What is social intent search? How to search 71 million LinkedIn engagements in plain English
Social intent search finds prospects by what they say and engage with in public. In traxy, one sentence searches 71M+ LinkedIn likes, posts and comments.
Social intent search is a way to find prospects by what they have said and engaged with in public, instead of by job title alone. In traxy, you type a sentence such as "find me founders looking for AEO help" and it searches a database of more than 71 million LinkedIn likes, posts and comments. Each match comes back with the person, their company, their ICP match and the post or comment that put them on the list. You can run the same search inside traxy, or from Claude or ChatGPT through the traxy MCP.
traxy launched social intent search on October 29, 2026 (launch post, product page). This guide explains what social intent is, what the search does, where each step happens in traxy, and where its limits are.
What is social intent?
Social intent is what people show through the things they like, post and comment on. traxy's own definition puts it plainly: someone who comments "we need help with AEO" is telling you what they need (product page FAQ).
On LinkedIn this shows up in three forms:
Posts. A founder writes "Who is good at AEO? We're invisible in ChatGPT answers."
Comments. Someone replies under a creator's post with "anyone know a good agency for this?"
Likes. A marketing lead likes a post comparing two podcast agencies.
These are different strengths of evidence. A post or comment asking for help says more than a like. A like says the topic caught their attention. traxy's docs make the same distinction and warn against reading budget or purchase intent into a generic interaction (outbound from qualified signals).
What is social intent data, and how is it different from other intent data?
Most B2B intent data works at the account level. It tells you a company's employees are reading about a topic somewhere on the web, without telling you who. Social intent data works at the person level. You see the named person, the public post or comment, and the topic.
That changes what you can do with it:
Account-level intent data | Social intent data | |
|---|---|---|
Unit | A company | A person |
Evidence | Aggregated research activity | A specific like, post or comment you can open |
What you can say in outreach | Little you can reference | The topic they raised, at the strength they raised it |
Main risk | Guessing who at the company cares | Treating a weak engagement as a buying decision |
Neither replaces the other. If you sell to named buyers who talk about their problems on LinkedIn, social intent data is the more direct evidence.
What does traxy's social intent search do?
Over the past couple of months, traxy scanned millions of LinkedIn likes, posts and comments. Social intent search puts those engagements into one searchable database. At launch it held more than 71 million engagements, and the product page describes it as growing (product page; launch post).
The search works in three steps, as described on the product page:
You ask in plain English. Describe the buyer and the need the way you would say it to a colleague. There are no Boolean strings or filter trees to build.
traxy searches the database. Your question runs against the LinkedIn likes, posts and comments in it.
You see who, and why. Each match shows the person, their company, their ICP match, and the post they commented on, liked or reposted.
The "why" is the useful part. Each name arrives with its evidence, so you can decide whether it is strong enough to act on.
Where does each step happen inside traxy?
Here is the path from a sentence to a first conversation, using the areas named in traxy's docs.
1. Set who fits
Before you search, check your Lead criteria. This is your ICP: roles, departments, seniority, location, company size and industry (lead criteria). Social intent search shows an ICP match for each person, and that score comes from these criteria. If your criteria are vague, your matches will be too.
Then open Signals and review your Signal Brief: the problems, topics and buying situations connected to your offer (signals).
2. Ask trax
Open trax, the agent chat, and describe who you want. Example prompts, adapted from the product page and launch post:
"find me founders looking for AEO help"
"heads of RevOps unhappy with their CRM"
"ecommerce founders asking about retention emails"
"marketing leaders comparing podcast agencies"
"CFOs talking about month-end close"
Be specific about the need. "Founders interested in marketing" will match a large, loose crowd. "Founders asking who can help them show up in AI answers" matches people who said something close to that.
3. Read the results and open the evidence
Results open beside the chat as a table with each person, their company and the signal, such as "Commented on a post about AEO" or "Liked a post about AEO agencies". Open a lead to see its details, and use the lead's Signals tab to read the recorded observations and source links (review qualified leads).
Read the source and its date before you use it. A comment asking for a recommendation is stronger than a like on a post about the same topic.
