What happens if a client's LinkedIn account gets restricted?
You do, commercially, whatever the vendor's terms say. The tool loses a seat; you lose a relationship and possibly the retainer. That asymmetry is the reason agency-side rules should be stricter than anything a vendor enforces, and the reason it belongs in the contract before the first invite.
The asymmetry nobody prices in
When a client's account gets restricted, the vendor's exposure is one churned seat and a support ticket. Yours is the relationship, the retainer, and a conversation where the client is angry about something they only half agreed to. The incentives are not aligned and pretending otherwise is how agencies get caught out.
This is also why vendor reputation is the only real enforcement mechanism in this category. Tools that got people restricted lost the market to tools that did not, and that happened through reviews and word of mouth rather than through anything contractual.
Have the conversation before the first invite
The client should know, in writing, three things: that no third-party outreach tool is compliant with LinkedIn's User Agreement, that the realistic downside is an invitation restriction rather than a closed account, and what you will do if it happens.
Agencies skip this because it sounds like talking yourself out of the deal. In practice it does the opposite. A client who signed knowing the trade is a client who stays when the warning arrives; a client who finds out from LinkedIn is a client you lose that week.
Pair it with the operational question of whose account it is. Running outreach from a founder's primary profile with ten years of relationships is a materially different bet from a seat created for the campaign, and the pricing should reflect which one you are being handed.
The rules that keep it from happening
Ramp every new client seat rather than starting at full volume. A seat that has never sent an invitation should not send twenty on day one, and the ramp should be automatic rather than a note in someone's calendar.
Never let two seats approach the same company in the same week. On a multi-client roster with overlapping ICPs this happens constantly, it looks exactly like coordinated automation from the company's side, and it is the collision no spreadsheet catches.
Keep suppression global across every seat you run, not per campaign. The person who said no to one client should not hear from another one a fortnight later.
Halve volume on any seat whose acceptance rate drops below 25% in a week. Falling acceptance with steady volume is the pattern detection is built for, and the correct response is to send less while you fix targeting.
What to ask a vendor before you put client seats on it
Are the caps enforced server-side or are they a number in a settings field? A settings field is not a cap, it is a default, and the person who raises it will be a junior on your team at 6pm on a Thursday.
What happens when a cap is reached? There are three behaviours: refuse and tell you, queue and send tomorrow, or overshoot and report success. The third is the one that costs you an account and it is never in the marketing page.
Is there a browser extension anywhere in the architecture? Extensions that modify the page are the one thing the platform can detect directly rather than infer.
Is there a published restriction rate with a methodology, denominator and date? A number without those three is not evidence. LinkedBoost currently reports measuring, because a defensible production percentage is not yet available.
Who carries what when a seat gets restricted
| Party | Exposure | Mitigation that is actually available |
|---|---|---|
| Client | Their own profile and network | Informed consent in writing, before the first send |
| Agency | The relationship and the retainer | Enforced caps, ramping, collision and suppression rules |
| Vendor | One seat, and reputation over time | Architecture, enforced limits, published methodology |
Questions people ask next
Should agencies tell clients that LinkedIn outreach tools break the terms?
Yes, in writing, before the first invitation. LinkedIn's User Agreement section 8.2 prohibits automated messaging and contact-adding, and the client is the account holder. A client who signed knowing the trade stays when a warning arrives.
Can an agency be held responsible for a client's restricted account?
Contractually it depends on what you signed, and most vendor terms push the risk to the account holder. Commercially it lands on you regardless, which is the number that matters when you price the work.
What is the safest way to run outreach on a client's seat?
Ramp new seats rather than starting at volume, keep suppression global across the roster, never let two seats hit the same company in a week, and cut volume when acceptance drops rather than pushing through it.
Why this page exists: Agency objection cluster; the commercial-asymmetry point is the one raised most often in founder and agency communities when a tool is recommended.
LinkedBoost is the LinkedIn MCP server: your agent sources, drafts, sends and works the inbox, inside caps the server enforces rather than suggests.
Related answers
Yes. LinkedIn's User Agreement section 8.2 prohibits bots and unauthorised automated methods, and every third-party outreach tool sits against that clause, ours included. What the clause does not say is that using one gets you banned. Enforcement is behavioural and it arrives as a ladder, not a switch.
One company, one seat, one week. It is a practice rather than a setting: before enrolment the agent checks company interaction history and searches active sequences for that company, and every CSV you import turns past outreach into suppression. Nothing in the product silently blocks a second seat for you.