---
title: "Batch Testing for Customer Agent Across Chat, Email, and Voice"
url: https://www.sidekickstrategies.com/hubspot-updates/hubspot-customer-agent-batch-testing-chat-email-voice
description: "HubSpot's Batch Testing now runs your Customer Agent across Live Chat, Email, and Voice at once so you can catch answer gaps before your customers do."
type: article
updated: 2026-11-16T23:19:05Z
canonical: https://www.sidekickstrategies.com/hubspot-updates/hubspot-customer-agent-batch-testing-chat-email-voice
---

# Batch Testing for Customer Agent Across Chat, Email, and Voice

> **Tip**
>
> TL;DR: HubSpot's Batch Testing now lets Customer Agent users run one test set across Live Chat, Email, and Voice at the same time, so you can find channel-specific answer gaps before a real customer does. Available in public beta to any Professional or Enterprise hub subscriber.

Here's a situation we see constantly in portals: a team spends real time tuning their Customer Agent, runs a solid round of chat tests, marks everything green, and ships it. Then a customer emails the same question and gets a completely different answer.

It's not a fluke. It's a structural gap in how most AI agents get tested. Chat, email, and voice each process context differently. Humans reach you on all three. If you only test one, you only know one-third of what your agent is actually doing.

## What This Update Actually Is

HubSpot has added multi-channel support to Batch Testing inside Customer Agent. Previously, you could run a test set on a single channel at a time. Now you pick which channels to include, run them all at once, and see each channel's response side by side in one test case view.

The three channels now available in the dropdown are Live Chat, Email, and Voice. HubSpot creates a separate thread for each channel for every question in your test set. You review each thread through a channel toggle inside the detail view, and you rate responses as usual.

One important note on rating: when you rate a response, that rating applies to the entire test case across all channels you selected, not to individual channel responses. Keep that in mind when you're using ratings to track quality over time.

This shipped as a public beta on November 17, 2026. Everyone in your account is already enrolled. You can read the official release note at the [HubSpot product update page](https://app.hubspot.com/product-updates/474711/all-updates?rollout=374943).

## Why HubSpot Shipped This

The external problem is straightforward: your customers don't choose one channel and stick to it. They chat on your site, email your support address, and sometimes call. Your AI agent handles all three, but until now you couldn't validate all three in a single testing session.

The internal frustration is subtler. When your agent gives a confident answer over chat but a vague answer over email, you don't always catch it right away. You catch it when a customer escalates, or when a support rep has to clean it up. That's costly, and it erodes trust in the tool you worked hard to configure.

HubSpot is building toward a world where AI agents are genuinely reliable across every surface. This update is a quality-control mechanism designed to close the gap between how you think your agent performs and how it actually performs.

## How to Use It Step by Step

1. Open Customer Agent in your HubSpot portal and navigate to the Batch Testing section.
2. Open the test set you want to run. If you don't have one yet, build a set of questions that reflect real inbound inquiries across your top support topics.
3. In the Channels column, click the dropdown and select the channels you want to test: Live Chat, Email, Voice, or any combination of the three.
4. Click Run. HubSpot generates a separate thread for each selected channel for every question in the set.
5. Open any test case and use the channel toggle in the detail view to switch between Live Chat, Email, and Voice responses for that question.
6. Rate the response as you normally would. That rating covers the full test case across all selected channels.
7. Document any channels where the answer drifts from your expected output. Use those gaps to refine your agent's knowledge sources, response tone, or escalation rules per channel.

## What It Touches in Your HubSpot Strategy

This update sits inside Customer Agent, but it has ripple effects across how you think about AI quality assurance in your whole portal.

Service Hub is the obvious home. If your team uses Customer Agent for support deflection, you now have a repeatable process for checking agent consistency before you push changes to production. Build a standard test set, run it after every major knowledge base update, and treat the results like a release checklist.

> **Key takeaway:** Build a standard Batch Testing set that mirrors your top 10 to 15 support questions. Run it across all three channels every time you update your agent's knowledge sources or instructions. That rhythm catches drift before customers do.

Voice is the channel most teams are still figuring out. If you're already using Customer Agent for SMS conversations, the same logic applies here: each channel has its own context window and tone. Our related update on [Customer Agent for SMS](https://www.sidekickstrategies.com/hubspot-updates/hubspot-customer-agent-sms-conversations) covers how the handoff model works, and that same handoff model applies when Voice responses go sideways and a human rep needs to step in.

Revops and reporting teams should pay attention too. Consistent agent quality across channels is a data quality issue as much as a service issue. If your agent gives different answers on email versus chat, you're generating inconsistent interaction records. That muddies your reporting on resolution rates, escalation triggers, and customer satisfaction scores.

