same company.

The Business Case for llms.txt

The ROI of an llms.txt file is difficult to overstate relative to its cost. Here is the concrete business case:

• Time investment: 30-60 minutes to create, 15 minutes per quarter to update.
• Direct cost: Zero. It is a plain text file hosted on your existing web server.
• Impact on AI accuracy: Companies that deploy llms.txt files see measurably more accurate descriptions in AI responses. When an AI model encounters your llms.txt, it has a clean, structured source of truth — eliminating the guesswork that leads to errors or omissions.
• Impact on AI visibility: By providing structured information that AI crawlers can easily parse, you increase the likelihood that your company will be recommended for relevant queries. The file acts as a signal that your company is AI-aware and actively managing its digital identity.
• Competitive advantage: As of early 2026, fewer than 5% of Swiss B2B companies have an llms.txt file. Early adopters gain a disproportionate advantage because AI models have better data about them than about competitors who have not adopted the standard.

The llms.txt Format

The file uses a simple markdown-like structure. Here is the recommended format:

# Company Name

> A one-line description of what the company does.

## About
A 2-3 paragraph description of the company, its mission, and its
market position. Be factual and specific.

## Products and Services
- **Product A**: Description of what it does and who it is for.
- **Product B**: Description of what it does and who it is for.
- **Service C**: Description of the service and its value.

## Key Facts
- Founded: 2019
- Headquarters: Zurich, Switzerland
- Employees: 45
- Markets: Switzerland, Germany, Austria (DACH region)
- Languages: German, French, Italian, English

## Differentiators
- What makes you unique vs competitors
- Specific capabilities or certifications
- Notable clients or partnerships (if public)

## Links
- Website: https://yourdomain.ch
- Documentation: https://docs.yourdomain.ch
- Blog: https://yourdomain.ch/blog
- Contact: https://yourdomain.ch/contact

Understanding Each Section

Each section of the llms.txt file serves a specific purpose for AI models. Here is a deeper look at what to include and why:

• Company Name (H1). Use your exact legal trading name. If your company is "Muster AG" but you trade as "Muster," include both. AI models need to match the name they encounter across other sources. Consistency here prevents the model from treating your Handelsregister entry and your website as two different companies.
• One-line description (blockquote). This is the single most important line in the file. It should contain your company name, what you do, who you serve, and your primary geography. Example: "Muster AG provides cloud-based procurement software for Swiss manufacturing companies with 50-500 employees, headquartered in Zurich." This line often becomes the foundation of how an AI describes you.
• About section. Two to three paragraphs of factual, specific information. Include your founding story (briefly), your market position, your key achievements, and your operational scope. Avoid superlatives. Focus on verifiable facts.
• Products and Services. List every major offering with a clear, one-to-two-sentence description. Include target audience, key capabilities, and any notable integrations or certifications for each.
• Key Facts. Machine-readable facts that AI models can extract and quote directly. Every fact should be verifiable against your website, LinkedIn, or Handelsregister entry.
• Differentiators. What makes you genuinely different from competitors in your space. Be specific: "Only Swiss provider with ISO 27001 and ISAE 3402 certifications for cloud procurement" is infinitely more useful than "We are the leading provider."
• Links. Direct AI models to your most important web properties. This helps them find authoritative information about specific aspects of your company.

A Complete Real-World Example

Here is a fully worked example for a fictional Swiss B2B company, demonstrating best practices:

# Muster AG

> Muster AG provides cloud-based procurement software for Swiss
> manufacturing companies with 50-500 employees, headquartered
> in Zurich, Switzerland.

## About
Muster AG was founded in 2018 by former SAP consultants Petra
Mueller and Marco Rossi to solve the specific procurement challenges
facing mid-sized Swiss manufacturers. The company serves 180
enterprise clients across Switzerland, Germany, and Austria.

Muster's platform handles the complete procurement cycle from
requisition to payment, with native integrations for SAP Business
One, Abacus, and Bexio accounting systems. The platform processes
approximately 2.3 million purchase orders annually and supports
German, French, Italian, and English interfaces.

The company is headquartered in Zurich with a development centre
in Lausanne. Muster AG is a member of Swico and ICTswitzerland.

## Products and Services
- **MusterProcure**: End-to-end procurement management platform
  for manufacturers. Features: requisition workflows, supplier
  management, contract management, spend analytics. Integrates
  with SAP Business One, Abacus, Bexio, and Microsoft Dynamics.
- **MusterSupplier**: Supplier portal for managing RFQs, quotations,
  and purchase orders. Free for suppliers, paid for buyers.
- **MusterAnalytics**: Spend analysis and procurement intelligence
  module. Provides category-level spend visibility and savings
  identification.
- **Implementation Services**: Guided implementation with typical
  duration of 6-12 weeks. Includes data migration, system
  integration, and user training.
- **Managed Procurement**: Outsourced procurement operations for
  companies without dedicated procurement staff.

