An independent research tool tracking AI adoption across HR functions. Scores companies across 7 domains using public signals — earnings calls, job postings, vendor case studies, governance announcements, and more. This reflects only what’s publicly visible, not the full picture. Use it as a starting point for your own analysis and a reason to reach out to peers. No company sponsors or endorses this tool. Created & maintained by Andrew Helms.
Spot an error? Use the flag button on any signal.
Industry Median Maturity
Signal Type Breakdown
All Companies filter via column headers
Select two companies to compare side by side.
Companies with maturity > 3 shown. Color intensity represents domain score (0-100).
Highest Positive Momentum
Governance Leaders
Companies ranked by HR AI governance activity score.
HR AI Tech Landscape
All technologies tracked across HR functions — from core HRIS to specialized AI tools.
Signal Explorer
Browse and filter all signals across every tracked company.
Methodology
This tracker evaluates HR AI adoption across companies by collecting, classifying, and scoring public signals. Below is a complete breakdown of how scores are computed.
Every signal gets a point value calculated from three independent factors multiplied together:
points = Signal Type × Source Credibility × Time Decay
Signal Type captures what kind of evidence was found — a confirmed deployment carries far more weight than a press release mention. Source Credibility adjusts for how trustworthy the source is (an earnings call vs. a social media post). Time Decay ensures recent signals matter more than old ones.
Together, these three factors determine how much each signal contributes to a company’s maturity score.
Why this matters: A deployment (weight 10.0) from an earnings call (source 1.0) scores far higher than a PR mention (weight 1.0) from social media (source 0.3). The difference in evidence quality is reflected directly in the score.
Each signal is classified by type. The base weight reflects how strong this type of evidence is as proof of real adoption:
| Type | Weight | Example |
|---|---|---|
| Deployment | 10.0 | Enterprise-wide rollout of Eightfold AI for recruiting |
| Pilot / Limited Rollout | 6.0 | Testing AI recruiting tool in one division |
| Leadership Hire | 5.0 | Hiring VP of AI for HR |
| Upskilling | 4.0 | Company-wide AI training for HR team |
| AI Talent Hiring | 3.0 | Job posts for ML engineers, data scientists |
| Announcement | 1.0 | Press release about AI strategy |
Governance is scored as an HR domain, not a signal type. Signals in the Governance domain receive a 6x domain multiplier to reflect the strategic importance of AI oversight activities.
Different source types carry different credibility multipliers:
Signals lose relevance over time using exponential decay. Most signal types use a 180-day half-life — a signal from 6 months ago contributes half the weight of an identical signal today. Deployments and Governance signals use a 730-day (2-year) half-life because they represent durable institutional facts: a confirmed tool deployment or an established AI governance framework remains relevant long after the initial announcement.
Classified based on where signal weight concentrates:
Measures how deeply and broadly a company has adopted HR AI:
points = base_weight × source_credibility × time_decay
7 domains: Recruiting, Talent Development, People Analytics, Comp/Benefits, HR Service Delivery, Governance, AI Hiring. Each scored 0–100 independently with diminishing returns within a domain (1/√rank). Domains with at least one signal score a minimum of 1.
base = (avg_all_7 × 0.6) + (avg_top_3 × 0.4)60% rewards breadth across domains, 40% rewards depth in strongest areas.
final = base × √(size_norm) × √(regulatory_adj)Smaller firms get a boost per signal (capped at 2×). Highly regulated industries get credit for operating in constrained environments.
Measures whether adoption is accelerating or decelerating:
raw = (change × 0.6) + (trend × 0.4)momentum = tanh(raw / 4) × 10tanh smoothly maps raw values to −10 … +10, compressing extreme values. In practice, scores rarely exceed ±8 because tanh flattens as it approaches its limits.
Companies are ranked relative to the current filter selection. If you filter by industry or sub-industry, ranks recalculate within that subset. Use the column header filters on any table to focus on specific segments.
Company attributes (revenue, employee count, market cap) come from different sources depending on whether the company is public or private:
Public companies: Revenue and employee count are sourced from the most recent earnings report or SEC filing (10-K/10-Q). Market cap is sourced from financial data providers. All public company attributes are refreshed weekly. These values link directly to the source.
Private companies: Revenue, employee count, and market cap (where applicable) are estimated using publicly available data — press reports, industry analyses, LinkedIn headcount, and comparable public peer multiples. These values are marked with an est. tag.
Update frequency: Revenue, employee counts, and market cap for public companies are refreshed weekly from the latest available filings, reports, and financial data. Private company estimates are reviewed periodically as new data becomes available.
All estimated figures should be treated as approximations. If you spot an inaccuracy, use the flag button on any signal or reach out directly.