ADP
Sorbet currently classifies ADP as the dark_horse in its HRIS & payroll bucket. The detailed selection note is not yet available.
No recent material read
Based on 0 material signals in the last 30 days: 0 positive, 0 negative.
Evidence summary, not a buy / sell rating
The role and pillars explain what job this holding has inside the System-of-Record basket—not whether it suits an individual investor.
ADP sits on one of the most structurally durable record sets in enterprise software: payroll runs, tax filings, benefits elections, time-and-attendance logs, and worker classification histories that span decades for some customers and carry legal retention obligations that make migration existentially costly. Those records are unambiguously P1-positive — every employee is a typed, owned, versioned object with a hire date, a compensation history, and a chain of approval events — and P3-positive, because "who is responsible for this person's pay and compliance" is a hard field with jurisdictional enforcement behind it, not an inference from a Slack thread. ADP's dark-horse designation reflects that it is not the thesis's primary HRIS vehicle — that role belongs to more-modern-stack challengers and leaders — but it represents a floor of agent-infrastructure value that is extremely difficult to dislodge: any agent that needs to act on workforce records (trigger an onboarding sequence, verify employment status, compute a severance liability) must eventually read or write through the system that legally owns those records, and for a large fraction of mid-market and enterprise employers, that system is ADP. The compounding moat here is compliance depth: payroll tax rules across fifty states and dozens of countries are encoded in ADP's calculation engine and updated continuously, creating a queryable, auditable substrate that no greenfield agent platform can replicate without years of regulatory exposure.
A smaller, higher-uncertainty paper position testing an emerging or less-consensus expression of the bucket thesis.
The thesis properties this holding is selected to test. These labels are portfolio classifications, not standalone proof.
The thesis for ADP breaks structurally — not merely underperforms — if one or more of the following conditions materialise. First, if a hyperscaler (most plausibly Microsoft, given its deep enterprise HR integrations via Viva and Teams) bundles a payroll-and-people-record layer into its M365 or Azure commercial suite at a price point that causes ADP's mid-market retention rates to visibly erode in consecutive quarterly cohort disclosures, that is a direct strike on P1 and P3: the durable employee record and the ownership/approval chain would migrate to the bundler's substrate. Second, if ADP fails to ship a credible agent SKU — meaning no announced consumption-priced product that exposes its people-record API or MCP-compatible surface to autonomous workflows within the portfolio's three-year horizon — and peer HRIS platforms (Workday, Rippling) demonstrably capture agent-driven HR automation wins that ADP cannot reference, ARR mix evidence would confirm the monetisation-failure invalidator. Third, if open-source payroll and HRIS tooling (e.g., self-hosted alternatives gaining traction among mid-market buyers) drives a sustained compression in ADP's Employer Services gross margin below its long-run band, pricing power on the compliance moat should be treated as structurally impaired rather than cyclically soft. The concrete signal to watch is ADP's Employer Services revenue retention rate and any disclosed API or agent-platform attach metrics in quarterly filings — a multi-quarter retention decline coinciding with margin compression, not a single-quarter miss, is the falsifying pattern.
Daily closes are shown for context. Reference levels are deterministic outputs of Sorbet's published ATR method, not analyst targets or advice.
2025-07-28 → 2026-08-31
$286.16
Revenue growth + FCF yield, not the conventional FCF-margin formulation. Visible for judgment, but not wired into selection.
Price is above both the 50d and 200d averages, with the faster average leading.
ADP's trajectory in the agent economy runs through its unmatched depth of payroll and compliance records — if it can expose that substrate via stable, machine-readable APIs rather than keeping it locked inside its own UI surfaces, it becomes genuinely difficult to displace as the people-record layer an agent queries before taking any action involving a human. The critical execution requirement is completing the transition from a service-bureau model, where ADP historically owned the workflow on the customer's behalf, to a platform model where third-party agents and orchestration layers can read and write structured people-state through documented, versioned interfaces. What remains unproven is whether ADP will open P2 and P4 — making its approval chains and HR verbs (onboard, terminate, promote, change-role) legible to an agent as named state-machine transitions rather than routing requests through human-mediated service channels. Until that surface is credibly agent-addressable, ADP's P1 and P3 strengths are necessary but not sufficient to confirm it as a full-pillar SoR substrate in the agent economy.
No material signals for ADP in this period.
Durable records
Typed ownership
Short-term levels use 14d ATR and the 200d average. The 3y projection compounds the current revenue-growth rate with no multiple expansion. Missing inputs stay blank.
| Quarter | Revenue YoY | Gross margin | FCF margin | Operating margin |
|---|---|---|---|---|
| 2026-03-31 | 8.55% | 48.40% | 34.42% | 30.06% |
| 2025-12-31 | 7.29% | 46.26% | 18.50% | 26.15% |
| 2025-09-30 | 7.88% | 45.37% | 9.69% | 25.75% |
| 2025-06-30 | 8.33% | 45.22% | 25.43% | 23.54% |
| 2025-03-31 | 6.63% | 47.85% | 25.00% | 29.41% |
| 2024-12-31 | 9.07% | 45.88% | 20.13% | 25.75% |
Missing data. No owner-reviewed NRR disclosure is available. Unreviewed AI extraction is never shown publicly.
No approved, rejected, or executed decisions reference this holding yet.