Sub-accounts and agent extensions: keep clients and buyers separate

Create one sub-account per client or buyer, then add agents with unique extensions and choose each agent’s endpoint (browser softphone or SIP softphone). That setup keeps every client or buyer isolated so they see only their own numbers, calls, and recordings, while agents answer on the app you pick. Your first client setup takes longer while you learn the pattern; after that it goes faster as you reuse it.

By the end of this how-to, you’ll have a clean layout. Two client brands and one pay-per-call buyer sit in their own sub-accounts. Agents in India and the Philippines answer on a browser or a SIP app. You’ll route calls by brand and campaign, cap volume to buyers, and keep reports separate so there’s no cross-client data bleed.

How do I set up sub-accounts and agent extensions for clients and buyers?

Create a sub-account for each client or buyer, then add agents with extensions and set their endpoint (browser softphone or SIP softphone). Each sub-account holds its own numbers, campaigns, targets, and recordings so only the right people can view them. In our running example, the agency runs two brands, Plumbing and Solar, plus a buyer team. Agents are split between India and the Philippines. On CallFlow, a sub-account is a separate workspace, so the Solar brand never sees Plumbing data and buyers see only leads routed to them.

Plan the extension ranges now. Give Plumbing 101–199, Solar 201–299, and the buyer team 301–399. Decide which agents will use a browser softphone and which will use a SIP softphone or desk IP phone. That choice affects setup time and network checks later.

Account architecture for agencies: clients, buyers, and teams

Sub-accounts are separate workspaces. Put each client brand and each buyer in its own sub-account to fence off ownership, recordings, and reports. Client users log into their brand’s sub-account and see only their numbers and call logs. Buyer users log into a buyer portal tied to the buyer sub-account and see only leads routed to them. Numbers, campaigns, targets, and recordings live inside a sub-account and do not appear in any other.

Keep the call path simple. Each brand’s tracking numbers point to the brand’s router in that same sub-account. The router tries active agent extensions for a set time (20 seconds is a good start), then sends the call to the queue and, if still unanswered, to voicemail.

Step-by-step: sub-accounts setup and access roles

Invite clients and buyers without exposing other data

Follow this once for Plumbing, then repeat for Solar and your buyer team. The screen names may vary slightly by country, but the flow is the same.

1. Create the sub-account

• Why: This is the wall between brands and buyers. It keeps numbers, campaigns, recordings, and reports separate.
• Do: From the agency dashboard, add a sub-account. Use a clear name like “Plumbing – US” or “Solar – CA North.” If you run regional teams in India or the Philippines, include the city or shift window in the name, such as “Buyer Team – IST Day.”

1. Provision numbers inside the right sub-account

• Why: Numbers must live where the routing and reporting live.
• Do: Open the Plumbing sub-account and buy US toll-free or local numbers (or numbers in other supported countries) from the prepaid wallet. Name or label each number by campaign, for example “Plumbing_GoogleAds_Houston.”

1. Attach numbers to a clean entry point

• Why: Every number should feed the right router or IVR so DNIS (the number dialed) stays tied to one brand and campaign.
• Do: In Plumbing, point numbers to the Plumbing router, or to a short IVR if you need simple menus first. Keep IVR options short: “Press 1 for Service, 2 for Sales.” If you operate a single “front door” line for multiple brands, host that IVR in a neutral “Front Door” sub-account. Make each option forward to the brand’s published number inside its own sub-account.

1. Add users with the least access they need

• Why: This protects client data and keeps buyers focused on the leads they paid for.
• Do: Add agency admins to all sub-accounts. Add client users to their brand’s sub-account. Add buyer users to the buyer sub-account so they can see only their lead flow in the buyer portal.

1. Build campaigns and targets inside each sub-account

• Why: Campaign-level reporting must not mix clients.
• Do: Create campaigns per brand and state/region if you route by geography. Add targets that point to agent extensions or buyer endpoints in that same sub-account. For routing patterns and when to use priority, weighted, or round-robin, see routing methods.

1. Schedule email reports per sub-account

• Why: Clients and buyers should receive only their reports.
• Do: In Plumbing, set daily or weekly scheduled reports to the client’s list. In the buyer sub-account, add the buyer’s list. Keep subject lines clear by brand and region. See examples in scheduled call reports.

1. Plan so you do not have to move assets later

• Why: Moving numbers or recordings across sub-accounts is painful. It breaks links and reports.
• Do: Before you buy numbers, list the brands, regions, and buyer groups you expect. Create sub-accounts first, then place assets inside the right one from day one.

Confirm it worked: From the client’s login, you should see only that brand’s numbers, calls, and recordings. From the buyer portal, you should see only the buyer’s routed leads. No other brand or buyer data should appear.

