NDA vs Non-Compete vs Non-Solicit: Which Restrictive Covenant Do You Need?

An NDA protects information. A non-compete restricts where someone can work. A non-solicitation restricts who they can approach (your customers or your employees) after they leave. Three tools, three problems. The expensive mistake in-house teams make is reaching for a non-compete to do a job an NDA or non-solicit would do. They then discover the non-compete is void in the employee's state, and the real protection was never drafted.

If you want the foundation, start with what an NDA is and how confidentiality works. This post assumes that and goes to the harder question: given a specific person in a specific state, which covenant (or combination) actually holds up.

TL;DR

• NDA / confidentiality: protects defined information. Broadly enforceable, low-risk, use it almost everywhere.
• Non-solicitation: bars poaching customers and/or employees for a set time. Usually enforceable if narrow, but California voids most of them outside trade-secret protection.
• Non-compete: bars working for competitors. The hardest to enforce; void in California, Minnesota, Oklahoma, and North Dakota, and gated by salary thresholds in Washington, Colorado, Illinois, Oregon, and others.
• No federal ban. The FTC's 2024 non-compete rule was struck down and the FTC dropped its appeal in September 2025, so enforceability is purely a state-law question.
• They stack. Most employment and vendor situations use an NDA plus a customer non-solicit; the non-compete is the exception, not the default.
• Choose by role, state, and access, not by habit. An overbroad non-compete creates false comfort; the narrow, enforceable core is usually the NDA plus a scoped customer non-solicit.

NDA vs non-compete vs non-solicit: the practical difference

Each covenant answers a different question, and each faces a different enforceability bar.

The three are not interchangeable. A court reviewing a non-compete asks whether you had a legitimate interest that a lesser restraint could not protect. If an NDA and a customer non-solicit would have done the job, the non-compete looks like a naked restraint on competition, and that is where it gets struck.

The 2026 legal backdrop: no federal ban, and state law that keeps moving

For a stretch of 2024 it looked like non-competes might be banned nationwide. The FTC issued a rule in 2024 that would have voided most existing non-competes and barred new ones. That rule never took effect. In Ryan LLC v. FTC, the Northern District of Texas set it aside. The court held the FTC lacked authority to issue it. In September 2025 the FTC voted to drop its appeal and accept vacatur. The practical result: there is no federal non-compete ban, and none is coming from that rule. Enforceability is a state-law question, full stop.

State law is where the action is, and it splits three ways.

Outright bans. California voids nearly all employee non-competes under Business and Professions Code section 16600. That statute makes any contract restraining a lawful profession or trade void to that extent. Minnesota banned new employee non-competes in 2023. North Dakota and Oklahoma void them by statute. In these states, an employment non-compete is void on its face, so budget your protection around the NDA and the non-solicit instead.

Salary-threshold gates. A growing group of states enforces non-competes only above an income floor, and the floors rise most years. Washington sits around $123,000, Colorado around $127,000 for non-competes (and a separate, lower floor near $76,000 for non-solicits), Oregon around $116,000, and Illinois at $75,000. Massachusetts ties enforceability to FLSA exemption status rather than a flat number and adds a garden-leave or other-consideration requirement. Below the threshold, the non-compete is void regardless of how narrowly you drafted it. This is easy to miss when a template travels from a headquarters state to a lower-paid role in a threshold state.

Reasonable-restraint states. Delaware, New York, Texas, and Massachusetts (above its threshold) will enforce a non-compete that protects a legitimate interest and is reasonable in scope, duration, and geography. Texas additionally ties enforceability to consideration reasonably related to the interest, which is why Texas non-competes often ride on trade-secret access rather than a bare promise. The same clause that is void in California can be fully enforceable in Delaware and merely narrowed in New York, which is exactly why the employee's state of work and residence has to drive the draft, not the company's home state.

Practitioner judgment calls

The rules are only half the job. These are the calls that decide whether a covenant helps you or just sits there looking protective.

