Client Operations
Leads marked as sent
Today0
This week0
Last week0
This month0
Filter Leads
Visible Columns
Focused workflow
Outreach Queue
Only leads with a generated outreach message that has not been marked as sent.
0 ready
Leads marked as sent
Today0
This week0
Last week0
This month0
Choose a client to load the queue.
Keyboard shortcuts
J/K select
O open LinkedIn
G check Sent
M mark sent
B mark bad fit
R regenerate
Ops Now
Queued or running OutreachBot work: manual searches, scheduled searches, ingest runs, score recalculations, and delivery jobs.
| Tenant | Status | Stage | Progress | Updated | Run ID |
|---|
Provider-side Exa Webset searches created by company-pattern runs. A Webset search can be active even when no OutreachBot pipeline run is currently running; once complete, it queues an ingest run.
| Tenant | Status | Exa Webset ID | Queued Ingest Run | Poll Attempts | Next Poll | Updated | Actions |
|---|
Diagnose
Scoped to selected client.
Recent completed, failed, cancelled, and active runs for the selected client.
| Created | Status | Candidates | Leads | Tier A / B | Est. Cost | Run ID | Actions |
|---|
Audit trail for the scheduler deciding whether to queue a run. A “Run queued” row is historical; check Active Pipeline Runs or Client Run History for the actual run state. Auto-refreshes every 60s.
| Time | Tenant | Outcome | Mode | Detail | Run ID |
|---|
Selected run dump collapsed by default
No run selected.
Recover
Repair queue for copying OpportunityLeads workspace changes back into legacy v3. These are not lead-search runs; failed items usually mean the new workspace saved but the old v3 mirror write needs retry.
| Tenant | Status | Attempts | Updated | Last Error | Entry | Actions |
|---|
Platform Config
Platform-wide — all clients
Platform-wide RSS feeds monitored for RFP opportunities. Individual clients match against these using their own RFP Signals trigger settings.
Influencer comment eval prompt platform-wide · collapsed by default
LLM prompt used to evaluate comments on influencer posts for pain signals. Use
{pain_phrases} as a placeholder — it will be replaced at runtime with the tenant's configured distress phrases. Leave blank to use the built-in default.Power Tools
Client JSON (advanced) power tool
Active client is selected from the left sidebar.
Start Search (power tool)
Same as Client Settings → Run Test Scrape; scoped to selected client.
Client Maintenance
Admin only. Refreshes company matches, company size checks, and disqualifying company-role evidence, then recalculates scores for existing leads without running a new search.
LISN bulk search jobs. Auto-refreshes every 10s while visible.
Create LISN bulk search collapsed by default
| Tenant | Name | Status | Imported | Queued | Processed | Leads | Skipped | Failed | Actions |
|---|
This checklist tracks the operational setup needed before a client can run reliably in the managed-service workflow.
Queue one test lead-search run for the active client using the current settings on this page.
Choose a client to queue a test scrape.
Active client is selected from the left sidebar.
General Settings
Basic client configuration: name, search limits, and the minimum score threshold for AI-generated outreach drafts.
Active when automatic runs are enabled and at least one enabled delivery endpoint is ready.
When enabled, the pipeline generates a 3-email outreach sequence for leads above the minimum score. Disable to skip this step and reduce OpenRouter costs.
Ideal Customer Profile
These are the settings you will touch most often when shaping who the client should target.
Requires a verified current target role, nonprofit organization type, annual revenue, and a cited revenue source. Hospitals, hospital systems, colleges, and universities are rejected.
Trigger Methods
Configure the lead-generation methods this client should run. Influencer comments remain the current path, and the other tabs add new signal-based triggers.
Influencer Comments
Monitor influencer profiles and surface any commenter who matches the ICP — no comment-text filter. Casts a wide net across the full audience. For a filtered variant that only fires on comments containing specific keywords or distress phrases, see Influencer Comments + Keyword.
The core switch for this trigger. When on, people who comment on posts from monitored profiles are evaluated as potential leads. Must be enabled for the option below to work.
Add-on — requires the option above. For each commenter found, runs additional searches to surface more adjacent prospects. Increases candidate volume; also increases run time and API cost.
When off, influencer commenters are admitted exactly as today. When on, commenters must match this title/headline gate before enrichment/research/scoring.
Add the influencer or creator profiles whose posts should be scraped for commenters. The Active toggle on each row controls whether that profile is included when a listening run happens. A listening run only happens when at least one of the comment-based trigger switches above is on — if all three are off, no scraping occurs regardless of Active.
Use this for saved searches, feed URLs, or other LinkedIn sources that should surface relevant posts.
Advanced JSON
Optional fallback for uncommon fields. If you edit this directly, those values will be saved even if the visual rows do not update until after reload.
Detect newly relevant leaders showing up in repeated people searches and leadership monitoring.
Search Variants — add multiple saved searches to run together. When variants are configured they run instead of the single URL above.
Look for hiring bursts or role-specific job postings that suggest a new initiative is underway.
Track important page changes like pricing, integrations, product pages, or case studies.
