HubSpot Customer Agent: Review, Pricing & Better Alternative (2026)

The HubSpot customer agent promises to offload common support questions to an AI agent right inside your HubSpot chat, email, WhatsApp and (in beta) phone channels. Launched in 2025 as the Breeze Customer Agent, it now lives in HubSpot Agent Hub and is simply called the customer agent. This guide was re-checked against HubSpot’s own documentation on 12 October 2026, including the changes announced around UNBOUND 2026.

In this guide, we’ll explain what HubSpot’s customer agent is, review how it works, and cover its pros and cons. We’ll break down pricing (including some unexpected costs) and share best practices on training, security, tone of voice, multi-language support, and analytics.

You’ll also learn where the HubSpot customer agent falls short and in what situations a dedicated alternative like Resolve247 can be a better fit for growing teams. The Customer Agent is just one of several ways to run a chatbot in HubSpot; our HubSpot chatbot examples guide shows how it compares with rule-based and third-party builds. Let’s dive in.

What is HubSpot’s Customer Agent?

Breeze was the umbrella name for HubSpot’s AI features when they launched in 2025. On 23 July 2026 HubSpot moved its agents into Agent Hub, which today lists a Customer Agent, a Prospecting Agent, a Data Agent, a Deal Progression agent and an AEO agent, plus custom agents built in Agent Builder (our Agent Hub explainer covers the rename). In simple terms, the HubSpot customer agent is an AI agent that answers your customers from your content. It started in the website chat widget and conversations inbox, and HubSpot’s channel list now covers live chat, email, WhatsApp, Facebook Messenger, forms and, in beta, calling.

The customer agent uses large language models (LLMs). Once enabled it can use your existing knowledge base articles, website pages, and other documents to answer users in a conversational way. Some early adopters have reported that the customer agent was able to significantly reduce handling time for queries. Essentially, it acts as a 24/7 virtual support rep that tries to answer FAQs and routine issues instantly.

A note on names: you will still see “Breeze Customer Agent” in older guides and in search results. HubSpot’s product page and knowledge base now call it the customer agent, and the Breeze name survives only on Breeze Assistant (the in-app assistant) and Breeze Intelligence (data enrichment). This article uses “customer agent” and “HubSpot AI agent” interchangeably.

HubSpot initially rolled out Breeze AI in HubSpot in beta in early 2025, and by June 2025 the customer agent became generally available to all Professional and Enterprise tier customers across HubSpot’s major hubs. HubSpot’s AI features consume HubSpot Credits. The customer agent was previously served by separate credits, but now pulls from the same credit pool as HubSpot’s other AI features.

In summary, HubSpot’s customer agent is an AI-driven chatbot within HubSpot’s platform that can chat with your visitors and customers, answer their questions by drawing on your content, and help route or resolve issues without human intervention. Next, let’s evaluate how well it works in practice and what limitations to be aware of.

HubSpot Customer Agent Review

The customer agent has generated significant interest as a built-in AI support solution. Here’s an in-depth look at its strengths and weaknesses:

Integration with HubSpot Ecosystem

A major advantage of the customer agent is that it’s native to HubSpot. If you’re already using HubSpot’s CRM and Service tools, enabling the AI agent is straightforward – no coding or third-party platforms needed. It plugs into your existing Chatflows (live chat widget) and shared inboxes, so AI-generated answers appear right in the usual conversation threads.

This integration means the AI can also create support tickets or trigger HubSpot workflows when needed, and hand off to your team without friction. The setup is guided by HubSpot’s UI, making it quick to deploy once you have access. Overall, the HubSpot AI agent scores high on convenience for HubSpot users.

Conversational AI, Not Rule-Based

The customer agent is an LLM-based AI, so it is different to HubSpot’s rule-based chatbot (which is still available for situations that warrant a strict rule-based flow). The benefits of LLM-based AI is that it can interpret free-form questions (not pick-and-choose) and can handle multiple languages. It can pull answers from your website, knowledge base or other documents. Do note that the quality of answers will still depend on having strong content.

Actions

The customer agent supports some light action-taking if you configure this with a bit of coding. For instance, if connected appropriately, it could perform tasks like sending a password reset email or checking an order status. The actions available are dependent on the API endpoints in your existing software. If you already have code to complete these functions at the click of a button, it is likely you would be able to use HubSpot’s Actions feature to allow the customer agent to complete these actions for your customers.

Limited Availability (High Plan Requirement)

The biggest drawback of the HubSpot customer agent is who can use it. HubSpot restricts this feature to customers on Professional or Enterprise plans. If you’re on HubSpot’s Free or Starter plans you cannot enable the customer agent at all (likewise many of the chatflow features are disabled). This puts the native HubSpot AI agent out of reach for many small businesses unless they upgrade to a much more expensive HubSpot tier. In short, the customer agent is geared toward customers already on higher plans; it’s not an all-inclusive feature.

Usage Limits and Credit Costs

Even if you have an Enterprise/Pro subscription, the customer agent isn’t an included resource. HubSpot uses a credit system for AI, and each resolved customer agent conversation consumes 50 credits. On a Professional plan you get 3,000 AI credits per month (Enterprise gets 5,000) – equivalent to 60 resolutions/month on Pro (or 100 on Enterprise).

After those 60 or 100 resolutions in a month, you have to purchase more credits to keep the AI responding. Additional HubSpot credits cost $10/month per 1,000 extra credits (or $9 if bought annually), which works out to $0.50 per resolved conversation ($0.45 annually).

