The goal was to make it easier for visitors to find relevant blog articles without a traditional search bar. We built and embedded an AI chatbot, connected it to the blog’s live content and fine-tuned its responses, aiming for the assistant to answer open-ended questions and suggests articles with short summaries and direct links.
Microsoft Copilot Studio
JSON metadata integration
Integration
Automation
Artificial Intelligence Solution
User ExperienceEnhancement
Our blog didn’t have a search bar. And while that’s not unusual, it can make things harder for new visitors who don’t know exactly what they’re looking for.
We started thinking: what if a user could just describe what they’re interested in, and the site would guide them to a few relevant posts?
That’s how the idea for an AI chat-based search assistant came about.
Make it easier for visitors to find relevant blog posts without a traditional search bar.
Use AI to handle open-ended questions and guide users to the right content.
Make it easier for visitors to find relevant blog posts without a traditional search bar.
We chose Microsoft Copilot Studio for its no-code setup, built-in GPT-based AI, and ability to connect to external data sources without extra plugins.
We chose Microsoft Copilot Studio for its no-code setup, built-in GPT-based AI, and ability to connect to external data sources without extra plugins.
We prepared and uploaded a JSON file with all article titles, links, and short descriptions. This gave the bot a reliable reference when generating answers.
In Copilot Studio we fine-tuned the bot’s system prompts so replies stay short, clear, and always include direct article links.
The Copilot embed script was added to the blog page, and the chatbot window styling was adjusted with CSS so it matched Rocksoft’s site design.
We chose Microsoft Copilot Studio for its no-code setup, built-in GPT-based AI, and ability to connect to external data sources without extra plugins.
We chose Microsoft Copilot Studio for its no-code setup, built-in GPT-based AI, and ability to connect to external data sources without extra plugins.
We prepared and uploaded a JSON file with all article titles, links, and short descriptions. This gave the bot a reliable reference when generating answers.
In Copilot Studio we fine-tuned the bot’s system prompts so replies stay short, clear, and always include direct article links.
The Copilot embed script was added to the blog page, and the chatbot window styling was adjusted with CSS so it matched Rocksoft’s site design.
Step 1
Step 2
Step 3
Step 4
The bot works!
Visitors can now search the blog just by asking a question — no need for a search bar.
The bot gives 2–4 relevant articles in seconds, each with a short description and a direct link.
It understands open-ended queries like “articles about cybersecurity best practices” or “how Rocksoft uses DesignOps”.
The feature is fully live on the Rocksoft blog and works on both desktop and mobile.
Setup required no custom backend — everything runs through Microsoft Copilot Studio.
Try it yourself on our blog page! If you're not sure where to start, just ask a question, it might save you a few clicks.
80
Articles in database (and still counting...)
4
+
People involved
6
tools and integrations used
Let's chat about your ideas and see how we can bring them to life. Drop us a message, and let's start this awesome journey together!
We’ll show you real examples of how the right technology choices can help your business grow faster and work better.
.svg){kind=icon}
.svg){kind=icon}
