Casework produces conformity files, Fundamental Rights Impact Assessments, and Article 9 risk documentation for AI hiring tools — aligned to the EU AI Act, Colorado AI Act, Illinois HB 3773, NYC LL 144, and the California rules taking shape behind them. One AI tool, one workflow, one production cycle.
The EU AI Act classifies hiring tools as high-risk under Annex III and binds deployers from August 2026. Colorado, Illinois, and California impose impact-assessment and disclosure duties. NYC LL 144 is already in force. Each regime defines its own documentation primitives, and a single hiring stack typically falls under three or more at once.
Design decisions live in Slack threads. Model cards exist if the vendor felt like writing one. There is no FRIA, no Article 9 risk file, no Annex IV-aligned technical documentation, no post-market monitoring plan. Agile delivery practices and regulator-acceptable records have never been the same artifact.
Their methodology is general-purpose AI governance, billed by a pyramid of staff who will learn hiring AI on your time. By the time the conformity file is delivered, the regulatory clock has run, the vendor has shipped two new model versions, and the partner has rotated off.
Each Pack is one production cycle: a structured intake, two weeks of discovery and drafting, a week of internal review, a week of revision and handoff. You receive the file, a walkthrough, and 30 days of revision support. Then we are done.
Each engagement is a structured production: intake, discovery, drafting, review, handoff. The methodology is published in full — what we ask, what artifacts we generate, what your team is on the hook for, and what the file contains when we close out. Read it before you commit.
Hiring AI only at launch. We do not file conformity packs for credit, healthcare, education, or generative tooling. Every engagement deepens a single methodology rather than spreading thin across verticals.
Fixed scope, fixed price, fixed timeline, fixed deliverables list. No procurement theater, no change-order economy, no junior staff billing into the weekend. You know exactly what you are buying before you sign.
Production is augmented end-to-end — discovery synthesis, regulation mapping, draft assembly, citation verification. Humans own judgment; the machinery handles the rest. That is how four weeks is possible at this price.
The regulatory window for hiring AI is narrow and well-defined. Between NYC LL 144 in 2023, Illinois HB 3773 in January, Colorado in June, and the EU AI Act's high-risk obligations in August, the documentation requirements for any HR team running an AI assessment have gone from optional to mandatory in roughly thirty months.
The firms positioned to help — Big 4 advisors, full-service law firms, AI-governance consultancies — are priced for the Fortune 500 and staffed for general-purpose work. For the mid-market deployer with one or two hiring tools in scope, that is the wrong instrument. Casework is the right instrument: one productized pack, one specialty, fixed scope, fixed price.
The firm is editorially anchored. We operate HireAIScore.com, an independent directory of AI hiring vendors. That work pays for our knowledge of the vendor landscape and keeps us honest about who is shipping what. The disclosure runs on every page of this site for a reason.
If your question isn't here, the intake form has a free-text field. We read all of them.
Law firms write memos. We produce documentation. The conformity file, the FRIA, the risk register, the monitoring plan — these are records a regulator opens, not arguments a partner makes on your behalf. Casework does not provide legal advice; we produce the artifacts your counsel, your auditor, and your regulator need to see.
If you need a legal opinion on whether your tooling is a high-risk system, retain counsel. If you need the file regardless of the answer, retain us.
The major consultancies sell general-purpose AI governance frameworks. We sell one productized engagement for one use case. Their pyramid bills your time learning hiring AI. We have already done that learning, and we price like a specialist firm with no leverage to monetize.
Their proposals take six weeks. Ours is on this page.
Vendor cooperation is a covered risk in the Article 9 file we produce — and a common one. Where the vendor declines, the file documents the request, the refusal, the workaround we engineered (often a deployer-side conformity record plus an explicit residual-risk acceptance), and a vendor-engagement letter you can send if you choose to escalate. We have produced conformity packs against uncooperative vendors before.
Our methodology covers AI assessment platforms, sourcing tools, conversational-AI screeners, resume-screening models, and workforce-analytics tools used in hiring. If your stack includes something outside that scope, we will tell you so on the intake call, before you commit. Out-of-scope tooling is a hard scope boundary, not a billable surprise.
You receive the file, a 90-minute internal walkthrough, and 30 days of revision support. After that, the engagement ends. We do not sell retainers, ongoing managed governance, or annual subscriptions. Repeat engagements happen — usually a year later, for the next tool — but they are sold the same way: fixed scope, fixed price.
Not at launch. The methodology is purpose-built for hiring AI, and we will be wrong about adjacent domains for at least a year. If we expand, it will be deliberately, with a separate productized pack and a separate body of regulatory work behind it.
No. Casework is a specialist services firm. We do not practice law and we do not issue legal opinions. The artifacts we produce are technical and procedural records intended to be reviewable by counsel, by internal compliance, and by regulators. Retain qualified counsel for legal opinions.
Tell us what tool you are documenting, what regimes you are exposed to, and when you need the file in hand. We respond within one business day with a scoped quote or a clear "not the right fit."
Start your intake →