This pricing can become expensive for any significant volume of chats. For example, 500 customer chats in a month works out to around 350 resolutions (using HubSpot’s stated 70% average resolution rate), which means $145 extra in credit purchases (350 resolutions at $0.50 each, minus the 60 included on the Pro plan). Heavier usage like 10,000 chats means roughly 6,500 resolutions, which would mean an extra $3,220 in credit fees on top of your base subscription. Many teams will find the included limit insufficient, meaning the true cost of the customer agent can escalate quickly (we’ll break down pricing more in the next section).

Custom Instructions: Now Available, Within Limits

Until mid-2026 this was the customer agent’s biggest gap: beyond a role and a tone you could not tell it how to behave. That changed in June 2026. HubSpot’s Guidelines give you four free-text boxes under Train > Guidelines: Tone, Response Style, Scripted Responses (reusable greetings, apologies and follow-ups) and Guardrails (topics to avoid and what to say instead). A separate free-text Handoff guidelines field lets you name the phrases and situations that must go to a person, such as “cancellation” or “refund”.

So “if the user asks about enterprise pricing, offer to connect them to sales” is now something you can write into a handoff guideline. Two caveats remain. HubSpot’s own documentation warns that AI-generated responses “may not always exactly match the specified formatting or wording”, so treat guidelines as steering rather than rules. And all of it sits behind the Professional and Enterprise gate. Specialist AI integrations such as Resolve247 offer the same kind of instructions on every HubSpot plan, Starter and Free included.

Insights and Analytics: Reports Added in 2026, Pro and Enterprise Only

When using an AI, you need to know how it is performing and what your customers are asking. HubSpot has filled much of this gap during 2026. Its performance reports now show resolutions, deflections and handoffs over time, a Knowledge gaps list of questions the agent could not answer, Coaching opportunities (knowledge gaps, conflicting sources, failed actions and thumbs-down answers) and a source-usage view of which content is pulling its weight. Grouping conversations by subject needs Topics, which is in beta and needs a Service Hub Professional or Enterprise seat on top of the agent itself. The remaining limitation is where it all sits: every one of these views is Professional and Enterprise only.

Quality Depends on Content (Mixed Early Feedback)

The effectiveness of the HubSpot customer agent hinges on the quality of your content and the AI’s training on it. In straightforward FAQ scenarios, it can perform well – providing answers drawn from help articles and even citing the source URL for transparency. However, if your documentation is sparse or the question is nuanced, the AI might stumble. HubSpot has no explicit accuracy guarantee on the AI’s answers; it may occasionally misinterpret things or “hallucinate” an answer not actually supported by your data.

Early user feedback on the customer agent’s answer quality has been mixed. Some users found it “mostly poor”, with reports that “We tried it and it is not worth it” and “they need to do so much better here”, stating that the offering is a “series of compromises”.

That said, others have had great experiences with basic Q&As. Expect to invest time in training and tweaking HubSpot’s AI with quality content and testing it (as we cover in Best Practices) to achieve good results. HubSpot will likely continue to improve the customer agent, but as one feature in a large pool of other features, its responses may still feel rough around the edges.

Customer Agent Auto-Resolution Rate: What to Expect

HubSpot markets the customer agent with bold claims about auto-resolution rates – the percentage of customer conversations resolved by the AI without needing a human agent. Some HubSpot marketing materials suggest the AI can automatically resolve up to 50% of tier-1 support queries, and you may have seen claims of 90% resolution rates circulating online.

In practice, the actual auto-resolution rate you achieve with the HubSpot customer agent depends heavily on several factors. These include the quality and coverage of your knowledge base, the complexity of questions your customers ask, and whether your content is formatted in the short Q&A style that the customer agent performs best with. Teams with comprehensive, well-structured help docs tend to see higher resolution rates. Those with sparse or long-form documentation may struggle.

Real-world user feedback on HubSpot AI agent auto-resolution paints a mixed picture. Some users on the HubSpot Community and Reddit report reasonable rates on straightforward FAQs, while others find the rate disappointing – particularly for nuanced or multi-step queries. Guidelines (covered above) help with edge cases, but HubSpot’s own note that responses may not follow them exactly means some tuning still happens in your content rather than in the instructions.

If auto-resolution rate is a key metric for your team, it is worth comparing the customer agent against specialist AI providers. Resolve247, for example, currently averages an 82% resolution rate (compared to HubSpot’s stated 70%), with custom instructions on every plan and a guarantee behind its accuracy: if the AI ever makes something up, that month is free. Better AI accuracy means more conversations are fully resolved without human intervention.

Onboarding Cost

One often-overlooked factor is that upgrading from a lower plan to enable the customer agent not only includes the cost of the plan itself but requires a mandatory onboarding fee of $1,500 for Professional or $3,500 for Enterprise – a one-time services fee that HubSpot requires for high-tier plans. This means the true cost of “unlocking” HubSpot’s customer agent could include thousands of dollars in subscription and onboarding costs, even before you consider the per-chat credit fees. We’ll detail pricing next, but it’s clear that for budget-conscious teams, the HubSpot customer agent might be cost-prohibitive.

In summary, the HubSpot customer agent is a powerful AI chatbot for HubSpot users with the right plan, offering native integration and strong potential to automate answers to common queries. It’s easy to set up and can deflect a significant portion of repetitive questions when fed good content.

However, the feature comes with notable limitations: it’s restricted to top-tier plans, has limited included usage and per-resolution costs, its guidelines steer rather than guarantee behaviour, and its answer quality is only as good as your knowledge base (with some users reporting underwhelming results).

