This one gets asked every time, and it deserves a real answer. HubSpot's third-party AI providers are contractually prohibited from using your data for model training. Full stop. For HubSpot's own models, there's an AI Model Training setting that's on by default. Super Admins can turn it off under Settings → Account Management → AI → Access. Opting out doesn't disable any AI features — it just means your data isn't used for model improvement. If your account has Sensitive Data enabled, it's automatically opted out of AI model training by default. For HIPAA-specific requirements, we'd encourage you to review your applicable agreement terms and HubSpot's HIPAA documentation directly rather than relying on any general summary — the details matter and the right answer depends on your specific setup.

Yes. Customer Agent supports uploaded files including CSV and XLS, and HubSpot specifically notes those formats are optimized for structured data like product catalogs and pricing information. The key is keeping source files clean, labeled, and current — the agent works from what you give it, so the quality of your inputs drives the quality of its answers.

In some cases, HubSpot can recognize returning visitors based on existing browser and cookie context. That said, this isn't guaranteed in every scenario — for workflows where identity matters, the agent may still need confirming information like an email address. Think of it as smart when it can be, careful when it needs to be.

No — and this is worth saying plainly. It's built to handle repetitive, high-volume conversations and route more complex issues to your team. Think of it as the first line that never sleeps, not a replacement for the people who do the real work.

Like any AI system, Customer Agent is only as reliable as the content it's working from. Start by giving it access to approved support documentation, FAQs, and product information, then monitor conversations and refine that content over time. Human handoff rules add an important safety net for more complex or sensitive questions.

This came up live and it's a fair question given how HubSpot's other chat tools work. No — Prospecting Agent is entirely internal. It works behind the scenes inside HubSpot: identifying buying signals, sourcing contacts, mapping buying committees, and drafting personalized outreach for your reps. Visitors never see it. Your sales team does.

Probably, yes — but it's a quick fix. The new experience (buying signals, contact sourcing, account discovery) is currently in Public Beta, and a Super Admin needs to opt the account in. Look for "Prospecting Agent Gets Smarter: Buying Signals and Contact Sourcing" in your HubSpot beta portal. If you're not enrolled, you're still on the older version.

In most cases, results improve when your plays, target audiences, personas, signals, and CRM context are more clearly defined. Prospecting Agent performs best when the inputs behind targeting and outreach are specific and current — if outputs feel generic, that's usually the first place to look.

It depends on how much you're configuring. A lightweight deployment can come together quickly, but a more complete setup — one that includes call scripting, handoff rules, routing logic, and HubSpot workflow mapping — naturally takes longer. Our team can walk you through what a realistic build looks like for your use case. And yes, that voice in the demo was the real agent. Not dubbed, not edited.

In the HubSpot integration model, Aloware works from CRM records and workflow-driven automation. In practice, that means the contact needs a usable phone number in HubSpot, and the trigger logic is defined through HubSpot workflows plus Aloware configuration. Call activity, qualification answers, outcomes, and notes sync back into HubSpot automatically — no manual data entry.

Most teams use AI voice agents primarily for qualification, routing, follow-up, and appointment-setting, with reps stepping in for deeper sales conversations and closing. The goal is to get the right conversations started faster so your team can focus their time where it actually moves deals.

It's designed to complement existing outreach motions — handling early-stage calls, SMS, and workflow-based follow-up while syncing all activity back into HubSpot so reps have full context when they pick up the thread. No gap between what the AI did and what the rep sees.

Good starting points are HubSpot's documentation on AI in Workflows and the Data Agent actions, plus HubSpot Academy content on AI-powered prospecting and AEO. For hands-on learning, the workflow builder itself is often the most practical place to start — you can use Breeze-assisted trigger and action generation and test AI actions directly in your workflow setup.