Join the Community

Clay Announcements

Clay Announcements

:jammies: New feature drop: Find Company Lookalikes! :jammies:

·

New feature drop: Find Company Lookalikes! When you're building lists on Clay, you can now start with any company's LinkedIn URL and automatically find more companies like that one using an LLM search! Check out this video to see it in action 👇 How it works:

If your table has a LinkedIn company URL, you'll see a button on the bottom that says "Find Company Lookalikes" 👀
Tick the box for one of the rows in your table & click the button to launch a lookalike search for that company
Review the results to add to your table ✅

What this means: As long as you know a few companies that fit your ICP, you can find others from our LinkedIn dataset—in seconds. How it works: We take a variety of factors from the company you provide (description, industry, size, etc) and feed them to a custom large language model (similar to OpenAI) to compare to all the other companies on LinkedIn. We then rank the results reported by the LLM and return what we deem to be the top 30 in terms of quality. Shoutout to Mark H. and Brian L. for bringing this feature to life ❤️

20 comments

Gleb K.
·
Awesome feature! Will be looking to check it out 👍
Luke B.
·
I'm still not seeing the box at the bottom on all of my tables. Is there anything special I need to do to set it up?
Mishti S.
·
Luke B. are you looking at tables with LinkedIn company URLs in them? If so, and you're still not seeing the lookalikes box, can you share a link to your table and we'll check it out? cc Arturo O.
Arturo O.
·
Hey Luke B., let me know if that helped or I can check if it's enabled for you! Feel free to DM me your email or a table URL
Daniel K.
·
Can you explain how this works in the background?
Jamie B.
·
Marcin
Mark H.
·
@Daniel We take a variety of factors from the company you provide (description, industry, size, etc) and feed them to a custom large language model (similar to OpenAI) to compare to all the other companies on LinkedIn. We then rank the results reported by the LLM and return what we deem to be the top 30 in terms of quality.
Ajay G.
·
Mishti S. is it possible to get this working with sales nav links as well? will make workflows quicker :)
Mishti S.
·
Ajay G. DM'd you about this but we'll note this and potentially add in the future, cc Mark H.! Thanks for the feedback 🙂
Niklas N.
·
Has anyone tested this in a campaign already? At times the LinkedIn companny suggestions are quite off and not accurate.
Mark H.
·
Niklas N. These are AI-generated lookalike suggestions (vs something more hand-curated) based on each company's LinkedIn presence, so it's quite possible that accuracy for some companies would be lower than others. We'd be interested in feedback if you notice consistent poor performance for certain types of companies or industries, which we can use to continue to improve our models. Do you have some examples at hand that you wouldn't mind sharing?
Neil S.
·
Very cool! Would love a similar lookalikes feature for People. And it'd be great to be able to use more than one company as input, to avoid over-generalizing
Ajay G.
·
Agree with being able to use more than one company as input ^
Mark H.
·
Multi-company lookalikes is something we're working on!
Mike F.
·
Mishti S. - sorry, where can I find the ‘Find Look-Alike’ button?
Arturo O.
·
Hey Mike F., once you have a company LI URL in a column in any sheet, the lookalike button should pop up on the lower left side of your table. To use it, you have to select a single row with a URL and it will become clickable as shown in the video. Let me know if you run into any issues!
Armand P.
·
Flagging this. Getting it when I select one or multiple company
Brian L.
·
You need a column LI URL in a column like Arturo O. mentioned - will work on improving this experience + error messaging 🙏
Arturo O.
·
Yup! For reference, this is what that means by having a column with the value, Armand P.. This feature can't currently read what's stored inside the integration/enrichment.
Armand P.
·
in business, thanks Arturo O.