In the next sections, we’ll break down the pricing in detail, cover setup and best practices, and discuss how an alternative solution might make sense if the customer agent isn’t meeting your needs.

What Changed at UNBOUND 2026

HubSpot’s annual conference ran from 16 to 18 September 2026 under its new name, UNBOUND, and the customer agent picked up a run of additions in the weeks around it. Everything below comes from HubSpot’s own knowledge base, product page and credits rate sheet, checked on 12 October 2026.

• Voice, in public beta. The customer agent can now answer inbound calls through a HubSpot calling channel and hand the caller to a person. The rate sheet lists voice at 50 credits per minute of conversation, which is $0.50 a minute at HubSpot’s $0.010 per credit. A six-minute call therefore costs about $3, against $0.50 for a resolved chat.
• Segments. One agent can serve up to 25 audiences, defined by contact properties, visitor context (such as language or country) or CRM lists, each with its own extra knowledge sources and guidelines.
• Handoff guidelines and a visible handoff status. Free-text rules decide what must go to a person, and Help Desk now shows a “Customer Agent ticket status” of Pending, Awaiting Human Assignment or Handed off (handoff documentation), so your team can see what the agent is still holding.
• Knowledge generated from your tickets (beta). The agent can read resolved, human-answered tickets, group similar questions and draft short answers for you to review (listed as a beta in HubSpot’s setup guide). The separate Knowledge Base Agent, also in beta, drafts full help articles from the same tickets at 200 credits ($2) per article.
• Reporting and coaching. Resolution, deflection and handoff trends, a Knowledge gaps list, Coaching opportunities and source-usage reports, as covered above. Topics, which groups conversations by theme, is in beta and needs a Service Hub Professional or Enterprise seat.
• Billing default. Credits bought since 16 September 2026 default to pay-as-you-go overage, invoiced monthly at $0.010 a credit with an automatic spend limit 50% above your monthly credits, instead of auto-upgrading your subscription. Packs bought earlier keep the auto-upgrade default (see the pricing section below).
• What did not change. The price is still 50 credits ($0.50) per resolved conversation, and the agent is still Professional and Enterprise only. Starter and Free accounts gained nothing from UNBOUND.

For a chat-first small team, the honest reading is this: the customer agent now reaches more channels and gives you more control, but the two things that decide whether it fits (the plan gate and the per-resolution meter) are exactly where they were before UNBOUND.

HubSpot Customer Agent Pricing Explained

One of the most important considerations with the customer agent is pricing – which can be confusing due to HubSpot’s credit system and plan structure. We have a dedicated article covering HubSpot AI agent pricing, otherwise keep reading where we clarify what you’ll pay (directly and indirectly) to use the customer agent.

Plan Requirement and Base Costs

• HubSpot Plan Requirement: First, you must be on a Professional or Enterprise plan in at least one of HubSpot’s Hubs to unlock the customer agent. Most commonly, companies would use it as part of the Service Hub (customer service), but the customer agent can be used with any Hub product. Service Hub Pro starts around $450/month (billed annually) for 5 seats (or $90/seat, billed annually).

  Note that HubSpot displays different pricing depending on whether you are logged in. Logged-in users see plans starting at 5 seats ($450/mo paid annually), while incognito visitors see the plan offered with no minimum seats ($90/mo/seat paid annually). After talking with their sales team, there is no minimum number of seats – you have to ask. This means you can save on your monthly cost by paying for the exact number of seats you need.

• Onboarding Fee: If you are upgrading to a Professional or Enterprise tier for the first time, be aware of HubSpot’s required onboarding fee. This is $1,500 (one-time) for Professional, or $3,500 for Enterprise, which HubSpot charges to provide setup support and training for their product. This fee is required regardless of the AI – it’s part of signing up for a high-tier HubSpot plan.
  It’s an unexpected upfront cost if you’re not already on Pro/Enterprise. There are reports that you can negotiate this down to an Onboarding Foundations course for $600 (one-off). Some have even managed to waive it completely by passing their free Software Certification course. Make sure to push your HubSpot sales rep to let you know all the options available to you.

Credit System and Per-Resolution Costs

• Includes 2 resolutions per day on Pro: HubSpot includes a small number of resolved AI conversations for free, in the form of included HubSpot credits, then they are charged per resolved conversation with HubSpot Credits. Each resolved conversation consumes 50 credits. Pro plans include 60 resolutions per month (3,000 credits per month) and Enterprise plans include 100 resolutions per month (5,000 credits per month). Most support teams will quickly exceed these volumes, so you need extra credits. Note how “resolved” is defined: the agent shared a source or took an action, and the visitor did not ask for a person within 72 hours of their last message. A visitor who reads the answer and leaves counts as resolved, and is billed.
• Extra Credits, a slippery slope: Additional resolved conversations cost $0.50 each on pay-as-you-go (the default for credits bought since 16 September 2026, invoiced monthly in arrears at $0.010 a credit). Committed packs of 1,000 extra credits for $10/month work out at the same $0.50 only if you use every credit; anything unused pushes the real cost per resolution higher.
  • There is a discount if you commit to annual credit bundles (e.g. you can pay $9/1,000 credits on an annual basis instead of $10 month-to-month), which brings the cost to $0.45/resolved conversation.
  • Unused credits expire at the end of each month. So cost per resolution can sky rocket if you have unused capacity.
  • If you use all of your free credits (without having bought any extra credits) then the AI functions that use credits will stop working until your next reset date.
  • If you bought credit packs before 16 September 2026, the default is still auto-upgrade: use them all and HubSpot will auto-add another Credit Pack for the rest of your contract term. Packs bought since then default to pay-as-you-go overage instead, with a spend limit set automatically at 50% above your monthly credits. Either setting can be switched in billing.
  • Downgrades are only allowed at your next renewal by signing a new order form, to take effect at the end of your contractual commitment.
  • With committed packs there is a careful balancing act between buying enough credits to handle months that have peak demand, while accepting that you will need to pay for that unused capacity for the rest of the year.