If you've been using AI signals to understand how customers communicate across channels, the broader picture of multi-channel AI quality is worth pairing with what HubSpot is now doing with [Segment Sentiment Analysis across calls, emails, and notes](https://www.sidekickstrategies.com/hubspot-updates/hubspot-segment-sentiment-analysis-calls-emails-notes). That feature surfaces how humans feel about their interactions. Batch Testing tells you how consistently your agent is showing up for them.

> **Key takeaway:** Pair multi-channel Batch Testing with your post-deployment review process. Any time agent instructions change, treat a full three-channel test run as a required step, not an optional one.

One thing to watch: rating a test case rates the entire case, not individual channel responses. If your email response is great but your voice response is weak, the single rating doesn't capture that nuance. Keep a simple tracking doc or spreadsheet alongside your test results to flag channel-specific issues that the aggregate rating won't surface.

## Who Should Care Most

Service Hub admins and customer experience leads are the primary audience. If you own the Customer Agent configuration, this is now part of your quality assurance workflow.

- Service Hub managers running AI-assisted support across multiple channels who need a faster way to validate agent performance before changes go live.
- RevOps leaders who care about data consistency and want AI interactions to produce clean, comparable records across all channels.
- Marketing ops and CX teams at companies scaling AI-first support who need a repeatable, documented testing process to maintain quality as their knowledge base grows.
- Business owners at Professional or Enterprise tier who've turned on Customer Agent but haven't yet built a formal testing rhythm. This gives you a structured starting point.

If your company handles support volume across all three channels and relies on Customer Agent to deflect tier-one tickets, this update is not optional reading. It's the difference between an agent you've verified and one you've assumed is working.

## George's Take

What excites me about this update isn't the testing feature itself, it's what it signals about how HubSpot thinks about AI responsibility. Shipping an AI agent without a cross-channel testing process is like publishing a website without checking how it looks on mobile. You just wouldn't do it. The fact that HubSpot is building quality assurance directly into the product tells me they're serious about helping humans use these tools with confidence, not just speed. We've seen too many portals where Customer Agent is on but nobody really knows if it's doing a good job. Now there's no excuse.

> Shipping an AI agent without a cross-channel testing process is like publishing a website without checking how it looks on mobile. Now HubSpot's given you no excuse to skip that step.
>
> _George B. Thomas_

If you're configuring Customer Agent and want a second set of eyes on how it's set up across your channels, our team at Sidekick Strategies does exactly this kind of portal audit. We'll help you build a test set that actually reflects how your customers reach you, and we'll make sure your agent is giving consistent, accurate answers before it touches a real conversation. Book a strategy call and let's look at it together.

## Frequently Asked Questions

### What is HubSpot Batch Testing for Customer Agent?

Batch Testing is a quality assurance tool inside HubSpot's Customer Agent that lets you run a predefined set of questions and evaluate how the agent responds. The new multi-channel update lets you run that same test set across Live Chat, Email, and Voice simultaneously so you can compare responses and find inconsistencies before customers encounter them.

### Which HubSpot hubs include Customer Agent Batch Testing?

Batch Testing for Customer Agent is available to Professional or Enterprise subscribers across Content Hub, Data Hub, Marketing Hub, Sales Hub, Service Hub, and Smart CRM. If you have any one of those at the qualifying tier, you have access.

### Can I test Voice and Email separately from Live Chat in Batch Testing?

Yes. The channel dropdown in Batch Testing lets you select any combination of Live Chat, Email, and Voice. You don't have to run all three at once. You can run just Email and Voice, or just Voice and Chat, depending on where you want to focus your quality review.

### How does rating work in multi-channel Batch Testing?

When you rate a response in multi-channel Batch Testing, the rating applies to the entire test case across all selected channels, not to each channel individually. If you need channel-specific quality tracking, maintain a separate log alongside your HubSpot test results to capture per-channel notes.

### Why does my Customer Agent give different answers on email versus chat?

Chat, email, and voice each process context differently. Tone expectations, response length, and how the agent interprets the same question can all vary by channel. That's why HubSpot built multi-channel Batch Testing: to let you catch those gaps in a controlled test environment before a real customer experiences them.

### Is HubSpot Customer Agent Batch Testing available now?

Yes. The multi-channel Batch Testing feature entered public beta on November 17, 2026. Everyone in your HubSpot account is automatically enrolled. No manual opt-in is required, though it's still in beta, so you should validate outputs carefully before treating ratings as final quality benchmarks.

## Key facts

- **Format:** HubSpot Update
- **Published:** November 16, 2026

---
Source: https://www.sidekickstrategies.com/hubspot-updates/hubspot-customer-agent-batch-testing-chat-email-voice
Sidekick Strategies · HubSpot consulting and content operations
