How Expert Networks Screen for Non-Competes and Garden Leave: A Field Guide for Buy-Side Compliance

A consulting expert bound by an active non-compete, non-solicit, or garden-leave clause is not a hypothetical compliance issue. It is a documented source of enforcement risk for the expert, the employer, the expert network, and the buy-side client that commissioned the call. Every major network has built a screening stack for this problem, and the stacks differ in ways that matter to research-ops leads and compliance officers picking a vendor or reviewing a workflow.

This is a field guide to those layers , the legal backdrop they sit against, the controls networks apply, and the questions a buy-side team should be asking when it evaluates coverage.

The Legal Backdrop: A Patchwork, Not a Rule

Post-employment restrictions on experts sit on shifting legal ground. In the US, non-competes are enforceable in most states under a reasonableness standard, but banned or heavily narrowed in California, Minnesota, Oklahoma, North Dakota, and , as of the most recent round of state-level reform , Washington. The Federal Trade Commission published a rule in 2024 that would have banned most non-competes nationally, but the Northern District of Texas set it aside in Ryan LLC v. FTC in August 2024, and the case is on appeal. Until the appellate ruling lands, the patchwork stands.

Non-solicit clauses, confidentiality agreements, and trade-secret obligations survive independent of any non-compete framework, and they often bite harder in the expert-network context because a single consultation can plausibly touch customer identity or proprietary process detail.

Outside the US, the picture is different but not simpler. UK courts enforce garden-leave provisions and post-termination restrictive covenants under a reasonableness test refined by the UK Supreme Court in Tillman v Egon Zehnder, which clarified how a court will sever an unenforceable clause from an otherwise valid covenant. European jurisdictions vary widely on duration and compensability. A network operating globally is screening against a moving target in every region.

The practical implication: no single onboarding attestation can carry the weight. Networks have built layered controls because the underlying law is too fragmented for a single gate.

Layer 1: Self-Attestation at Onboarding

The base layer at every major network is the expert agreement itself. When an expert joins GLG, AlphaSights, Guidepoint, Third Bridge, Dialectica, or Coleman, they sign terms that require self-certification that no active non-compete, non-solicit, NDA, or confidentiality obligation prevents the specific consultation being arranged. The language is contract-of-adhesion standard and has been referenced in SEC enforcement matters against individual experts and their institutional counterparties.

Self-attestation is a necessary layer, not a sufficient one. It shifts contractual risk to the expert if they misrepresent, but it does nothing to catch experts who genuinely do not know whether their old employment agreement covers the topic at hand , a common situation for anyone who left a role more than three years ago and never re-read the paperwork. Networks that treat attestation as the primary control are running thinner than networks that layer additional screening on top.

Layer 2: Current-Employment and Recent-Employer Screening

The second layer is structural: networks screen for experts who are still employed by, or recently departed from, a company in the topic area. Most operate a lookback window , typically six months, sometimes twelve , during which an expert cannot be consulted on their current or recent employer.

This control is not designed to catch non-competes specifically. It is designed to catch the higher-order risk that a currently-employed expert will disclose material non-public information about their own company, which is the fact pattern the SEC's expert-network enforcement history has repeatedly punished. The 2013 case against Diamondback Capital and Level Global involved a chain of consultations with technology-company insiders. The 2011 action against a healthcare consultant tracked a similar pattern. In both, the expert's employment status was the compliance red flag that the network's screening was meant to catch.

The practical overlap with non-compete screening is that current-employer restrictions are usually the strictest, so any control that blocks calls involving a current employer sweeps up a large share of the active-restriction risk as a side effect.

Layer 3: Cooling-Off Windows Beyond Contract

Some networks impose their own cooling-off periods on top of whatever the expert's contract says. Depending on the network and the client's own policy overlay, a former employee of a public company may be blocked from being consulted on that employer for three, six, or twelve months after departure, regardless of whether their original non-compete has expired.

The rationale is that even a lapsed non-compete does not necessarily lapse the underlying confidentiality obligation, and even a lapsed confidentiality obligation does not necessarily lapse the reputational risk to the network if a call goes badly. Cooling-off windows are a network's own belt-and-suspenders layer, not a legal requirement. Buy-side teams evaluating coverage should ask specifically what the window is, whether it varies by expert seniority, and whether the client can request a longer window as a policy overlay.

Layer 4: Topic Scoping

The fourth layer is the most nuanced and probably the highest-leverage. Compliance teams scope the call topic so it stays outside the specific restrictive covenant an expert is subject to. An expert with a customer non-solicit can be consulted on industry-level commentary but not on named customer relationships. An expert with a narrow technology non-compete can discuss market dynamics but not product roadmap detail from their prior employer.

Topic scoping is where the network's compliance team earns its keep. It requires reading the actual covenant (or the expert's summary of it, with appropriate caveats), understanding the buy-side team's research question, and finding the intersection where the call adds value without crossing a restriction. This is not a scalable, automatable control , it depends on trained reviewers, and its quality varies with the compliance team's depth. It is also the layer that most differentiates networks from each other in practice, though not one that lends itself to a public comparison.

Layer 5: Client-Side Disclosure and Re-Screening

Networks pass employment history and, where known, active restrictions to the buy-side client before the call. The client's compliance function is expected to re-screen , not as a redundant check, but as an independent obligation.

The SEC's enforcement pattern treats the buy-side firm and the network as separately accountable. A network's clean screening does not immunize a hedge fund whose own compliance workflow failed to catch a problem. Firms that treat network disclosure as the end of the compliance chain are misreading how the regulator has approached these cases historically. The re-screen typically checks the expert against the firm's restricted list, the covered-company list for the requesting analyst, and any firm-specific policies on former employees of portfolio companies or short targets.

Layer 6: Chaperoning and Post-Call Review

For higher-risk experts , those with known active restrictions, or those whose topic sits close to the edge of a covenant , networks add a live layer. A compliance chaperone joins the call, the session is recorded (with consent), and the transcript is reviewed after the fact. If the conversation drifts into restricted territory, the chaperone can intervene in real time, and the post-hoc review provides a defensible record if a question later arises.

Chaperoning is expensive and slow. Networks reserve it for the cases where the topic-scoping layer alone is not enough confidence. Buy-side teams should understand which of their calls are being chaperoned by default and which are being run on attestation-plus-scoping alone.

How to Read a Network's Stack

The six layers are not equally implemented across networks, and no network publishes a detailed breakdown of how its controls interact. A buy-side compliance officer or research-ops lead evaluating coverage should be asking specific, comparable questions rather than accepting a general assurance that screening exists.

The questions that matter: What is the recent-employer lookback window, and can we lengthen it? Who reads the actual restrictive covenants when a match is flagged, and what is their compliance training background? How is topic scoping documented so we can audit it later? What is the default trigger for chaperoning, and who bears the cost? How does the network hand off employment-history data to us, and in what format does our own compliance system ingest it?

Answers to those questions expose the real shape of a network's stack. General statements about a compliance program do not.