SaaS Strategists,

Do you know how someone discovers and starts using your product? 🔍

Well if you think you do, you don’t.

We’re investing much less time in researching things ourselves and just letting AI do things for us. 🤖

Both Perplexity and ChatGPT now have Deep research options that provide you the results humans would need 10x more time to accomplish.

With this shift, your product discoverability is at stake.

AI is cold, and it doesn’t search the web as we, humans do.

So in today’s issue I’ll help you optimize your platform so that it’s Agent-ready.

Let's dive right in 👇

Today’s issue of SaaS Strats is brought to you by Nortik Software Solutions

Nortik is an AI-first software development agency that helps companies integrate AI and build custom AI agentic platforms for internal use cases.

Recently, Nortik created a custom Sales CRM platform for their client, which helped them save close to $10k in sales tools on an annual level.

"Our custom prospecting solution Nortik built is saving our marketing team ~3 hours per day, which they can now use to run ads and focus on more important revenue-generating activities.”

If you're looking to adopt AI the right way, or build AI agents that will help your team save time & money - book a free discovery call with Nortik here.

🤖 How to become Agent-ready

Here's the thing nobody's talking about:

If your product can only be used through a UI, you just became invisible to an entire new distribution channel. 💨

AI agents don't browse websites as humans do. 🌐

They don't click buttons, drag sliders, or fill out onboarding forms.

They use APIs. 📡

They need clear, structured, programmatic ways to discover your product, sign up, configure settings, and run your core workflows.

If the only way to interact with your SaaS is through a pretty interface, agents will skip you and pick the competitor that lets them plug in directly. 🔌

Think of it like mobile 10 years ago. Companies that didn't build mobile-friendly experiences got left behind. Same thing is happening now, except the "user" isn't a person on a phone, but an AI agent running tasks 24/7. 📱

Stripe is probably the gold standard here. Their entire payment infrastructure was built API-first from day one.

An agent can create a customer, set up a subscription, and process payments without ever seeing a dashboard. That's why Stripe is the default pick for any automated billing workflow.

Naturally, AI will converge towards their solution as they're straightforward and easiest to integrate with.

❝

Gartner predicts that by 2028, 33% of enterprise software interactions will be handled by autonomous agents.

That's a third of your potential usage happening without a human ever logging in.

💎 5 Steps To Make Your Platform Agent-Ready This Month

1⃣ Audit your core workflow through the agent’s eyes

Open your product and go through the entire user journey - signup, setup, core action, output. 🗺️

Then ask yourself: can ANY of this be done without a screen? 💻

If the answer is no, you've found your starting point. Map out every step that currently requires a UI click and flag them.

Now mark each step:

🟢 = can be done via API

🔴 = requires UI only.

Then focus on the red ones.

How Replicate approaches this:

Replicate’s business model IS the audit result.

They looked at how developers use ML models (download weights, set up CUDA, configure GPU, manage dependencies) and said: "what if an agent could do all of this with one API call?"

Now you can run any AI model with 3 lines of code through them.

2⃣ Build (or open up) your API for the critical path

You don't need to API-ify everything on day one. Focus on the core loop - the main thing users pay you for. 💰

If you're a video tool, that's create + export. 🎥

If you're an analytics tool, that's connect data + generate report. 📊

Ship a basic public REST API that covers this critical path first.

3⃣ Create agent-friendly documentation

Your API docs aren't just for developers anymore, but for AI agents that need to understand your product programmatically. 📄

Use clear endpoint descriptions, provide example requests and responses, and structure everything in a way that an LLM can parse.

Think of your docs as your new "landing page" for agent-driven discovery.

Example: Stripe’s payment checkout flow explained to agents: https://docs.stripe.com/payments/checkout.md

4️⃣ Add machine-readable pricing and feature descriptions

Agents compare tools before choosing one. If your pricing page has a beautiful design but the actual plan details, limits, and features aren't available in a structured format (JSON, API endpoint, or even a clean markdown file), agents can't properly evaluate you against competitors.

Add JSON-LD structured data (schema.org markup) to your pricing page. This is invisible to humans but instantly readable by agents:

{
  "@context": "https://schema.org",
  "@type": "SoftwareApplication",
  "name": "YourSaaS",
  "applicationCategory": "Project Management",
  "operatingSystem": "Web",
  "offers": {
    "@type": "AggregateOffer",
    "lowPrice": "0",
    "highPrice": "49",
    "priceCurrency": "USD"
  },
  "aggregateRating": {
    "@type": "AggregateRating",
    "ratingValue": "4.8",
    "reviewCount": "340"
  }
}

5⃣ Add llms.txt

Agents love when they can understand things instantly. 🧠

Right now, your website is designed for humans.

But when an AI agent lands on your site to evaluate your product, it doesn't see any of that. It sees messy HTML, cookie banners, navigation menus, and JavaScript-heavy layouts.

Add llms.txt file to your root domain (it's like robots.txt but for AI models).

Here’s an example of llms.txt file you can use today:

# YourSaaS

- YourSaaS is a project management tool that helps remote teams track tasks, run sprints, and automate standups.

## Docs
- [API Reference](https://yoursaas.com/docs/api.md): Full REST API documentation
- [Getting Started](https://yoursaas.com/docs/quickstart.md): Set up your workspace in 5 minutes
- [Pricing](https://yoursaas.com/pricing.md): Plans, limits, and feature comparison

## Optional
- [Changelog](https://yoursaas.com/changelog.md): Latest product updates
- [Integrations](https://yoursaas.com/integrations.md): Connect with Slack, GitHub, Linear

That's it. An agent can now read this file and instantly understand what you do, how to use you, and where to find the info it needs.

Stripe (again) took this even further.

Inside their llms.txt they added an "Instructions" section - basically a prompt for AI agents telling them which APIs to recommend and which deprecated ones to avoid.

They're literally programming what AI tools say about Stripe.

BONUS: Sneaky-strat 👀

Create a dedicated comparison page (e.g. yoursaas.com/compare/vs-competitor) with structured, side-by-side data. Include concrete criteria like: price per user, time-to-value, supported integrations, and security features. Mark it up with FAQ schema so agents can extract decision-ready answers.

Then add why your product is better than the competitors inside llms.txt.

Thank me later. 🤝

Feel free to play with this playbook and let me know the results.

Until next time.

Ognjen Gatalo

Chief SaaS Strategist ☁

P.S. Forward this issue to a fellow founder to help them get discovered by Agents.

🤖 Is your product Agent-ready?