Multi-tenant outreach.
Keep every client separate.
You run outbound for five to twenty clients. Each one wants their data siloed, their reporting branded, their recipes adapted to their ICP. Coil gives you one workspace per client, org-scoped keys, and the same primitives your operator team uses across the portfolio. Row-level isolation at the database. No cross-tenant leakage by construction.
For agencies running outbound for clients
Who this is for.
You run a small outbound agency. Three to fifteen people, mostly operators, mostly working asynchronously across a client portfolio of five to twenty accounts at any given moment. Your delivery shape is a mix of list building, copy drafting, sequence management, and weekly reporting back to each client. Your tooling shape is the accumulated debt of two years of point solutions: an Apollo seat for one client, a Clay graph for another, a homegrown n8n flow for the third, and a shared Notion that nobody updates after Wednesday.
The painful part of your week is not the work; it is the context switching between client environments and the constant fear of cross-tenant mistakes. Sending the wrong ICP to the wrong client. Showing one client's reply data to another. Reusing copy that was supposed to be exclusive to the first client. Coil exists because the organisation primitive belongs at the platform layer, not in your head and your shared spreadsheets.
The four problems that come with agency-shaped outbound.
Agencies have a different problem shape than in-house teams. Naming them helps decide whether Coil maps to your operating model or whether you should keep what you have.
1. Multi-tenancy.
Every client wants their lead lists, their scrapes, their automation results, and their reply data isolated from every other client on your roster. Tools that bolt tenancy on top of a single-tenant model leak. Coil's organization model enforces the org_id filter at every query in the service layer. One Coil workspace per client, switched at the navbar, with no shared row anywhere in the database between two client orgs.
2. White-label expectations.
Some clients want the deliverable to look like it came from your agency, not from a SaaS dashboard they have never heard of. Coil does not ship a white-label dashboard tier in 2026. The pattern that works today is to use Coil as the back-of-house tool your operators run, and ship the client-facing deliverable from your own template (Notion, Google Sheets, a custom client portal) with the data pulled via the agent API. If a true white-label tier matters more than that, this is worth flagging in the feedback widget before signing up.
3. Client reporting.
The reporting weight is per-client and per-week, which means the agency needs a programmable export rather than a single shared dashboard. Coil exposes complete lead runs as formula-safe CSV through the dashboard, API, and CLI. The agency scripts the reformat step; exports for scrapes, automation results, and general activity are not shipped yet.
4. Scope creep.
Clients who watch the dashboard often ask for “just one more thing.” A new segment, a new sequence variant, a new signal type. The agency response that keeps margin healthy is to maintain a master library of recipes and clone them per client rather than writing each campaign from scratch. Coil's recipe primitive is built around that pattern; the cross-org library is still a roadmap item, so the cloning step is scripted today.
How an agency uses Coil.
The pattern that has worked in the beta is one Coil workspace per client, with the agency's operator team carrying org:admin role in each one. Coil's multi-tenancy is native at the platform layer, so the isolation work is done before your operators ever click into a client org.
Onboarding a client. You create a new Coil workspace, invite the client (or keep it operator-only depending on your delivery model), and issue an org-scoped API key for any agent traffic that will run inside that tenant. The key format is ak_... and the org boundary is enforced at the auth layer, not at the business logic layer.
Building lists and copy. Your operator opens a chat with the AI of their choice, switches to the client org context, and runs the prospect skill or the Sales Nav scraper. Recipes from your master library get cloned into the client org via the agent API before the campaign starts. The same agent flow your founder clients run themselves also runs here, behind your operators.
Running the sequences. Sequences queue into the sending tool the client already pays for (Smartlead, Instantly, or whatever sender the client has warmed up). Coil holds the list and the audit log; the sending tool holds the deliverability story. Per-key rate limits contain any runaway agent inside the client org rather than letting it burn the agency's minute budget across the portfolio.
Reporting at end of week. The operator runs the CLI against the client org, exports the lead run as CSV, and feeds it into the client's reporting template. Each export has a durable actor-attributed receipt; general activity-log export and automation-result export are not shipped yet.
Where Coil is not the right answer for agencies.
If you need a single sign-on experience that bridges all your client tenants, where your operators authenticate once and Coil sees them as one identity across every client org, that is not what we ship today. Operators sign into each client org through the standard Clerk membership flow. Cross-org operator-identity federation is not on the 2026 roadmap; verify against your operator-experience requirements before signing up.
If your delivery model depends on a fully white-labelled client-facing dashboard with your agency's brand on it, Coil is not there yet. The deliverable today is your own client report generated from Coil's exports. If the dashboard branding is the product, wait for the white-label tier or consider a platform that started life as agency software.
Agency questions, answered honestly.
What agency leads ask in the first conversation.
Is Coil really multi-tenant for agencies today?
Coil isolates every organization at the database row level by an org_id filter on every query. The agency pattern is one Coil workspace per client; you switch between them from the Coil org switcher. There is no cross-tenant data leakage by construction. What does not exist yet is a parent-agency view that aggregates across all client orgs from one screen.
Can I white-label the dashboard for clients?
Not in 2026. The dashboard is branded as Coil. Agencies can use Coil as a back-of-house operator tool and feed its lead-run CSV into their own reporting template. A real white-label tier is still an open roadmap question; use the feedback widget if it is a requirement.
How do I share recipes across client orgs?
Recipes are saved per org today, so the pattern is to maintain a master recipe template and clone it into each client org via the agent API or the CLI. A native cross-org recipe library is a roadmap item; for now the cloning step is what you script. The CLI returns JSON for every command, so the script is small.
How does client reporting work?
Lead runs export per client org as formula-safe CSV through the dashboard, API, or CLI. You can reformat that file into a client-specific template in Notion, Google Sheets, or a custom dashboard. Scrape, automation-result, and general activity exports are not shipped yet.
Keep reading.
The operating method behind Coil is in the Coil methodology pillar. Recipes and the cloning pattern are documented on the recipes page; the agent surface that powers cross-org scripting is the agent API. Isolation, audit log, and the role gates live on the security page.
One Coil org per client. Same operator team behind all of them.
Native Coil workspace isolation. Org-scoped keys per client. Recipes and automations clonable via the agent API. Formula-safe lead CSV export for client-shaped reporting.
For agencies