4. Keep the good ones and teach traxy what bad looks like
Mark leads Good fit or Bad fit with a reason (how qualification works). Save the people you want to contact as a view in Leads, so you can reuse the audience (filter and save audiences).
5. Reach out from the evidence
Ask trax to prepare a campaign for that saved view. You test the plan, inspect the facts and messages, and review the launch before anything sends (create a campaign; review and launch). Sending uses your own connected LinkedIn account and mailbox. Replies land in Inbox.
Example prompt: "Create a campaign for my AEO founders view. Connect on LinkedIn first, then follow up by email if they don't reply."
Can you run social intent search from Claude or ChatGPT?
Yes, according to the launch post and the product page. Connect the traxy MCP to Claude or ChatGPT, then ask the same question in that chat. You search the same database you would search in the traxy app (product page).
To connect, add traxy's remote MCP endpoint in your client and authorize it with OAuth (connect an AI client). A traxy REST API key is not accepted as an MCP credential. For a fuller walkthrough of what Claude can do with your traxy workspace, see how to find and review LinkedIn leads from Claude with traxy.
Campaigns, senders and sequences stay in the traxy app. They are not part of the public MCP tool catalog (MCP tools).
What are the limits?
It is LinkedIn only for now. Instagram and Facebook groups are listed as coming soon on the product page. Plan around LinkedIn.
A match is evidence, and it can be weak. A like on a popular post can come from someone with no interest in buying. Read the signal before you contact anyone.
Fit still matters. A person asking the perfect question at a company you can't serve is not a lead for you. traxy's docs treat fit and signals as separate checks for this reason (lead criteria).
Some people fish for pitches. One reply to the launch asked how to filter out people who post "anyone know a good agency?" only to collect pitches (reply). traxy doesn't claim to detect this. Check the person's role and company, look at what else they have posted, and keep your first message useful rather than salesy.
How does social intent search fit with the rest of traxy?
Search is one way into traxy. The others keep running after you set them up:
Discovery runs find new people from your lead criteria and signals. You can review each run in trax > Discovery runs (discovery runs).
Followed profiles surface people who engage with the posts of competitors, creators or teammates you choose (track people, posts, and companies).
A single post can be checked for the people in its audience who fit your ICP. See find leads from any LinkedIn post with the traxy agent.
Search is for the question you have today: "who is asking about this right now?" The other sources keep adding people over time.
FAQ
What is social intent?
It is what a person shows through what they like, post and comment on. A comment like "we need help with AEO" is social intent. traxy makes those LinkedIn engagements searchable.
What is in traxy's social intent database?
More than 71 million LinkedIn engagements at launch: likes, posts and comments that traxy has scanned (product page).
Do I need to write Boolean searches?
No. You describe the buyer and the need in a sentence, such as "heads of RevOps unhappy with their CRM".
What comes back with each result?
The person, their company, their ICP match, and the signal behind the match: the post they wrote, commented on, liked or reposted.
Does it cover Instagram, Facebook groups or X?
The current database is LinkedIn only. The product page lists Instagram and Facebook groups as coming soon, and X is not mentioned.
Does traxy write the outreach for me?
You build a campaign with trax from a saved audience, then test it and review the messages and the launch before anything sends. You approve what goes out.
Do I need to connect my LinkedIn account to search?
traxy's lead sourcing does not need a LinkedIn connection. Sending outreach does, because it goes out through your own LinkedIn account and mailbox (sender accounts).
Want to try it with your own buyer? Start at traxy.ai/social-intent-search and run one search for the problem your buyers talk about most.
Keep reading
All Buyer Intent & Signals guides
Buyer Intent Signals (2026): What Buying Signals Are, Examples by Type and How Sales Teams Act on Them
Buyer intent signals explained: what buying signals are, examples by type, how reps act on each one and a simple scoring framework to prioritize accounts.
Social listening vs social intent search: which one finds B2B leads on LinkedIn?
Social listening tracks mentions and sentiment for marketing. Social intent search finds named people who fit your ICP and show a need. Which to use.
How to find people on LinkedIn asking for a recommendation for what you sell
Search for the request, not the job title. How to find LinkedIn posts and comments asking for a recommendation for what you sell, by hand or with traxy.