Frequently Asked Questions

Hamming is a QA and testing platform for AI voice agents. It helps teams test their voice agents pre-launch, monitor them in production, and continuously improve performance, reliability, and compliance.

Hamming is designed for anyone building and scaling voice agents. Whether you're a hardcore engineer or a UX obsessed product manager, Hamming gives you the right tools to make your voice agents production-ready.

You can use Hamming to measure AI voice agent performance, simulate real-world conversations, stress-test agents with edge-case failures, monitor latency and compliance 24/7 in production.

Both. Hamming enables pre-production testing and 24/7 production monitoring, helping you catch issues early and keep agents reliable in live environments.

No. Hamming is designed so anyone can test and monitor voice agents.

From pre-production testing to post production monitoring, Hamming offers tools to optimize your voice agents throughout its lifecycle.

Hamming is used in finance, healthcare, insurance, retail, travel, and customer support industries.

Yes.

Yes. Hamming supports multilingual testing and monitoring, including regional language variations like Castilian Spanish and Mexican Spanish, and more.

Setup takes just minutes. Depending on whether you’re dialing SIP numbers or connecting via LiveKit/Pipecat for WebRTC, you can run your first test call in 15 minutes.

Yes. You can add users to your workspace once you’re onboarded, with role-based access to manage permissions.

You can sign up here

You can book a demo here

Yes. You can add and test multiple agents from your workspace.

Yes. Hamming supports testing in staging environments as well as 24/7 monitoring in production.

Once you’re onboarded, you can add users to your workspace and assign them roles with the right access permissions.

Yes. Our documentation is available to customers and can be accessed through your Hamming workspace after onboarding.

Yes. We provide onboarding support, including a dedicated Slack channel for direct access to our team.

Voice agent testing is the process of simulating conversations to evaluate an AI voice agent’s performance and reliability.

Full information about your usage is shared during onboarding.

Yes, you can test your voice agent error boundaries to ensure that your fallback prompts are working as expected.

Yes, Hamming supports multi-turn conversations and you can test your voice agents with multi-turn agents.

Yes, you can red-team your voice agent with Hamming. We recently red-teamed Ani, Grok’s AI companion.

Yes, you can test and measure customer and voice agent interruptions.

The number of test cases you can run per month depends on your plan. All plan information and usage is shared during onboarding.

Yes, recordings are stored securely and can be accessed from your dashboard.

Yes, you can define custom scenarios. You can customize the test's scenario, conditions and user profile to reflect real life conditions.

Yes, you can test authentication flows and monitor whether your voice agent is compliant.

Yes, you can replay calls from your dashboard.

Yes. Hamming was designed to stress test voice agents.

Yes, you can export test results via PDF, which are emailed to you.

Yes, you can run scheduled tests from your dashboard.

Yes, Hamming supports voice agent regression testing.

Voice agent monitoring is the real-time tracking of your voice agent’s performance and reliability during live calls.

Testing happens before voice agent deployment, while monitoring tracks real-world calls after launch.

Yes, Hamming monitors live production calls.

Yes, Hamming sends real-time alerts to notify your team when thresholds are breached.

Yes, Hamming detects and monitors authentication failures.

Yes, you can customize your dashboards and alerts to your needs.

Data is streamed in real time, and monitoring results are available immediately after the call ends.

Yes. Hamming can simulate background noise issues to indicate how your agents may perform in production environments.

Yes, Hamming can detect when agents go off-script and how well agents stick to the script (prompt adherence).

Yes, you can monitor multi-language agents with Hamming. For instance if a voice agent starts the conversation in English and switches to Spanish, you can monitor and test the voice agent's performance across both languages.

Hamming’s monitoring metrics are benchmarked against industry standards, however, you can also set specific benchmarks from your dashboard.

Voice AI risks include exposing personally identifiable information (PII), PCI DSS violations and HIPAA violations.

You can simulate compliance edge cases pre-production and monitor agents in production to detect compliance violations.

Yes, Hamming tests agents against PCI DSS compliance use cases to identify risks and help ensure that your voice agents handle sensitive payment information securely.

Yes. Hamming can test agents against HIPAA-related use cases to identify risks and verify that protected health information (PHI) is handled securely.

Yes. Hamming monitors and detects when agents risk exposing personally identifiable information (PII).

Yes, you can set specific guardrails based on compliance rules tailored to your industry and regulatory requirements.

We are in the process of SOC 2 compliance.

Yes. Hamming can simulate red-team compliance attacks to safely uncover vulnerabilities.

Latency is the time between a user's command and the system's response.

Latency affects the voice user experience. Poor latency and silence gaps can make customers feel like the agent has abandoned the call.

TTFW is the time until the agent begins responding after a user finishes speaking.

Industry benchmarks suggest 1.5 seconds.

Yes.

Yes, you can also set alerts to be notified when latency spikes occur.

Yes, poor latency can lead to call drop-off and lost conversions.

Yes, via PDF, which are emailed to you.

Yes, Hamming monitors live latency in production.

Error handling is how an agent detects, responds to, and recovers from failures during a conversation.

Without proper error handling, it affects the voice user experience, leading to frustration and call abandonment.

Common issues include ASR misrecognitions, API timeouts and context loss.

By politely asking for clarification, rephrasing, or escalating when needed.

By using fallback responses like “I’m having trouble accessing that right now.”

By summarizing the last known state of the conversation and, if needed, politely restarting the flow.

Yes, you can test for fallback prompts.

Escalation paths are defined rules that hand off conversations to human agents.

Yes and you can test when and how handovers are triggered.

Yes.

Hamming is platform agnostic. We integrate via telephony SIP and outbound for numbers and via WebRTC as well (LiveKit, Daily).

Yes.

Yes, alerts can be pushed to Slack.

Yes you can do everything you do in our dashboard through the API, including scheduling test runs, fetch test results and more.

Yes.

Yes, you can setup webhooks to get notifications.

Hamming provides a voice agent analytics dashboard designed to measure performance and reliability across your voice agents.

Yes, you can create custom views, metrics, and filters.

Yes, dashboards can be shared with role-based access controls.

Yes.

Yes, you can compare performance across all of your voice agents.

Yes, you can view performance trends over days, weeks, or months.

Yes, you can export via PDF, which will be emailed to you.

Yes, dashboards show p50, p90, and p99 latency breakdowns.

Yes, you can see your call volume analytics.

Hamming offers tailored pricing based on your team’s needs and usage. Detailed pricing information is shared during your demo call, so we can recommend the best plan for you.

We don’t offer a free trial in the traditional sense, but we do provide ways for new teams to experience Hamming and see how it works before committing to a plan.

Plans are primarily based on test volume and usage, not per seat.

Yes, we offer enterprise pricing. Please get in touch for more information.

We offer email and live chat support for all customers, a dedicated Slack channel during onboarding, and enterprise-level SLAs for teams that need 24/7 coverage.

We provide detailed evidence for why the test case is failing - prompt, knowledge base, or other issues. Given we have broken a lot of agents and helped teams fix them, feel free to reach out to us - we're happy to work with you to fix them.

Use the “Forgot Password” option on the login page.

Yes. Our team will work hard to accommodate your request, we prioritize features based on customer needs and roadmap alignment.

Response times depend on your SLA, ranging from 10 minutes to 2 hours.

We can help you with onboarding support for new members to your workspace.

Yes, we’ll provide support for integrations.

Custom integrations can be requested and scoped. Please reach out to us for more information.