Watch for public events that create urgency, like funding, acquisitions, launches, or leadership announcements.
Stub for an Exa-based press-release search that looks for precursor events and infers the downstream buying need.
Stub for monitoring RFP RSS feeds and matching opportunities against client criteria. Admin-level feed configuration is separate; this controls per-client matching rules.
Compound trigger: first scrape a LISN company search (via Vine) to find fast-growing companies, then identify ICP-matching people within those companies. The goal is net-new company discovery → people identification.
Async Exa Websets company discovery. Best for company-pattern hypotheses like “small independent RIAs with a founder plus a clear #2.” A run may create or poll a remote Webset job, then fan out to people once matching companies are ready.
0 preserves legacy behavior. Use 7+ for static market-discovery searches.
To revert immediately, choose Websets — current / rollback mode and save. The next pipeline run will use the old Websets path.
Stub for a direct LinkedIn Sales Navigator people search, scraped via Vayne and metered to only add a set number of new results per month.
Use the single URL below, or add named variants to run multiple saved searches. Variants override the single URL when present.
Search Variants — add multiple saved searches, each with its own caps.
Filter social listening comments by text substring match. Uses the same scraping infrastructure as v3-outsiders-edge social listening. Useful for identifying commenters whose words signal pain, intent, or ICP fit — e.g., people commenting on "false hope" influencer content who reveal they are actively searching for help.
Like Influencer Comments, but only surfaces people whose comment text contains one of the configured keywords or distress phrases. Use this when you want to narrow the audience to commenters who are actively signaling pain, intent, or a specific topic — rather than everyone who commented. The influencer URL list here is configured independently from the broad Influencer Comments sources, so you can target different profiles or the same ones.
Optional phrases to prioritize commenters whose words signal pain or active problem-awareness. Leave empty to include all commenters.
Scans recent Google Maps reviews for businesses whose customers mention pest, maintenance, or other configurable keywords. Phase 1 discovers places via keyword+location search; Phase 2 fetches recent reviews and filters client-side; matching businesses fan out to Exa people search to find a contact.
Prefix-aware: “cockroach” matches “cockroaches”, “rat” matches “rats”, etc.
Searches posts across a LinkedIn industry using Apify LinkedIn post search, then mines comments and matches comment text against configured substrings. Useful for finding ICP-fit commenters in a specific vertical without needing individual influencer profiles. Cost caps apply per Apify actor run.
Cost caps: each keyword triggers a separate Apify actor run. Max posts per keyword caps the number of posts fetched per search. Max comments per post caps comment mining depth. Max total posts is the overall cap across all keywords.
Delivery Workflow
Choose whether this client keeps the default leads-first delivery flow or uses internal opportunity review before handoff.
Lead Delivery
Once targeting is set, this is the next thing most clients need: where approved leads should land.
Enable any combination — leads will be delivered to every enabled destination.
Paste a Notion page URL to auto-create a database there, or an existing database URL to use it directly.
No per-destination delivery config is active yet.
Frozen Inventory Campaign
Snapshots the currently undelivered qualified lead backlog. New leads are excluded. Creating a campaign disables the tenant-wide drip so only this inventory can deliver.
No frozen inventory campaigns loaded.
Google Sheet and Email Setup
1. Share your Google Sheet with the service-account email below as an editor. 2. Save your settings. 3. Test sheet access and email delivery.
Automation Schedule
The worker can enqueue runs automatically for this client, so you do not need an external Render cron service.
Cost Estimate Overrides
These are usually lower-frequency admin settings. Use your actual contract rates here when you need better cost reporting.
Outreach Messaging
Configure per-lead AI-generated outreach messages. When enabled, the pipeline generates a personalized message for each qualified lead above the minimum score, then optionally creates drafts in connected mailboxes.
Templates
Boilerplate
LLM Settings
Destinations & Delivery
Generated outreach messages always appear in the Leads view when outreach message generation is enabled. Use these toggles only to control whether outreach fields are sent to delivery destinations.
Mailbox Connections
Connect a Google Workspace or Microsoft 365 mailbox to create drafts. Drafts are never sent as live email.
Mailto Link
Adds a clickable mailto: draft link to enabled Google Sheets or Notion outreach-message delivery fields.
Tenant Gmail Inbox
Associates this client with a Gmail account index for compose links and Check Sent searches in both the Outreach Queue and Chrome extension.
Added to the BCC field when opening a mailto or Gmail compose draft.
Not connected
Not connected
Backfill Existing Leads
Generate outreach messages for existing leads that do not yet have one. Uses the Exa Agent for personalization (~$0.025/lead at low effort). No messages are sent — this only populates message fields.
Generating a full batch of 10 leads may take several minutes.
Informational only: the live pipeline still uses the server-side `.env` keys. This screen stores only non-sensitive status details, such as whether a key is configured and its last 4 characters.
Vayne LinkedIn Auth
Live admin control. Paste the raw `li_at` value or a full `li_at=...` cookie string to refresh Vayne's LinkedIn session. OutreachBot does not store this cookie.
Connection Status
Choose a client to load API key details.
Current Settings Snapshot
Choose a client to load current settings.