Real-World Examples

• Example Scenarios with committed credit packs: In real-world use cases, the pricing model can make the extra costs uneconomical if you aren’t aware of what to look out for. On pay-as-you-go you simply pay $0.50 for every resolution over your allowance, invoiced monthly. The committed-pack route (auto-upgrade) is where the surprises live, so here are a few scenarios of its costs (using HubSpot’s stated 70% average resolution rate to translate conversations into chargeable resolutions):
  • If you generally have 8-10 conversations per day (240-300 conversations/month, ~168-210 resolutions at HubSpot’s 70% rate):
    • you might have committed to 200 resolutions for $100/month (10,000 extra credits beyond Pro’s included 3,000).
    • if you have a quiet month and only resolve 160 conversations that month, your cost per resolution increases by 25% to $0.625/resolution.
    • if you have a busy month and resolve 240 conversations that month, HubSpot will auto-increase your ongoing monthly extra credits to $120/month for 240 resolutions/month (and prorate the charge for that month). In the following months, this would increase your costs to $0.55-$0.75/resolution.
  • If you generally have 60-80 conversations per day (1,800-2,400 conversations/month, ~1,260-1,680 resolutions at HubSpot’s 70% rate):
    • you might have committed to 1,500 resolutions for $750/month (75,000 extra credits).
    • if you have a quiet month and only resolve 1,170 conversations that month, your cost per resolution increases by 28% to $0.64/resolution – or an overspend of $165.
    • if you have a busy month and resolve 1,700 conversations that month, HubSpot will auto-increase your ongoing monthly extra credits to $850/month for 1,700 resolutions/month (and prorate the charge for that month). In the following months, this would increase your costs to $0.50-$0.73/resolution.

Watch out: To prevent credit costs spiralling out of control, monitor any month you go over your credit allowance. Contact your HubSpot Sales Rep to remove the extra credits that would have been auto-added, or switch the account to pay-as-you-go overages so a busy month is a one-off charge rather than a permanent upgrade. We have put together an in-depth article to help you reduce your HubSpot credit costs.

Bottom Line: Using the HubSpot customer agent is not cheap. This HubSpot AI agent costs orders of magnitude more than many alternative high-quality AI chatbot services that can reduce your support costs.

If you’re already on Enterprise or Professional plans, the included 60-100 resolutions/month are useful to do some testing with. But any serious usage will incur meaningful extra costs. Companies should project their expected chat volume and budget for the overage, whether it arrives as a pay-as-you-go invoice or an automatic pack upgrade. This might be fine for a large company, but isn’t practical for a small business.

In later sections, we’ll discuss alternatives that can be more cost-effective, or you can see our full article on HubSpot AI chatbot alternatives.

HubSpot AI Agent Setup: Step-by-Step & Best Practices

Ready to give the HubSpot customer agent a try? This section walks you through how to enable and configure this HubSpot AI agent, then outlines a best-practice checklist to get the most out of your new AI helper.

Setting up the basics can be done in 10-15 minutes. If you need to upgrade your plan, the required onboarding will add time on top of this. Tuning HubSpot’s AI for optimal performance is an ongoing process – HubSpot’s case studies refer to the process taking weeks to get quality results. We’re going to help you keep this as short as possible and recommend alternatives that don’t require a lengthy process (hint: AI doesn’t have to be that hard).

How to Set Up the Customer Agent

Setting up the HubSpot customer agent involves creating an agent in your HubSpot portal, choosing content sources to train it on, assigning it to a chatflow, configuring handoff rules, and testing before going live. You’ll need admin access on a Professional or Enterprise account with Agent Hub access.

We’ve written a detailed step-by-step walkthrough in our HubSpot chatbot builder guide, which covers the full setup process including content source selection, channel assignment, and testing.

Once your agent is live, the real work begins. The best practices below will help you get the most out of it.

Best Practices for Customer Agent Success

Enabling the feature is just the start. To truly succeed with HubSpot’s customer agent, keep these best practices in mind:

• Only train on data you want to be shared: When activating the customer agent, you grant it access to certain data (like conversations or files). Be mindful of what content you allow – only include sources that you’re comfortable having the AI use in responses. Do not upload sensitive internal documents that you wouldn’t want shared with customers. If a topic is off-limits, make sure that info isn’t in the training set.
• Maintain Your Knowledge Base: The AI is only as good as your content. So effort put into adjusting and improving the knowledge base will result in better results from the AI.
  • Keep it up-to-date: If you push a new product release or policy change, update those help articles promptly – and re-train or refresh the customer agent’s content if needed. HubSpot’s agent will automatically pick up changes to included sources.
  • Occasional Audits to catch missing data: With HubSpot, it’s wise to periodically review what it’s trained on – HubSpot’s case studies include information going missing and they recommend conducting audits to make sure information hasn’t dropped out of your AI’s knowledge. This shouldn’t happen, and may have been resolved by HubSpot now.
  • Re-write content to suit HubSpot’s AI: It’s recommended that content is amended/added in a format to suit HubSpot’s AI to maximise benefit from the customer agent. Such as creating short Q&A articles for any common questions that you discover the AI couldn’t answer. It has been said that short answers are essential to success with the customer agent.
  • Fill knowledge gaps: The customer agent flags unanswered questions for you, which is a great way to identify knowledge gaps, so that same question can be answered automatically in the future. By continuously grooming your knowledge base, you’ll improve the AI’s accuracy and usefulness.
• Test (with real questions): The questions you might think to use to test a chatbot might be skewed because you know your company well – you might use different wording or phrasing to your customers. A customer might be unclear, they might be frustrated, or they might not provide full context – all of which are real situations that will affect the results from the chatbot, and need to be handled appropriately. To ensure the AI is properly trained, take a variety of your recent real conversations and see how the AI would respond. Make sure to test for the following situations:
  • Common questions: Start with frequently asked questions, to make sure the AI responds exactly how you need it to. As these are frequent questions, it’s worth making sure the AI responds accurately.
  • Unanswerable questions: Ask a question that you know it doesn’t currently have the information for.
  • Frustrated customers: If a customer is unhappy, we want to make sure the AI doesn’t do anything to make the situation worse! For example, sending them round in circles or not providing a way forward.
  • Similar but not quite the same: There might be some loosely related topics, that might have different answers. Test this by asking a specifically unclear question, or intentionally omitting some context to see if the AI appropriately responds without confusing the similar topics.
  • Known tricky situations: If there are delicate or complicated situations that you need to handle, it is worth testing these out to make sure the AI can handle these in a robust way – in line with your business requirements.
  • Try alternative words: A modern AI (an LLM) should handle this fine (as it should understand the context of the words) but check that any common alternative words still capture the topic.
  • Known confusing/sales-y topics: Some of your content might be intentionally written in a positive or non-specific way in order to generate leads/interest. This kind of “implied” language can confuse some AI. Think about any parts of your business where you might be glossing over any missing features or work-arounds and check that the AI responds in a way that is factually accurate, but also matching how you would want to answer the question.
• React to Test results: Based on the findings of the above, the solution with HubSpot’s AI is to amend the knowledge base to catch that situation, and retest.
  • Missing information: If the missing information is a missing piece of knowledge, you can add Short Answers instead of needing to amend webpages (but still make sure that all your information is consistent).
  • Behavioural/business logic: If the issue is behaviour based, add or tighten a Guideline (tone, response style, scripted responses, guardrails) or a Handoff guideline, publish it and retest. HubSpot notes the agent may not follow the wording exactly, so check the result rather than assume it. If a behaviour still will not stick, the fallback within HubSpot is to force a handover to a human for that situation. Specialist providers such as Resolve247 offer the same kind of Custom Instructions on every HubSpot tier, including Starter and Free, which we come back to later in this article.
• Human Oversight: Don’t set and forget the AI. Especially in the early phase, have your support team keep an eye on live chats. They can intervene if the AI goes off track. Also, encourage your team to proactively update content based on what they observe. A best practice is to have a feedback loop: support agents should flag if they see the AI give a wrong or suboptimal answer, so you can correct it. Over time, as confidence in the AI grows, you can let it handle more on its own. There should always be an easy path for customers or agents to reach a human when needed. Ensure the chat widget offers an option like “connect me to an agent”.

By following these best practices you’ll set up HubSpot’s customer agent for success. Some early adopters learned that launching the HubSpot customer agent is not a one-and-done task, but an iterative process. So, if you are already on HubSpot’s higher plans, not cost-sensitive, and have time to shape your knowledge base to suit HubSpot’s AI, then the HubSpot AI agent can provide highly efficient support assistance to help your Customers, and your Team!

Next, let’s address where even a well-tuned customer agent might still fall short of your needs – and how an alternative solution can fill those gaps.

Where the Customer Agent Falls Short (and How Resolve247 Helps)

HubSpot’s customer agent is a convenient solution, but as we’ve seen, it has notable shortfalls. For many growing businesses, these limitations become pain points. Here we recap the main issues with the customer agent and explain how Resolve247 (a specialist in AI that integrates directly into HubSpot) addresses each:

• Limited to Enterprise/Pro Plans: The HubSpot AI agent is not available on Free or Starter HubSpot plans. This excludes smaller businesses from using it unless they upgrade to expensive tiers. On the other hand, Resolve247’s AIChatbot works with any HubSpot plan. You don’t need to change your HubSpot subscription at all. Resolve247 integrates directly into your existing HubSpot chat widget regardless of tier, so startups and small teams can have an AI support agent without enterprise software costs.
• High Cost per Resolution: HubSpot’s pay-per-resolution credits (~$0.50 per resolved conversation) can blow up your budget quickly. In contrast, Resolve247 offers cost-effective flat rate pricing with generous allowances included. For example, their Starter plan at $35/month includes 2,000 AI responses (roughly 600-800 conversations) – which works out to only around $0.04 per conversation. Even higher plans drive the cost per chat down to 2-3 cents. No confusing credits; you just pick a plan with enough responses. This model can save up to 92% of the costs you’d incur with HubSpot’s customer agent.
• Custom Instructions on Every Plan: HubSpot’s Guidelines (June 2026) narrowed the behaviour gap, but only for Professional and Enterprise accounts, and HubSpot itself notes the agent may not follow them exactly. Resolve247 was built to work out of the box and be customisable to suit your specific business needs – it allows you to set Custom Instructions for your chatbot on all plans. You can define business-specific guidelines, special prompts, or fallback behaviours easily.
  For example, you can instruct Resolve247’s AI to tell customers about special promotions, direct Enterprise customers to schedule a call, or gather feedback if someone asks for a refund. These instructions give you fine-grained control to align the AI with your unique processes, whatever HubSpot tier you are on. With HubSpot you get similar controls only once you are on a Professional or Enterprise plan, and HubSpot notes responses may not follow them exactly.
• Insights & Analytics: HubSpot’s reports now cover knowledge gaps, coaching opportunities and source usage, but grouping conversations by subject (Topics) is in beta and needs a Service Hub seat, and all of it is Professional and Enterprise only. Resolve247 analyses your conversations on every plan: grouped by subject and category so you can see what customers ask most, with knowledge gaps flagged so you can fill them.
• AI Quality and Focus: HubSpot’s customer agent is one feature among hundreds in their platform – it works decently for FAQs but has no special accuracy guarantees and might lag behind the latest AI tech. Resolve247 is a specialist in AI support, which results in higher response quality. They use the latest AI models and techniques and provide an anti-hallucination guarantee with money behind it: if the AI ever makes something up, that month is free.
  Resolve247 uses a multi-step process to double-check the AI’s response against your knowledge sources, ensuring it stays factual. HubSpot’s customer agent has no such guarantee. For companies where answer accuracy is paramount, this quality focus is a big advantage. See what our customers say about the difference.
• Support & Guidance: With HubSpot, you generally configure the AI yourself and rely on standard support channels or community if you need help – there’s no specialised assistance for the AI chatbot. Resolve247, by contrast, has AI experts on hand to help get your AI chatbot exactly how you want it and working to its full potential.
  Resolve247 also offers ResponseAssistant – an agent-facing tool that drafts replies in your inbox, giving your team a head start on every conversation. It’s built to be completely self-serve, so you can get going quickly. But if you do need help, their team can assist via screen-share or email for one-on-one guidance. For small teams, having provider support like this can be invaluable – and there’s no hefty onboarding fee. Their help is included for free, rather than the $1,500 or more HubSpot charges for onboarding.

In essence, Resolve247 was created as a high-quality specialist AI solution with a quick setup process to get great results with no effort, and without changing your existing systems. It fixes the issues that many HubSpot users encounter with the customer agent: AI chatbot on any plan, at a fraction of the cost, with custom instructions on every plan and dedicated support.

Importantly, it plugs into your existing HubSpot chat widget – so your end-users see no difference in the chat interface, and all conversation logs still reside in HubSpot. The difference is all under the hood: a more flexible and cost-effective “brain” powering the chatbot.

To decide which way to go, consider your specific situation. Let’s talk about when using HubSpot’s included customer agent makes sense, and when you might be better off opting for Resolve247’s solution.

When is the Customer Agent a Good Fit?

Despite its limitations, HubSpot’s customer agent can be a suitable choice in certain scenarios – particularly if you’re already an ideal HubSpot customer for it. Here are the criteria where using it would make sense:

1. You’re on HubSpot Enterprise or Pro: If your company is already paying for a HubSpot Professional or Enterprise subscription, then you have the AI agent available (without this, it’s not worth upgrading for the sake of the customer agent).
2. You have Low Volume: The customer agent works best for basic Q&A support at low volume. If your website only gets a handful of support questions per day you might be able to make use of the limited free credits HubSpot provide (Beyond which, costs of credits become uneconomical).
3. You have Simple Queries: If the questions you receive are mostly straightforward (like “How do I reset my password?”), the HubSpot AI can handle them well to save your team a bit of time.
4. You have Basic Needs, Little Customisation Required: If your goal is simply to provide a FAQ bot and you don’t have any complex logic or niche requirements, the customer agent could do the job, and Guidelines now cover tone, style and the topics it should avoid.
5. Budget Isn’t a Concern: In some large companies, the high cost of HubSpot credits might be within the allowed budget, and the potential savings might not be worth combating your internal policies to use a specialist third party integration. In the case where bureaucratic-convenience outweighs cost saving and quality improvement, the customer agent will be the appropriate choice.

If you meet all the above criteria, then HubSpot’s customer agent would be a good fit for you.

However – and this is a big however – growing teams might find the customer agent hindering their support quality and budget. This tends to happen as your usage grows or your needs become more advanced. That’s when looking at alternatives like Resolve247 becomes worthwhile. In the next section, we’ll outline when you should strongly consider Resolve247 as a more suitable solution.

When is Resolve247 More Suitable?

For many organisations, a third-party integrated solution like Resolve247’s AI Chatbot will be more aligned with their needs. Consider opting for Resolve247 in these situations:

• You’re on HubSpot Free/Starter: This is a no-brainer – if you aren’t on an eligible plan for the customer agent, Resolve247 gives you an AI chatbot without forcing a costly HubSpot upgrade. Even if you are on Pro/Enterprise, maybe you’d prefer not to pay HubSpot’s credit fees. Resolve247 works on any plan via a simple integration, so it’s the clear choice for those on lower tiers or those avoiding HubSpot’s upsell.
• High Chat Volume or Rapid Growth: If you anticipate a large number of AI conversations (hundreds or thousands per month), Resolve247’s pricing model will save you a lot of money. For example, handling 500 chats (estimated 350 resolutions) via the HubSpot customer agent would cost around $145 in credit fees (350 resolutions at $0.50 each, minus 60 included on the Professional plan); with Resolve247, the $35 plan would cover that (and would also resolve a higher percentage, current average resolution rate is 82%!). For a growing support team, these savings compound. It’s more predictable pricing too – you’re not getting a surprise bill because usage spiked one month.
• Need for Custom AI Behaviour: When your chatbot needs to do more than regurgitate FAQs (like follow certain business rules, promote specific offers, or handle different customers differently) then Resolve247 is more suitable for you. HubSpot’s Guidelines steer tone and topics on Professional and Enterprise plans, and HubSpot itself says responses may not follow them exactly. Resolve247’s custom instructions are built for exactly this kind of business logic, work on every HubSpot plan, and sit behind a guarantee that any month the AI makes something up is free.
  This is ideal if you want the bot to treat VIP customers differently, handle certain subjects in specific ways, or capture an email on certain queries. If you’re thinking “I just want the bot to do X when Y happens,” that’s a sign to test whether HubSpot’s Guidelines hold it reliably. With Resolve247, that is exactly what custom instructions are for.
• Quality and Accuracy are Paramount: Perhaps you operate in an industry where incorrect answers can be harmful (finance, healthcare, etc.), or you just have a low tolerance for the AI messing up. Resolve247’s emphasis on accuracy (with a guarantee that any month the AI makes something up is free) makes it a safer bet when you need trustworthy answers. Also, because Resolve247 is a specialist AI provider, our sole priority is high-quality AI responses, staying ahead of other providers and delivering more nuanced and context-aware responses. If your testing of the HubSpot AI agent showed too many “off” answers, an alternative like Resolve247 could likely perform better with the same data.
• Desire for Hands-On Support: Smaller companies might not have the time or expertise to optimise an AI bot. If you want a partner to help ensure the chatbot succeeds, Resolve247 provides that white-glove support if you want it. This can make a big difference in achieving value from the AI – it’s like having an AI consultant on call. If HubSpot’s self-serve approach feels daunting or if you simply don’t want to be alone in troubleshooting the bot, a provider like Resolve247 is more suitable.
• Integration with Private Knowledge Sources: If your crucial support information isn’t all on public webpages (such as if you’ve got your internal policies in Notion) then consider Resolve247. It can connect directly to your private Notion pages, whereas with the customer agent you’d have to copy that info into another source. For companies with extensive internal knowledge bases, this capability is a huge time-saver (and keeps sensitive info internal).

To sum up, Resolve247 is more suitable when you outgrow the customer agent’s confines or you’re not in its target demographic to begin with. It brings the power of an AI support agent to any HubSpot user, with more flexibility and usually at a much lower total cost. Many teams start with the customer agent and then realise they need more – more conversations, more control, more support – which is exactly where Resolve247 shines.

Below is a quick comparison highlighting the key differences between HubSpot’s customer agent and Resolve247’s AI chatbot solution:

Aspect	HubSpot Customer Agent	Resolve247 AI Chatbot Integration
Plan Availability	Only on Professional/Enterprise tiers of any Hub (not available on Free or Starter)	Works on any HubSpot plan (Free, Starter, Pro, Enterprise) – no HubSpot upgrade required.
Cost per Conversation	$0.45-$0.50 per resolved conversation. Plus requires high-tier HubSpot subscription and $1.5k onboarding fee.	$0.02-$0.05 per conversation (e.g. $35 for ~800 conversations). Much larger allowances, scales affordably.
Custom Instructions	Supported since June 2026 via Guidelines (tone, response style, scripted responses, guardrails) and handoff guidelines, on Pro/Enterprise only. HubSpot notes responses may not follow them exactly.	Supported. You can set custom AI instructions and business rules on all plans, tailoring the bot’s behaviour.
AI Response Quality	Uses AI as a generalist platform feature. No explicit accuracy guarantee; quality depends on your content and HubSpot’s default model tuning. Early user feedback on answer quality has been mixed.	Specialised AI provider focused on support quality. Uses latest models with an anti-hallucination guarantee: if the AI ever makes something up, that month is free. Continuously optimised for customer service use-cases.
Support & Onboarding	Standard HubSpot support (help articles, community). No AI-specific coaching; it’s self-serve configuration. Also requires a $1,500 (Pro) or $3,500 (Enterprise) onboarding fee (for general HubSpot onboarding).	Personalised support from AI Specialists included. Resolve247’s team can help with setup, and troubleshooting (even one-on-one screen share). Can be entirely self-serve or with free expert help as needed.
Knowledge Sources	Can index any public website content, HubSpot knowledge base articles, and uploaded documents. Cannot natively connect to private data sources like Notion without making them public or manual upload.	Can index public websites and connect to certain private sources (e.g. Notion) securely. More flexible in training on your existing knowledge repositories without exposing them.
Setup Effort	Fairly quick to enable if you have HubSpot Pro/Ent – integrated in UI using the Chatflow builder, no coding. But limited customisation in setup.	Quick integration (Click to connect to HubSpot). No code needed, no Chatflow changes needed. Works out-of-the-box with the HubSpot chat widget.

In summary, the customer agent is an easy add-on for those already in HubSpot’s ecosystem at the highest tiers, handling basic support queries with minimal setup. But for many businesses, Resolve247’s solution will be more cost-effective, accessible, and powerful – especially if you’re not on Enterprise, have higher volumes, or need more control over your AI assistant.

The right choice depends on your situation: if the customer agent fits your current needs and you have it available, it’s worth trying first. If you hit its limits or can’t use it, know that alternatives like Resolve247 are readily available to supercharge your HubSpot chat with AI on your terms. For a detailed side-by-side breakdown, see our HubSpot AI vs Resolve247 comparison.

HubSpot Customer Agent FAQ

Q: Is the HubSpot customer agent available on all HubSpot plans?