When a customer non-solicit beats a non-compete. If the real risk is a departing salesperson taking accounts, a customer non-solicit scoped to the accounts the person actually serviced is narrower and more enforceable than a non-compete. In most states it survives where a non-compete would not, because courts read it as protecting relationships you paid to build rather than blocking someone from earning a living. California is the big exception. There, even a customer non-solicit can be void under section 16600 outside trade-secret protection, so a California relationship risk points you to a tight NDA and trade-secret enforcement, not a non-solicit.

When an employee non-solicit still creates California risk. Non-solicits feel safe, but California treats even employee-facing non-solicits skeptically. After AMN Healthcare (2018), a California Court of Appeal struck an employee non-solicit as a section 16600 restraint, and Edwards v. Arthur Andersen had already rejected the "narrow restraint" theory other states use to save these clauses. In California, lean on trade-secret protection and a tight NDA rather than a non-solicit you may not be able to enforce, and do not assume a non-solicit is the safe fallback there.

When continued employment is or is not enough consideration. For a covenant signed at hire, the job offer is the consideration and there is no issue. For one signed by an existing at-will employee, states diverge. Some accept continued employment as sufficient; Illinois generally requires roughly two years of continued employment (or other consideration) before a covenant sticks; and Massachusetts requires garden leave or other mutually agreed consideration by statute. If you are rolling out covenants to current staff, "you still have a job" is not reliably enough. Pair the covenant with a raise, a bonus, equity, or a promotion to anchor the consideration.

When a sale-of-business non-compete belongs in its own bucket. A non-compete tied to selling a business is a different animal and courts treat it that way. Even California enforces sale-of-business non-competes within statutory limits, because the buyer is protecting the goodwill it just paid for, not restraining a mere employee. Keep it in the purchase agreement, tie it to the goodwill and the sale price, and do not let it get analyzed under the far stricter employment standard.

A covenant-selection framework

Skip the generic decision tree. Choosing well means walking a specific set of inputs before you pick a clause.

1. Role and access. Is this person client-facing, technical, or executive? A recruiter with no customer accounts needs different protection than a sales lead or a principal engineer with source-code access.
2. State of work and residence. Run the employee's state first, then yours. A void state (California, Minnesota, North Dakota, Oklahoma) kills the non-compete before you start. A threshold state kills it below the salary floor.
3. Trade-secret access. Real access to secrets is the strongest legitimate interest and the surest ground for the strongest restraint. No secret access usually means an NDA plus a customer non-solicit is your ceiling.
4. Customer control. Did this person own the relationship, or just support it? Scope any non-solicit to accounts they actually serviced; blanket "all customers" language invites narrowing.
5. Compensation level. Compare pay against the state threshold before you draft. Below it, the non-compete is void and you are back to NDA plus non-solicit.
6. Consideration. New hire? The offer covers it. Existing employee? Add real consideration and watch Illinois and Massachusetts specifically.
7. Duration and geography. Shorter and narrower survives. One to two years for a non-solicit, six months to two years for a non-compete where allowed, and geography no wider than where the person actually operated.
8. Blue-pencil risk. Know your state's rule, because it decides what happens to an overbroad clause. Reformation states let a court rewrite the clause to something reasonable; Texas and Florida courts routinely do this. Strict blue-pencil states let a court delete the offending words but not add new ones. Some states refuse to save an overbroad covenant at all and void it, an approach Virginia and Wisconsin courts have taken. In a void-it state, over-drafting can lose you the protection entirely, so draft to survive, not to intimidate.

Run those inputs and most situations resolve to NDA plus a scoped customer non-solicit. The non-compete enters only for senior, high-access roles in states that still enforce them, above any salary threshold, and even then it should be as narrow as the legitimate interest allows.

Three quick walkthroughs show how the same instinct produces different clauses:

• California sales rep leaving with CRM access. The non-compete is void, and so is the customer non-solicit outside trade-secret protection. Your enforceable core is a tight NDA plus trade-secret enforcement on the customer data itself. Do not paper over it with a non-solicit that will not hold.
• Washington mid-level engineer earning below the threshold. The non-compete is void because the pay does not clear the state floor, even though the engineer has real source-code access. Protect the code with a strong NDA and invention-assignment terms, and use a customer non-solicit only if it fits (an engineer usually has no customer relationships to solicit).
• Delaware founder selling a company. Now the non-compete belongs in the purchase agreement, tied to the goodwill and the sale price. Analyzed as a sale-of-business restraint, it is enforceable well beyond what an employment non-compete would allow, so keep it out of any side employment agreement where a court would judge it by the stricter standard.