Assign extensions to agents and choose endpoints

Create the agent records inside each sub-account, give each a unique extension, and pick the endpoint per agent. In Plumbing, assign 101–199. In Solar, assign 201–299. For the buyer team, use 301–399. Note each agent’s time zone and shift, then set per-target call windows so schedules line up later.

If an agent needs to take calls on a mobile during outages, add a mobile number you control as a fallback target. Keep ring timers short so you do not fall into a voicemail race. A 15–25 second ring per ring step, then queue, works well. Do not forward straight to personal mobile voicemail.

Browser softphone vs SIP softphone

You can let agents answer in a browser softphone (a call app in the browser) or a SIP softphone (a VoIP app that registers via SIP, or a desk IP phone). QoS (quality settings on your network) and NAT (network address translation) can affect how SIP apps behave. Pick per agent based on network and work style.

If you roll out a browser team first, follow the mic and device checks in browser team setup. For SIP apps or desk phones, confirm registration and keep-alive timers as in SIP softphone registration. In some networks in Bangladesh or Pakistan, carriers or ISPs may block or shape SIP. In that case, use the browser softphone, or run SIP over a network path your IT team controls with the right firewall rules.

Tip: Do not mix endpoints for a single agent without testing. If you register the same extension on a SIP app and also use the browser softphone, check ring order and voicemail behavior. It is safer to give the agent one primary endpoint and keep the other disabled unless needed.

Route calls cleanly across clients

Separate the entry points. Use different tracking numbers per brand and per campaign to avoid cross-client reporting bleed. Put each number in the brand’s sub-account so reports match billing and recordings.

If you must use a shared front door, keep the IVR short. “Press 1 for Plumbing, 2 for Solar.” Each option should forward to the brand’s published number that lives in its sub-account. That number then hits the brand router, which fans out to that team’s agent extensions. This keeps each sub-account in control of its own queue, voicemail, caps, and reports.

Pick routing methods by need. Use priority routing when a lead closer is on duty and backups sit behind them. Keep each ring step at 15–25 seconds to avoid personal voicemail. For a fair split across a team, pick weighted or round-robin routing. Use simultaneous ring only for short spikes, because it can trigger voicemail races if someone’s mobile grabs the call. If you must use it, make device voicemail timeouts longer than your queue timeout.

Respect time. Set per-target call windows (the hours an agent or buyer takes calls) so after-hours calls go to the queue or voicemail. Use a whisper to remind agents of after-hours scripts when they do answer. For a how-to on time-of-day controls, see per-target call windows.

Protect buyers with caps. Daily caps stop over-delivery. Concurrent caps limit how many calls a buyer or agent can handle at the same time. When a cap hits, overflow to an alternate buyer team, a backup agent pool, or the brand queue. Keep the overflow inside the same sub-account so data stays clean.

Operations: QA, monitoring, and common mistakes

Run a full end-to-end test before you turn up ads. Start with caller ID passthrough: place a call from your mobile and check that the agent sees the original caller ID, not the tracking number. Check the call logs tag the right brand and campaign, and make sure agents know which line rang for each brand. Let the call ring to the edge of your timer to confirm that queue overflow and voicemail trip when they should.

If you involve real people, get consent and follow your local recording disclosure rules during tests. Play test recordings from the client’s login. If you do not want buyers to access recordings, disable recording on those routes and send scheduled billable-call reports instead. Use a short IVR greeting to inform callers of recording; use a whisper only to remind agents of what to say.

Watch live call logs during your tests. Confirm which attempt answered, the ring timers, and the router used. The sub-account should display only the data that belongs to it. If you see Solar data while logged into Plumbing, stop and check your number and router placements.

Fix the two common endpoint pitfalls early. For browser softphones, grant mic permission and pick the right input device in the app. Headsets help avoid echo. For SIP softphones, confirm the register status, the expire interval, and NAT settings. If calls ring once then drop, you may have a firewall or NAT issue.

Keep one brand per sub-account. Mixing brands blends recordings and reports and makes buyer access risky. Do not mix endpoints for the same agent without testing, or you will see double-rings and voicemail races. Keep ring timers shorter than device voicemail so personal voicemail does not grab calls. When agents sit in India or the Philippines and your callers are in the US, set call windows in the agent’s local time and double-check schedules against each team’s local time.

Add failover routes inside the same sub-account to meet response targets. For example, if the Plumbing closer does not answer in 20 seconds, spill to the backup pool, then queue, then voicemail. Do not forward to another brand’s line or buyer line across sub-accounts. Cross-account transfers create reporting leaks and confuse ownership of the recording.

Put this live, then lock it down

Turn on a small ad group or source per brand and watch the first day’s calls in the live log. Confirm caps trip when expected and that buyers see only their leads. Share scheduled reports with clients and buyers so they get the right view without logging in every time. If you have not picked a platform yet and want to build this layout with a prepaid wallet and no contracts, you can start here: /register