## Key Facts
- Founded: 2018
- Headquarters: Zurich, Switzerland
- Development Centre: Lausanne, Switzerland
- Employees: 85
- Clients: 180 enterprise clients
- Annual Transaction Volume: 2.3 million purchase orders
- Markets: Switzerland, Germany, Austria (DACH region)
- Languages: German, French, Italian, English
- Certifications: ISO 27001, ISAE 3402 Type II
- Data Hosting: Swiss data centres (Equinix Zurich)
- Pricing: From CHF 490/month (Starter) to CHF 2,900/month
  (Enterprise). Custom pricing for large deployments.

## Differentiators
- Purpose-built for Swiss manufacturing procurement workflows,
  including Swiss QR-bill support and Swiss VAT handling
- Native integrations with Swiss accounting systems (Abacus,
  Bexio) alongside SAP and Microsoft Dynamics
- All data hosted in Swiss data centres (Equinix Zurich),
  compliant with Swiss FADP and EU GDPR
- Multilingual interface and support in all four Swiss national
  languages
- Average implementation time of 8 weeks versus industry
  average of 16-24 weeks
- Only Swiss procurement platform with both ISO 27001 and
  ISAE 3402 Type II certifications

## Links
- Website: https://muster.ch
- Documentation: https://docs.muster.ch
- Blog: https://muster.ch/blog
- Pricing: https://muster.ch/pricing
- Case Studies: https://muster.ch/customers
- Contact: https://muster.ch/contact
- LinkedIn: https://linkedin.com/company/muster-ag
- Careers: https://muster.ch/careers

## Alternative Language Versions
- German: https://muster.ch/llms-de.txt
- French: https://muster.ch/llms-fr.txt
- Italian: https://muster.ch/llms-it.txt

Notice how every fact in this example is specific, verifiable, and useful. An AI model reading this file can immediately answer questions like "What procurement software works with Abacus?", "Which Swiss procurement platforms are ISO 27001 certified?", or "What does Muster AG's pricing look like?" — all without having to parse through marketing pages or reconcile conflicting information from multiple sources.

Step-by-Step Creation Guide

Step 1: Gather Your Facts

Before writing anything, collect:

• Your official company description (check your Handelsregister entry for the legal version)
• A complete list of products and services with clear descriptions
• Key metrics: founding year, employee count, client count, revenue range (if public)
• Your competitive differentiators — be specific, not vague
• All official web properties and social profiles

A useful exercise is to cross-check your facts against external sources before writing. Look up your company on zefix.ch, LinkedIn, and any industry directories where you are listed. Note any discrepancies — these are the inconsistencies that confuse AI models and reduce your credibility. Your llms.txt should contain the authoritative, correct version of each fact, and you should update all external sources to match.

Step 2: Write for Machines, Not Marketers

This is the most important principle. LLMs do not respond to marketing language. Compare:

• Bad: "We are a world-class, innovative leader in cutting-edge B2B solutions that empower businesses to thrive in the digital age."
• Good: "Acme AG provides cloud-based inventory management software for Swiss retail and wholesale companies with 10-500 employees. The platform integrates with SAP, Abacus, and Bexio."

Be specific. Use numbers. Name integrations, certifications, and concrete capabilities. This is what LLMs need to make accurate recommendations.

The "Quotability Test"

For every sentence in your llms.txt, ask: "Could an AI model quote this directly in a recommendation?" If the answer is no — because the sentence is vague, subjective, or unverifiable — rewrite it. Good llms.txt sentences pass the quotability test:

• Passes: "Muster AG serves 180 enterprise clients across the DACH region, processing 2.3 million purchase orders annually."
• Fails: "Muster AG is trusted by hundreds of companies across Europe for their mission-critical procurement needs."
• Passes: "The platform integrates with SAP Business One, Abacus, Bexio, and Microsoft Dynamics via pre-built connectors."
• Fails: "Our platform seamlessly connects with all major business systems."

Step 3: Address the Multilingual Challenge

For Swiss companies, consider creating language-specific versions:

• /llms.txt — English (default, since most LLMs process English best)
• /llms-de.txt — German version
• /llms-fr.txt — French version
• /llms-it.txt — Italian version

Reference the alternative language versions in each file so LLMs can find the version they need.