A drafting checklist for the covenant you land on

Before you sign or send, walk the specific clause-level issues that decide enforceability:

• Definition of confidential information: specific and bounded, with the standard exclusions (public, independently developed, already known, lawfully received from a third party).
• Residuals carveout: present or absent by design, not by accident; it changes how much the NDA actually protects.
• Customer scope: limited to accounts the person serviced or learned about, not the entire book.
• Employee-solicit scope: limited to staff the person worked with, and check the state before relying on it at all in California.
• Duration: one to two years for a non-solicit; six months to two years for a non-compete where allowed.
• Geography: no broader than where the person actually operated.
• Consideration: identified and adequate, especially for covenants signed after hire.
• Notice or garden leave: required in some states (Massachusetts) for a non-compete to hold.
• Blue-pencil and severability: drafted for your state's narrowing rule so an overbroad phrase does not sink the whole clause.

For the clause mechanics, see the non-compete clause breakdown, the non-solicitation clause breakdown, and the confidentiality clause breakdown.

Where AI helps

Restrictive covenants are exactly the kind of clause worth checking against a standard before you sign or send. The review is mechanical once you know what to look for: identify the covenant type, confirm the employee's state allows it (and that pay clears any threshold), check that the non-solicit is scoped to serviced accounts, confirm the confidentiality definition has its exclusions, and flag blue-pencil risk in a strike-it-all state. That is a checklist, and checklists are automatable.

In Vaquill AI, that runs as a structured review against your playbook, flagging the covenant type, the scope and duration, and the state-enforceability risk so a routine agreement clears fast and only the real problems reach a lawyer.

FAQ

What is the difference between an NDA and a non-compete? An NDA protects confidential information: it stops someone from disclosing or misusing defined secrets. A non-compete restricts where and for whom a person can work after they leave. Different purpose, and the non-compete is far harder to enforce because it restrains someone's ability to earn a living.

What is the difference between a non-compete and a non-solicitation? A non-compete bars working in the field or for competitors at all. A non-solicitation only bars going after your specific customers or employees. The non-solicit is narrower and generally more enforceable, though California treats even non-solicits skeptically.

Can you use all three together? Yes, and senior-employee agreements often do. An NDA protects information, a non-solicit protects relationships, and a non-compete (where enforceable) protects against direct competition. Each should be scoped to its own legitimate interest rather than copied broadly, because an overbroad clause can drag down the enforceable ones.

Which restrictive covenant is the most enforceable? The NDA. Protecting confidential information is a well-accepted, low-risk restraint. Non-solicitations come next if narrow. Non-competes are the least enforceable and are banned or salary-gated in a growing list of states.

Are non-competes still enforceable in 2026? It depends entirely on the state, because there is no federal ban. The FTC's 2024 rule was struck down in Ryan LLC v. FTC and the FTC dropped its appeal in September 2025. Non-competes are void in California, Minnesota, Oklahoma, and North Dakota, and gated by salary thresholds in Washington, Colorado, Illinois, Oregon, and others. Check current state law before relying on one. See non-compete enforceability by state.

Is a non-solicitation agreement enforceable? Usually, if it is narrow: a defined group of customers or employees the person actually worked with, a reasonable duration (often one to two years), and a legitimate business interest. California is the big exception, where even employee non-solicits have been struck under section 16600. Blanket non-solicits covering everyone tend to be narrowed by courts.

Do I need a non-compete to protect my customer list? Usually not. A customer list is typically protected better by an NDA and trade-secret law, plus a customer non-solicit scoped to serviced accounts. Those survive in states where non-competes are void, so they are often the stronger play.

For more, see what an NDA is, the NDA vs confidentiality agreement explainer, and non-compete enforceability by state.

Arshita Anand

Co-Founder & CEO · Attorney

Arshita leads product and strategy at Vaquill, building the legal AI suite that solo, small-firm, and in-house US lawyers use to run a matter end to end.