Signal
Enterprise AI adoption
- Coverage
- US
- Billing
- 1 cr / call
This dataset tracks concrete enterprise AI applications disclosed by US-listed companies in earnings calls, including the use case, deployment stage and scope, disclosed value, metrics and supporting evidence. General statements about AI strategy are excluded.
Each row captures one application at one disclosure point. deployment_id identifies the row, application_id links the same application across calls, and evolution_summary records what changed. To follow an application over time, fetch by ticker and group by application_id.
Only metrics attributable to the application are included. evidence contains the shortest supporting excerpt. Uncertain classifications return unclear, and management targets are not treated as realized results.
Questions this endpoint answers
- Which concrete workflows does a company apply AI to?
- Which applications are still at the concept or pilot stage, and which have reached production?
- Does the value a company discloses come mainly from revenue growth, cost reduction or risk control?
- How do the deployment scope and business results of the same application change from quarter to quarter?
Query parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
| ticker | string | optional | At most 10 tickers, comma-separated; this endpoint covers US only, and a non-US ticker (including CN / JP) returns 404. |
| since | date | optional | Return only observations whose update_date is on or after this date; must use YYYY-MM-DD format. |
| limit | int | optional | Maximum 10. Default 10. |
| page | int | optional | Starts at 1. Default 1. |
Sorted update_date desc, deployment_id asc. Paged with limit and page (pages start at 1); fewer rows than limit means the last page.
Response fields
| Field | Type | Description |
|---|---|---|
| deployment_id | string | Stable public ID of this application disclosure observation, one per row |
| application_id | string | Stable ID of the same application, and the key that links its multiple disclosures across periods |
| ticker | string | The company's canonical ticker |
| company_name | string | Canonical company name |
| application_name | string | Short description of the application's identity and the workflow it sits in |
| first_report_date | date | The date the application was first recorded within the research window; not the date the company first adopted AI |
| update_date | date | The most recent disclosure date for this observation |
| summary | string | The application, results, financial linkage and limits of the judgment in the current disclosure. The text is mixed-language and the large majority is English ([Application] / [Current Result] / [Financial Link] / [Boundary] sections), with a small share in Chinese (【应用】…); the public-facing language is not yet unified |
| evolution_summary | string | What changed relative to the earlier disclosures of the same application; on a first disclosure, the field states explicitly that prior history is insufficient, without guessing |
| business_position | string | Where the application sits in the company's business; see OpenAPI for the full enum (for example support_function) |
| deployment_stage | string | Actual deployment stage supported by the disclosure: production / pilot / concept / unclear; unclear when the disclosure is not specific enough |
| deployment_scope | string | Organizational scope: enterprise_wide | multi_business | business_unit | limited | unclear; unclear when the disclosure is insufficient. New enum values are added, never changed, |
| value_type | string | The main type of value the application creates: cost_reduction | revenue_growth | risk_reduction | capital_efficiency | unclear; new enum values are added, never changed, ; when it comes from a management framework rather than a realized result, the summary text says so explicitly |
| metrics | object[] | Quantitative metrics that have passed the attribution check; may be an empty array. Each item contains metric_name, value, scope, metric_role, financial_metric, financial_attribution and evidence_id (matching evidence[].evidence_id) |
| evidence | object[] | The minimal set of source quotes supporting the current conclusion; each item contains evidence_id, quote (the original text), speaker and section; the latter two are null when unavailable |
No key yet? Create one under API Keys — it is shown once, so store it where your shell can read it. Connecting an agent over MCP instead? That signs in through the browser and needs no key at all.
Authentication, conventions, rate limits and error shapes are the same for every endpoint — they live on REST API.