This post is part of the AI design patterns series – I’m sharing new design patterns I’m seeing as AI enables solutions that weren’t possible before.
I’m testing and evolving the ICP definition for something we’re building. I wanted to reachout to creative agencies to test if they’ll be interested in sharing early feedback on our prototypes.
My first instinct was to see if there’s anyone in my first-degree LinkedIn network who runs a creative agency. I went to LinkedIn search, and tried looking up keywords like “agency”, “webflow”, “marketing”, “brand” etc. to find people.
There were people who were running Webflow development or WordPress agencies in my connections, but did not necessarily have these terms like “agency” mentioned in important fields within their profile. As a result, LinkedIn results were pretty bad.
Siddharth introduced me to Happenstance – a product that helps you run natural language searches on your connections on LinkedIn, Twitter, etc. You can see a sample search that I ran on Happenstance below – the product understood what I meant in a way that LinkedIn couldn’t, and got back to me with fairly good results.
Boardy was even better – I shared my requirements with Boardy in plain English through WhatsApp. Boardy asked me additional questions to refine the search I was attempting, and got back to me with fairly good results.
Both Happenstance and Boardy did a much better job at what LinkedIn search was supposed to be good at.
The lesson here is simple: consumers expect answers when they search – not simple results based on keyword or semantic relevance.
So, don’t build search like it was built until 2023. The search experience in your product needs to understand and answer actual questions from users about their data and provide true answers. If it doesn’t, your product risks irrelevance.
This post is part of the AI design patterns series – I’m sharing new design patterns I’m seeing as AI enables solutions that weren’t possible before.
Shankar is thinking 