A: No – the customer agent (formerly the Breeze Customer Agent) is only available on Professional and Enterprise tiers of HubSpot. This applies to Service Hub Pro/Enterprise (and as of 2025, also to Pro/Ent of other Hubs). It is not included on Free, Starter, or Basic plans. If you’re on a lower plan, you would have to upgrade to at least Professional to use the native AI chatbot (Alternatively, you can use a third-party integration like Resolve247 to get an AI chatbot without upgrading your HubSpot plan).

Q: How much does the HubSpot customer agent cost?

A: The feature itself doesn’t have a fixed price tag – it’s included if you have the required HubSpot plan. However, usage is metered by credits, where 50 credits equals 1 resolved conversation. You get 60 resolutions per month on Pro (100 on Enterprise). Beyond that, each additional resolved conversation (50 credits) costs $0.50 in credits. So, if you have a lot of conversations, costs can add up quickly. Remember you’re also paying for the HubSpot Pro/Enterprise subscription itself (hundreds or thousands per month) and a one-time onboarding fee ($1,500 for Professional, $3,500 for Enterprise) to upgrade in the first place.

Q: How do I set up the HubSpot customer agent?

A: You need to be on a qualifying plan first. Then, in HubSpot, go to Service > Customer Agent and click Create Agent. You’ll name your bot, choose a personality and languages, set guidelines, and then provide content sources (knowledge base articles, webpages, PDFs) for it to train on. After creating it, assign the agent to your chatflow or inbox (so it’s live on your site). Configure any handoff rules (what the bot should do if it can’t answer). Finally, test the bot using the preview tool and tweak content as needed. Once satisfied, turn it on in your website chat. For a full walkthrough, see our HubSpot chatbot builder guide.

Q: Can I customise the HubSpot customer agent’s behaviour?

A: Yes, within limits. Since June 2026 HubSpot’s Guidelines let you write free-text instructions for tone, response style, scripted responses and guardrails (topics to avoid), plus handoff guidelines naming the situations that must go to a person. HubSpot cautions that responses may not match the wording exactly, and the feature is Professional and Enterprise only. If you want instructions on any HubSpot plan, with a money-backed accuracy guarantee behind the answers, a third-party integration like Resolve247 offers custom instructions on every tier. See our roundup of the top AI chatbots for HubSpot for more options.

Q: What auto-resolution rate can I expect from the HubSpot customer agent?

A: The auto-resolution rate – the percentage of conversations resolved without human intervention – varies widely depending on your knowledge base quality, question complexity, and content format. HubSpot states a 70% average resolution rate, but real-world results are mixed. Teams with comprehensive, well-structured Q&A content tend to see higher resolution rates. If auto-resolution is a priority, specialist AI providers like Resolve247 often achieve better rates (Resolve247 currently averages 82%, compared to HubSpot’s 70%) backed by a guarantee that any month the AI makes something up is free.

Q: I’m using Fin on HubSpot. Is there a flat-rate alternative inside HubSpot?

A: Yes. Intercom’s Fin has run on HubSpot since May 2026: it answers HubSpot email tickets and Intercom Messenger conversations, hands off into your HubSpot agent queue with the history attached, and is billed at $0.99 per outcome (a resolution or a successful procedure handoff) on top of your HubSpot seats, according to Intercom’s help centre. HubSpot’s own customer agent is $0.50 a resolution but needs a Professional or Enterprise plan. Resolve247 is the flat-rate option: it answers inside the HubSpot chat widget on any HubSpot plan, including Starter and Free, for a fixed monthly price with no per-resolution meter, and a Contact a Human button sends the customer to your team. Our HubSpot vs Intercom comparison goes deeper on the two platforms.

Q: If I’m on HubSpot Starter/Free, can I still add an AI chatbot to my site?

A: You can’t use HubSpot’s native customer agent on those plans (since it’s locked to Pro+), but you can integrate a third-party AI chatbot into your HubSpot chat widget even on Free/Starter. Resolve247 is one such solution – it connects to your HubSpot chat widget and powers the same chat interface with AI responses.

From a visitor’s perspective, it looks and feels like any other HubSpot chat, but the intelligence behind it is from Resolve247’s AI. All the chats still log in your HubSpot inbox as normal. See our guide on how to add AI to your HubSpot chat widget for a full walkthrough.

Final Thoughts

The HubSpot customer agent is a solid option if you’re already on a Professional or Enterprise plan and need basic AI support at low volume. But for many growing businesses, the limitations quickly add up.

Here are the key takeaways from this guide:

1. Plan access is limited. The customer agent requires HubSpot Professional or Enterprise – it’s not available on Free or Starter plans.
2. Costs escalate fast. At $0.50 per resolved conversation (plus credits that expire monthly and, on older contracts, auto-upgrade), the pricing model is unpredictable and expensive at scale.
3. Customisation arrived, but only on the top tiers. Guidelines and handoff guidelines (June 2026) let you steer the HubSpot AI agent on Professional and Enterprise plans only, and HubSpot notes it may not follow them exactly.
4. Quality varies. Early user feedback is mixed, and there’s no accuracy guarantee. Success depends heavily on the quality and format of your knowledge base.
5. Alternatives exist that solve these problems. Resolve247 offers AI chatbot support on any HubSpot plan, at a fraction of the cost, with custom instructions and conversation insights on every plan, and dedicated support.

If you’re evaluating whether the HubSpot customer agent is right for your team, or if you’ve tried it and hit its limits, give Resolve247 a try. There’s a 30-day free trial, no credit card required, and it works with your existing HubSpot chat widget – setup takes minutes, not weeks.