HubSpot has launched Agent Hub and Agent Builder to help go-to-market teams manage AI agents from one CRM workspace.<\/p>\n
The public beta targets AI agent sprawl, a growing challenge for teams that have tested multiple automation tools without clear ownership, customer context, or performance visibility.<\/p>\n
Agent Hub gives teams one place to view, activate, and manage agents across marketing, sales, and service.<\/p>\n
HubSpot positions the product as a control layer for businesses that already use AI, but struggle when different agents work from disconnected customer data. That matters because scattered automation can create uneven customer journeys.<\/p>\n
The company says Agent Hub lets users track live agent status and performance. Teams can also activate new agents with one click and organize outcomes around goals such as building demand or winning deals. For Duncan Lennox, Chief Product and Technology Officer at HubSpot, the issue is no longer whether one agent can complete one task:<\/p>\n
\u201cThe problem isn\u2019t managing a single agent in isolation. It\u2019s that once you have multiple agents, they become fragmented, all working from different pictures of the customer, or even worse, no picture at all.\u201d<\/p>\n
That shared context sits at the heart of HubSpot\u2019s pitch. The company wants AI agents to work from CRM data instead of separate workflows that need manual setup or repeated field mapping.<\/p>\n
Why AI Agent Governance Now Matters<\/p>\n
HubSpot\u2019s launch lands as analysts and CX leaders warn that AI agents need clearer management models.<\/p>\n
In customer service, the stakes can rise quickly. A human agent error may affect a handful of interactions, while a faulty AI agent can reach thousands of customers before leaders spot the issue.<\/p>\n
Speaking to CX Today<\/a>, Kathy Ross, VP Analyst at Gartner, warned that businesses should manage AI agents as technology rather than human talent:<\/p>\n \u201cWe have to remember that AI agents are tools. They\u2019re very powerful tools, but they\u2019re not employees, they\u2019re not teammates, and they have to be managed like technology.\u201d<\/p>\n That warning fits the broader market shift HubSpot is trying to address. Once AI agents spread across marketing, sales, and service, leaders need a way to understand what each agent does and how it affects the customer.<\/p>\n Ross also pointed to the scale of potential failure. If an AI agent fails while serving hundreds or thousands of customers, the impact can hit operations, trust, and brand reputation at once.<\/p>\n For CX leaders, that makes management visibility more than an IT concern. It becomes part of customer journey design.<\/p>\n Agent Builder Brings No-Code AI Into the CRM<\/p>\n Agent Builder lets teams create custom AI agents using natural language instructions.<\/p>\n HubSpot says the tool can use deal history, contact records, call transcripts, and buying signals already inside its platform. Breeze Assistant then turns plain-language instructions into automated actions.<\/p>\n That could help teams automate routine work without pulling developers into every workflow. It also gives CX and revenue leaders a way to test AI agents closer to real customer data.<\/p>\n Agents can run on schedules, contact updates, webhooks, or third-party integrations. That gives teams flexibility while keeping agent logic inside HubSpot\u2019s customer platform.<\/p>\n In a separate CX Today conversation<\/a>, Rebecca Wettemann, Principal at Valoir, argued that leaders need real-time oversight before they push AI agents into customer-facing work:<\/p>\n \u201cLet\u2019s put the management and monitoring tools in place just like we would manage quality and monitor quality for human agents so we can understand in real time what these agents are doing and how do we expose them in areas of lower risk.\u201d<\/p>\n That idea aligns with HubSpot\u2019s emphasis on performance visibility. If teams can see what agents are doing and measure outcomes, AI becomes easier to scale with discipline.<\/p>\n What This Means for CX Teams<\/p>\n The launch shows how AI agents are moving from isolated experiments into managed customer workflows.<\/p>\n For CX leaders, the use cases are practical. A service team may want an agent to summarize customer calls. A sales team may want one to flag buying signals. A marketing team may want one to trigger follow-up based on campaign engagement.<\/p>\n Those use cases lose value when each agent works from a different customer record. They also create risk when leaders cannot see which agent changed what.<\/p>\n HubSpot\u2019s customer example from Ignite Reading shows the operational upside. The organization used Agent Builder to parse school district academic calendars, reducing a 15-20 minute task to seconds.<\/p>\n HubSpot says the workflow saves Ignite Reading more than 350 hours each year. That example highlights the type of back-office automation that can quietly improve frontline customer experience.<\/p>\n The bigger question now concerns adoption. Teams may like the promise of natural-language AI agents, but they still need clear ownership, reporting, and guardrails.<\/p>\n Suggested internal link: link \u201cAI agents in CX\u201d to a relevant CX Today AI agents explainer, AI automation hub, or HubSpot brand page.<\/p>\n AI Agent Sprawl Is Becoming a CRM Problem<\/p>\n HubSpot\u2019s launch suggests CRM vendors want to become the home for AI agent control.<\/p>\n That makes sense. CRM systems already hold customer records, interaction history, deal context, and service signals. AI agents need that context if businesses expect them to make useful decisions.<\/p>\n But technology alone will not solve AI agent sprawl. CX leaders will still need to decide who owns agent performance, how issues escalate, and which workflows are safe to automate.<\/p>\n The future of AI in CX will depend less on how many agents a business can launch and more on whether those agents understand the customer. HubSpot is betting that shared CRM context can turn scattered automation into measurable customer impact.<\/p>\n Join the conversation: Join our LinkedIn community (40,000+ members): https:\/\/www.linkedin.com\/groups\/1951190\/<\/a>
Get the weekly rundown: Subscribe to our newsletter: http:\/\/cxtoday.com\/sign-up<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"HubSpot has launched Agent Hub and Agent Builder to help go-to-market teams manage AI agents from one CRM…\n","protected":false},"author":2,"featured_media":120527,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[6],"tags":[179,509,510,405,7537,513,644],"class_list":["post-120526","post","type-post","status-publish","format-standard","has-post-thumbnail","category-agentic-ai","tag-agentic-ai","tag-agentic-ai-in-customer-service","tag-agentic-ai-software","tag-ai-agents","tag-artificial-intelligence-agents","tag-autonomous-agents","tag-crm"],"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/120526","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/comments?post=120526"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/120526\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media\/120527"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media?parent=120526"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/categories?post=120526"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/tags?post=120526"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}