BlueHive HUM
Use case

Voice agents that delegate follow-up work to background agents

Bea keeps the conversation going while background agents do the heavy lifting — looking things up, updating your systems, and finishing the follow-up your team used to do after every call.

The call ends. The work doesn’t.

Most calls create a to-do list — and today that list lands on your staff.

Callers wait while systems catch up

Put a caller on hold to look something up, and every extra second raises the chance they hang up before you have an answer.

After-call work piles up

Updating records, opening tickets, sending summaries, and chasing answers eats hours after the phone is back on the hook.

Automation is a black box

When follow-up steps are scripted somewhere else, nobody can see whether they ran, what they did, or what they cost.

How it works

How background agents work

Bea handles the conversation. Background agents handle what comes next.

  1. Step 1

    The caller asks for something

    Bea handles the conversation and recognizes when a request needs more work than a quick answer.

  2. Step 2

    Bea hands it off

    She passes the task to an approved background agent and keeps talking — no hold music, no “we’ll call you back.”

  3. Step 3

    The background agent works

    It uses your connected tools and knowledge to research, update records, open tickets, or draft replies.

  4. Step 4

    Watch and review

    Follow each run live, see every tool it used, and get results by webhook where your team already works.

What you get

Multi-agent follow-through, built into every call

Your voice agent stays focused on the caller while a team of background agents finishes the job.

Mid-call task handoffs

Voice agents start follow-up work during the call, and you choose which background agents they’re allowed to hand work to.

Tool-triggered follow-ups

When a selected tool succeeds — a booking, a lookup, a record update — the next step can start automatically.

Event-driven agents

Start agents when a voicemail arrives, a call completes, an outcome is captured, or your own systems send a webhook.

Live activity and full history

Follow each run as it happens, review every tool result, and download complete activity logs for later review.

Budgets and cost tracking

Set organization and per-agent daily limits, and see background-agent costs next to your calls and analytics.

You control the context

Choose exactly which caller details, transcripts, and tool results each agent receives. Only admins can change these settings.

FAQ

Common questions

What is a background AI agent?
A background agent is an AI agent that works on a task without a live conversation attached. Bea — your voice agent — can hand it a task mid-call, or it can start on its own when something happens, like a voicemail arriving or a call ending. It uses the tools and knowledge you connect, and every step it takes is recorded.
Does the caller have to wait while the background agent works?
No. Bea hands off the task and continues the conversation right away. The background agent works in parallel, and its results are delivered to your team, your systems, or a webhook when it finishes.
What kinds of work can background agents do?
Anything your connected tools allow: researching an account, updating your CRM or ticketing system, drafting a follow-up message, summarizing a voicemail for the right team, or escalating an issue. You decide which tools each agent can use.
Can follow-up work start automatically?
Yes. A background agent can start after a selected tool succeeds during a call, and you choose whether that tool’s details are shared with it. Agents can also start from voicemails, completed calls, captured outcomes, or your own systems through the API.
How do we keep costs and data under control?
Admins set daily limits for the organization and for each agent, and background-agent costs appear alongside call costs and analytics. Caller details and transcripts aren’t shared by default — you pick exactly what each agent receives.
What if the same task gets triggered twice?
When a call hands the same task to the same agent again within an hour, or your systems resend a request with the same ID, HUM returns the original result instead of starting the work over. If a run fails, the task can simply be tried again.
Related workflows
AI agent APIOverflow call answeringAfter-hours answeringAll featuresIndustriesPricing

Let Bea finish the job, not just the call

Connect a background agent to your first voice agent and watch the follow-up happen on its own — free for 14 days, no credit card required.