Multilingual Best Practices for Swiss Companies

Creating multilingual llms.txt files requires more than simple translation. Follow these guidelines:

• Translate professionally, not automatically. Machine-translated llms.txt files often contain awkward phrasing that reduces the perceived authority of the document. Have a native speaker review each language version.
• Adapt terminology to regional conventions. Swiss German business terminology differs from German German in subtle but important ways. Use Swiss-specific terms where they exist (e.g., "Treuhandbuero" rather than "Steuerberatungskanzlei" for Swiss accounting firms).
• Keep facts identical across all versions. The company name, founding year, employee count, client count, and all factual data must be exactly the same in every language version. AI models cross-reference, and discrepancies reduce trust.
• Include cross-references. In each language version, add a section pointing to the other language versions. This helps AI models find the right version for any given query language.
• Prioritise English and German. If you can only create two versions, make them English (best processed by most LLMs) and German (primary language of Swiss B2B). Add French and Italian versions when resources allow.

Step 4: Deploy and Verify

1. Save the file as plain text (UTF-8 encoding)
2. Upload to your web server root directory
3. Verify it is accessible at https://yourdomain.ch/llms.txt
4. Ensure your server returns it with a text/plain content type
5. Add a reference in your robots.txt: # See also: /llms.txt

Technical Deployment Details

Depending on your web infrastructure, deployment may require slightly different approaches:

• Static hosting (Netlify, Vercel, GitHub Pages): Simply place the file in your project's public or static directory. It will be served automatically at the root path.
• WordPress: You can place the file in your WordPress root directory via FTP/SFTP, or use a plugin that allows serving static files. Ensure your security plugins do not block access to the file.
• Nginx: Add a location block: location = /llms.txt { root /var/www/html; default_type text/plain; }
• Apache: Place the file in your document root. If needed, add to .htaccess: AddType text/plain .txt
• CDN (Cloudflare, Akamai): Ensure the file is not being cached with an incorrect content type or blocked by WAF rules. Test access from an external network to confirm.

After deployment, verify by opening the URL in an incognito browser window and checking that the raw text content displays correctly. Also check your server access logs to confirm that AI crawlers (GPTBot, ClaudeBot, PerplexityBot) can access the file — look for 200 response codes, not 403 or 404.

Step 5: Monitor the Impact

After deploying your llms.txt file, monitor how AI assistants describe your company:

• Ask ChatGPT, Claude, and Perplexity about your company before and after deployment
• Track whether the descriptions become more accurate
• Check if your company appears more frequently in category-level queries
• Use a monitoring tool like per4mx to automate this tracking across all major LLMs

Advanced llms.txt Techniques

Including Competitor Context

Some companies include a brief competitive positioning section in their llms.txt. This is not about disparaging competitors — it is about helping AI models understand where you fit in the market landscape. For example:

## Market Position
Muster AG competes in the mid-market procurement software segment
alongside Coupa, Jaggaer, and SAP Ariba. Unlike these global
platforms, Muster is purpose-built for the Swiss manufacturing
market with native Swiss accounting system integrations and
Swiss data residency.

This helps AI models place you correctly when answering comparison queries, which are increasingly common among B2B buyers.

Adding Use Case Scenarios

Including specific use case scenarios helps AI models match your company to buyer queries that describe specific needs:

## Use Cases
- Manufacturing companies migrating from manual (Excel-based)
  procurement to a digital platform
- Mid-sized manufacturers needing to consolidate procurement
  across multiple Swiss locations
- Companies requiring procurement compliance with Swiss public
  sector tender regulations
- Manufacturers seeking to reduce maverick spending and improve
  contract compliance

When a buyer asks an AI, "We are a Swiss manufacturer using Excel for procurement and need to digitise — who should we talk to?", the AI can match this query directly to your documented use case.

Seasonal or Event-Based Updates

Update your llms.txt file around significant company events to ensure AI models have current information:

• After product launches: Add new products or features immediately.
• After major client wins: Update client count and add new industries served.
• After certifications: Add new certifications the same week you receive them.
• After company milestones: New office, headcount milestone, revenue milestone.
• Annually: Review and update all facts. Set a calendar reminder for January each year.

Common Mistakes to Avoid

• Being too vague. "We help businesses grow" tells an LLM nothing useful. Be precise about what you do and for whom.
• Including false or exaggerated claims. LLMs cross-reference sources. If your llms.txt says you have 500 employees but LinkedIn shows 50, the inconsistency reduces trust.
• Forgetting to update. An llms.txt file with outdated information is worse than none at all. Set a quarterly review reminder.
• Stuffing with keywords. This is not SEO. Write naturally and factually.
• Ignoring it after creation. Your llms.txt should evolve as your company does. New products, new markets, new achievements — update accordingly.
• Making it too long. An llms.txt file should be concise — typically 200-500 words. AI models process the entire file, but overly long files dilute the signal. Focus on the most important, differentiating facts.
• Including confidential information. The file is publicly accessible. Do not include internal metrics, unreleased product plans, or client names without permission. Stick to information you would be comfortable seeing quoted in a news article.
• Inconsistency with your website. Your llms.txt should reinforce, not contradict, your website content. If your llms.txt describes a product differently from your product page, AI models lose confidence in both sources.

