Job Change Signals for AI Agents: Catching Buying Windows (2026)

Job change signals for AI agents catch a moment most CRMs miss. A VP of Engineering you sold to two years ago just switched companies. She landed at a 400-person org that never heard of your product, and she posted about it the day she started.

Your records still show her old title, old email, and old account. Nobody notices for six weeks, and by then a rival with sharper timing has booked the intro call. Her move was a buying window swinging open, and the agent that reads her post first wins the account. The job is to catch the move while it is still fresh, enrich the person, and act while the seat is warm.

Key Takeaways#

• A job change resets budget, mandate, and tooling at once, which makes it the sharpest B2B buying trigger.
• People announce moves in their posts, so a daily social-search routine surfaces job changes as live data with no server to host.
• A scheduled Claude agent can run the search, spot the move, and enrich the person for a fresh work email every day.
• Webhooks are the complement: use a monitor when you want push on a fixed watch list instead of a daily poll.
• The right reaction re-qualifies the person in their new seat instead of replaying the old deal.

Why is a job change the sharpest B2B buying signal?#

A job change resets a buying decision all at once. New person, new budget line, new mandate, and a stack of tools waiting to be filled. Someone who just started wants early wins, and they carry spending permission that a tenured employee rarely has. That mix makes the move sharper than a whitepaper download.

Timing carries the rest. A champion who moves takes their preference for your product into an account where you hold zero presence. When a buyer you were courting leaves instead, the deal you thought was live has gone stale, and a stranger now sits on the budget. One event, two very different plays.

The reverse case is just as useful. When a contact leaves an account you own, the seat they vacated still holds the spend you were working. A good agent treats one departure as two openings, the person and the seat, and opens a fresh thread on each.

How does an agent catch the move first?#

Run a daily social-search routine, because people announce their own moves. "Excited to join," "starting a new role at," and "thrilled to announce" show up in posts within hours of a start date, well before any database refreshes a title. Searching those patterns every day turns the public feed into a live stream of job changes you can act on the same afternoon.

This is the most accessible way to catch the signal because it needs no infrastructure. There is no webhook endpoint to host, no server to keep online, and no queue to manage. A single scheduled call to Search Social Posts (/search/posts) returns fresh matches on each run, and the agent enriches each person to confirm the move and pull the new work email. For the architecture behind reading a feed as fresh data, see our guide to real-time data for AI agents.

You get two search modes from one endpoint. Search the announcement phrases across the open feed to catch moves anywhere in your market, or search the recent activity of your own watch list to catch the people you already track. The first fills the funnel, the second guards the accounts you own, and a daily run covers both without a growing bill for checks that return nothing.

The reason to lead with search rather than a stored dataset is freshness. A title that a provider refreshes on a quarterly cycle can already be months stale by the time your agent reads it, while a post published this morning is the person telling you, in their own words, that they moved today. Live data closes the gap between the move and the outreach, and that gap is exactly where a sharper competitor slips in ahead of you.

Building the daily social-search routine as a Claude agent#

Wire the routine as a scheduled Claude agent that runs once a day. Each morning it calls Search Social Posts with your announcement phrases, filtered to the last day so it only sees fresh posts. The request is a short body: the keywords to match, a time window, and the source you want to read.

{
  "keywords": ["excited to join", "starting a new role", "thrilled to announce"],
  "since": "24h",
  "source": "social"
}

Each match names the author and the post, for example a profile at social.com/in/jane, so the agent knows exactly who to enrich next. Because the run is live, a post from this morning is in hand by lunch rather than a week later. Add your watch-list profiles as a second query on the same schedule, and one routine covers both open-market discovery and account defense. The DataForB2B buying signals API exposes job change alongside funding, founder moves, and post engagement, so the same routine can widen to other triggers later.

The agent's reaction workflow, step by step#

A detected move should fire work, not just update a field. Once the routine surfaces a post, the agent runs a short loop that turns a raw match into a qualified play. Four steps cover the path from the post landing to the first draft waiting for approval, and each one leans on a real endpoint rather than a manual review.

1. Read the match: the post text and author tell you who moved and where they landed.
2. Call Enrich Profile to fetch the new work email and confirm the current role.
3. Re-qualify the new account with Search Company, checking size, funding stage, and whether your pain still fits the seat.
4. Draft the play in your CRM for a human to approve: a warm reintroduction for a past customer, a fresh thread for the seat they left behind.

The worst version of this is a templated blast that congratulates the new title and pitches in the same breath. Enrich first, then write like you actually noticed. Most teams get this wrong by treating a job change as a contact-update chore: they sync the new title into the CRM, mark the record clean, and let the window close without a single touch going out. The first ninety days in a new role are when budgets get set and vendors get chosen, so a move you act on this week is worth far more than one you file. You can wire the whole loop against a live endpoint before spending a cent. Start on the free plan, schedule the routine, and watch a real post flow through enrichment to a drafted reintroduction.

When should you add a webhook monitor instead?#

Reach for a webhook when you want push on a fixed watch list rather than a daily poll. The daily social-search routine is the accessible default, but if you already run a server and want the change the instant it lands, a monitor delivers it. Send one POST to the monitors endpoint with three fields: the job_change signal, the watch list of profiles it applies to, and the url that receives events.

Every event lands in one envelope with the signal, the target who moved, and a data block holding the new company and title. Set a secret and each delivery carries an X-Webhook-Signature header signed with HMAC-SHA256, so the agent can verify a payload before acting. Failed deliveries retry with backoff, and the List Monitor Events call lets an agent replay anything it missed during an outage. The reaction loop is identical to the routine above: enrich, re-qualify, then draft for approval. Push simply removes the daily wait for teams that need the seconds.

The tradeoff is scope. A monitor watches the exact profiles you register, so it is precise but blind to a move outside that list, while the daily search sweeps the open feed and surfaces people you were not yet tracking. Most teams end up running both: the search routine to discover new names and the monitor to guard a short list of named champions and open opportunities. Start with the routine, since it costs nothing to stand up, and add a monitor only once a specific watch list earns the push.

How do job change alerts compare to intent signals?#

They answer different questions, so a strong agent runs both. A job change alert says a specific decision-maker moved, a person-level buying window. An intent signal, like someone engaging with a competitor's post, says a person is active on a topic right now. One tracks who holds the budget, the other tracks who is paying attention.

Neither wins alone. Intent without authority chases people who cannot buy, and a title change without topic interest can arrive before any need exists. Read together, a recent move plus fresh engagement points at a buyer with both the mandate and the motive. Scoring gets sharper when the agent stacks them: a job change alone might earn a moderate score, but the same person engaging with a rival's post two weeks later should push it higher. Decay the score when things go quiet, and raise it when signals cluster inside a short window, so the agent spends its touches where conviction is highest. DataForB2B carries both event types over one surface, so a single integration feeds the whole picture. See the plans and schedule your first routine this week.