Tagged: AI

Unlock AI-Ready Experiences with Optimizely

Over the past few months, almost every customer conversation has shifted from SEO to AI readiness.

The questions are no longer just:

“How do we rank higher on Google?”

They’re becoming:

“How do we make sure ChatGPT, Copilot, Gemini, or Claude can find and understand our content?”

If you’re running an Optimizely website, the good news is that you don’t need a complete redesign. A few small improvements can make your content much easier for AI systems to discover and consume.

In this post, I’ll walk through a few simple changes we’ve been discussing with customers, including:

  • Upgrading the Stott Security RobotHandler
  • Adding an llms.txt file
  • Structuring content for AI
  • Improving metadata
  • Preparing your Optimizely implementation for the next generation of search

Why AI Readiness Matters

Search engines return links.

LLMs return answers.

That means AI systems need to understand your content, not just crawl it.

If your website has:

  • messy content
  • outdated metadata
  • poor internal linking
  • no structured information

then AI assistants have a harder time understanding what your organization actually offers.

Fortunately, most of these improvements are things we should already be doing.

Step 1 – Upgrade Stott Security RobotHandler

We are currently using Stott.Security.Optimizely.RobotsHandler 3.x – CMS PaaS 12 

There isn’t necessarily anything wrong with version 3, but if you’re actively maintaining your solution, it’s worth upgrading to the latest release to benefit from ongoing fixes, compatibility improvements, and newer Optimizely support.

Updating is straightforward.

Remove the old package:

dotnet remove package Stott.Security.Optimizely.RobotsHandler

Install the latest version:

dotnet add package Stott.Security.Optimizely.RobotsHandler

After upgrading, verify that:

  • robots.txt is still generated correctly
  • your sitemap is referenced
  • custom rules still work
  • different environments behave as expected

It only takes a few minutes but keeps an important part of your site’s crawl configuration up to date.

Step 2 – Add an llms.txt File

You have probably heard about llms.txt recently.

Think of it as the AI equivalent of a sitemap.

While it isn’t an official web standard today, more organizations are publishing one to help AI systems discover their most important content.

A simple example looks like this:

# Contoso

Website:
https://contoso.com

Documentation:
https://contoso.com/docs

Products:
https://contoso.com/products

Support:
https://contoso.com/support
 

For smaller websites, a static file is perfectly fine.

For enterprise Optimizely implementations with thousands of pages, I’d recommend generating it dynamically so it always reflects your published content.

That’s something I’ll cover in a future post.

Step 3 – Structure Your Content

This is probably the biggest opportunity.

Many websites still publish pages that contain multiple unrelated topics, huge blocks of text, and headings like:

Learn More

AI models love structure.

Instead, organize pages with:

  • meaningful headings
  • summaries
  • FAQs
  • bullet lists
  • tables where appropriate

Good content structure helps humans and AI alike.

Step 4 – Don’t Ignore Metadata

Metadata is still important.

Make sure every page includes:

  • Page title
  • Meta description
  • Canonical URL
  • Open Graph tags
  • Schema.org structured data

Structured data gives AI additional context about your pages, products, articles, events, and organization.

Step 5 – Review Your Sitemap

Your XML sitemap should always be current.

Remove:

  • unpublished pages
  • redirects
  • expired content

Include:

  • important landing pages
  • articles
  • product pages
  • documentation

This benefits traditional search engines and AI crawlers alike.

Step 6 – Improve Internal Linking

One thing that’s often overlooked is internal linking.

If every page exists in isolation, AI has to work harder to understand the relationships between your content.

Instead, connect related pages naturally.

For example:

  • Blog → Product
  • Product → Documentation
  • Documentation → FAQ
  • FAQ → Support

These relationships help both users and AI navigate your website more effectively.

Step 7 – Expose Content Through APIs

If you’re building AI-powered experiences, don’t stop with web pages.

Optimizely already provides excellent options for exposing structured content, including:

  • Content Delivery API
  • Content Graph
  • Search & Navigation indexes

These services make it much easier to build Retrieval-Augmented Generation (RAG) applications and AI assistants that provide grounded responses based on your content.

Most of these are relatively small changes, but together they make your content significantly easier for AI systems to understand and surface.

As AI continues to change how people discover information, preparing your Optimizely website today will put you in a much stronger position tomorrow.

Happy Optimizing!

0

Optimizely: Our Journey in the Opal Innovation Challenge

When we signed up for Optimizely’s Opal Innovation Challenge, we knew it wasn’t just about building a solution—it was about solving a real marketing pain point. As marketing professionals and technologists, we’ve seen countless campaigns fail despite being well-written. Why? Because they didn’t resonate with the audience. This challenge gave us the perfect opportunity to tackle that problem head-on.

https://www.optimizely.com/operation-opal/#challenge

The Spark Behind the Idea

Marketing teams often struggle with persona alignment. A campaign might sound great internally, but fall flat with the target audience. For example:

  • Campaign 1: “Save more with Smart meters” – casual tone, generic message.
  • Campaign 2: “Embracing Green Living” – empathetic, informative, and resonates better.

This gap between content creation and audience expectation inspired our solution.

Our Team & Approach

We formed Infosys Transformers, a two-member team:

Our goal was simple yet ambitious: Make every piece of content count.

Challenges We Faced

  • Time Constraints: Building a multi-agent workflow in a hackathon timeframe was intense.
  • Balancing Creativity & Tech: Ensuring the solution was both innovative and practical.
  • Persona Complexity: Defining measurable persona attributes for scoring was tricky.

Timelines

Key Lessons Learned

  1. Persona Alignment is Non-Negotiable: Content without persona context is just noise.
  2. AI Can Bridge the Gap: Multi-agent systems can automate what manual reviews can’t.
  3. Collaboration is Everything: Combining tech and marketing expertise was our biggest strength.
  4. Start Simple, Scale Later: A modular approach helped us deliver a working prototype fast.

The Hackathon Experience

The Opal Innovation Challenge wasn’t just a competition—it was a learning marathon. We explored Opal’s agent ecosystem, experimented with orchestration, and realized how AI can transform marketing workflows.

Check out our video presented at the Hackathon – 

 

Closing Thoughts

Innovation happens when you combine empathy with technology. For us, this challenge was proof that marketing and AI are not separate worlds—they’re partners in creating meaningful experiences.

Top teams can win prizes like $5,000 in MDF, gift cards, and thought leadership opportunities.

Hope this helps.

Happy Optimizing!

0