ACTION_ID: company_news_events_from_predictleads
NAME: PredictLeads: Find News Events
CATEGORY: Company Enrichment
CREDITS: 1

Find news events published about a company domain, optionally bounded by
date, returning each article's title, body text, author, link and
publish date.

1. INPUTS

company_domain (type: url, required)
  Company Domain. Domain of the Company for which you want to find
  news events

date_from (type: string, optional)
  Date From. Returns NewsEvents found after given date

date_until (type: string, optional)
  Date Until. Return NewsEvents found before given date.

news_events_needed (type: number, optional)
  News Events Needed. The nuumber of news events needed(default 100)

2. OUTPUTS

news_events_count (number) — Number of total news events

news_events (structured_array) — News Events found

  Columns — reference as {{<action_instance_id>.news_events.<column>}}:
    id (string) — ID field
    author (string) — Author of the News
    title (string) — News Title
    news_text (string) — News Text
    image_url (string) — Image URL of the news event
    news_link (string) — News event link
    published_at (date) — News event published at timestamps

3. HOW TO CONFIGURE

Configure Action body:

{
  "inputs": {
    "company_domain": "{{input.company_domain}}",
    "date_from": "2026-05-01",
    "date_until": "2026-08-01",
    "news_events_needed": "25"
  }
}

Field-by-field:
  - company_domain      Required. Bare domain — no scheme, no `www.`.
  - date_from /         `YYYY-MM-DD` bounds on when the event was found.
    date_until
  - news_events_needed  How many events to return (default 100). Pass as
                        a quoted string — Configure Action ignores raw
                        JSON numbers.

4. KEY NOTES

- `news_events_needed` is a number field, so pass it quoted (`"25"`). A
  raw JSON number is dropped with an `unknown_field` warning.
- The credit cost is per call, not per event — pulling 100 events costs
  the same as pulling 5.
- `news_events` is a `structured_array`. To summarise the articles with
  an LLM, fan out with `push_data_to_sheet` or reference the list in a
  prompt; column refs do not resolve inside a `formula`.
- Coverage is PredictLeads' own crawl. Small and non-English companies
  frequently return zero events — check `news_events_count` before
  branching on content.

5. WHERE IT FITS IN A WORKFLOW

Pattern: company_news_events_from_predictleads -> filter
(news_events_count greater than 0) -> llm_models (classify the trigger)
-> push_data_to_campaign.

6. WHEN TO USE

Use company_news_events_from_predictleads to detect account-level
trigger events — launches, partnerships, expansions — as an outbound
timing signal.

7. WHEN NOT TO USE

The signal you want is hiring activity
  -> action-detail/company_open_jobs_from_predictleads.txt

The signal you want is a funding round
  -> action-detail/company_funding_and_acquisition.txt

Last updated: 2026-08-03.