Reps are using AI-powered assistants to draft outreach emails, summarize calls, and prioritize follow-ups—all directly within HubSpot. These tools are helping teams spend more time selling, and less time on admin.

“We used to ask of our reps...‘Hey, can you go look at all the deals and tell us what’s happening?’ And now, with AI, it’s not just about making reps faster—it’s about freeing up time so they can be more effective.”- Will Hughes, HubSpot

Agentic AI goes beyond one-off features. It enables you to build full workflows where AI agents proactively support tasks like lead routing, opportunity updates, and even forecasting nudges—tailored to your unique sales process.

“The difference with agentic AI is you’re not just using AI—you’re designing workflows where an agent takes action based on rules or real-time context. It’s more than automation. It’s delegation.” - Sam Mallikarjunan, Agent.AI

Start with the friction. Look at high-effort, repetitive tasks your reps do daily—like logging notes, chasing down contact info, or qualifying leads. Those are prime candidates for your first agent.

“Start where the friction is. What’s repetitive? What’s annoying? That’s where your first agent lives. For some teams, it’s follow-up emails. For others, it’s qualifying inbound leads. Start small, get a win, and build from there.” - Mathieu Grandmaître, SmartBug Media

Too often, teams try to do too much, too soon. Successful adoption starts with clean data, one strong use case, and internal buy-in—especially from frontline reps.

“I think the biggest mistake is trying to boil the ocean—teams want to go all-in on AI without knowing what they actually want to fix. Pick one problem. Solve it really well. Then look at the next one.” - Will Hughes, HubSpot

AI can surface deal risks, rep activity insights, and even call coaching suggestions—helping managers go from reactive to proactive. It's like having an assistant that flags what needs attention before it’s too late.

“Agents can surface patterns that managers don’t have time to dig for—like which reps are struggling with discovery calls, or where deals consistently stall. That makes coaching proactive, not reactive.” - Sam Mallikarjunan, Agent.AI

It’s a valid concern—and one we’ve seen before during the rise of cloud computing. The key is education, plus choosing the right architecture. For strict data needs, companies can run open-source models like Meta’s LLaMA locally to keep sensitive information inside their firewall. The trend toward edge AI (like Apple’s on-device processing) means more privacy-focused options are coming fast.

“You can use the data you need… and get the benefit from AI without actually losing control of the data your customers trust you with.” - Sam Mallikarjunan, Agent.AI