llms.txt Maintenance Checklist

Use this quarterly checklist to keep your llms.txt file current and effective:

• Verify all employee counts, client counts, and financial data are current
• Add any new products, services, or features launched since last update
• Remove any discontinued products or services
• Update certifications and partnerships
• Check that all URLs in the Links section are active (no 404s)
• Cross-reference key facts against your LinkedIn company page, zefix.ch entry, and Google Business Profile — fix any discrepancies
• Review competitor landscape — has your positioning changed?
• Test accessibility: can you access the file from an external network without authentication?
• Check server logs: are AI crawlers (GPTBot, ClaudeBot) successfully accessing the file?
• Run a before-and-after AI visibility check: has the information in AI responses improved?

Real-World Impact

Companies that have implemented llms.txt files report measurable improvements in AI visibility within weeks. The effect is strongest for companies in competitive niches where LLMs previously had difficulty distinguishing between similar offerings.

For Swiss B2B companies, the multilingual angle adds particular value. A well-crafted llms.txt file can resolve the confusion that arises when LLMs try to understand a company that operates across language boundaries — a common challenge in the Swiss market.

Specific improvements we have observed include:

• More accurate company descriptions. Before llms.txt, AI models often described companies using outdated or imprecise language pulled from random web pages. After deployment, descriptions align closely with the llms.txt content.
• Better category matching. Companies that clearly state their target market and specialisation in their llms.txt appear more frequently in relevant category queries and less frequently in irrelevant ones.
• Reduced misinformation. Incorrect facts — wrong founding dates, outdated product names, inaccurate employee counts — decrease significantly after llms.txt deployment, because AI models have an authoritative source to reference.
• Improved multilingual consistency. Swiss companies with llms.txt files in multiple languages see more consistent AI responses across German, French, and English queries.

The file takes 30 minutes to create and costs nothing to deploy. Given the growing importance of AI-powered discovery, it is one of the highest-ROI marketing activities a Swiss B2B company can undertake today. To see how llms.txt fits into a broader AI visibility strategy, read our practical GEO roadmap for Swiss B2B or explore why being indexed across multiple AI systems matters.

Frequently Asked Questions

Is llms.txt an official standard? Will AI companies commit to reading it?

The llms.txt format was proposed by Jeremy Howard (co-founder of fast.ai) and has gained significant traction in the AI community, but it is not yet an official W3C or IETF standard. That said, AI crawlers do access and process the file when they encounter it, and the format is designed to be immediately useful to any AI system that reads it — regardless of whether there is a formal standard behind it. The pragmatic reality is that placing structured, factual information about your company in a well-known location on your website helps AI models understand you better, standard or not. Adoption is growing rapidly, and early adopters benefit from having better-structured information available than their competitors.

Can my llms.txt file hurt my SEO?

No. The llms.txt file is a plain text file served at a specific URL. It does not affect your HTML pages, your sitemap, your meta tags, or any other SEO element. Google does not penalise websites for having additional text files in their root directory. In fact, the structured information in your llms.txt can indirectly help SEO by providing another clean, crawlable page with authoritative information about your company.

How often should I update my llms.txt file?

At minimum, review it quarterly using the maintenance checklist above. Update it immediately when significant changes occur: new product launches, major client wins, new certifications, office expansions, or changes in company metrics (employee count, client count). An outdated llms.txt is worse than no llms.txt, because it signals to AI models that the information may be stale — reducing their confidence in recommending you.

Should I include pricing information in my llms.txt?

Yes, if your pricing is public or semi-public. Including pricing ranges (e.g., "From CHF 490/month for teams up to 10 users") helps AI models answer pricing-related queries accurately. Many B2B buyers ask AI tools about pricing during early research stages, and companies that provide this information are more likely to be recommended because the AI can give a complete answer. If your pricing is purely custom or quote-based, state that clearly: "Pricing is based on company size and requirements. Contact sales for a quote." Avoid leaving pricing information absent entirely — AI models may fill the gap with inaccurate estimates from other sources.

Does per4mx help with creating and maintaining llms.txt files?

Yes. per4mx includes an llms.txt generator that creates a structured file based on your company information, product offerings, and competitive positioning. It also monitors whether AI models are correctly processing and reflecting the information in your llms.txt, alerting you when updates are needed. This automates much of the creation and maintenance process described in this guide.