At last year’s INBOUND conference, HubSpot introduced a wave of new AI tools and agents. While the announcements were exciting, they also created a common challenge for many teams: Where do we actually start?
Like many HubSpot users, we left the conference inspired, but also aware that adopting AI successfully requires focus. Rather than trying to implement everything at once, we decided to start with a specific problem that had long been inefficient for our team:
Following up with Marketing Qualified Leads (MQLs) to convert them into Sales Qualified Leads (SQLs).
What we found is that the new HubSpot Prospecting Agent solved a workflow that had historically been manual, inconsistent, and largely ineffective.
Early Results (in under two weeks):
Fewer than 10 emails sent, 1 meeting booked, and 1 MQL successfully converted into an SQL.
Before using Prospecting Agent, our MQL follow-up process was fully manual and depended on individual review inside HubSpot where we would:
1. Review our list of MQLs
2. Manually check the activity feed for each contact
3. Decide whether outreach made sense
Either:
Enroll them in a pre-built email sequence, or
Draft a one-off email tailored to their activity
In theory, this approach allowed for thoughtful, contextual outreach. In reality, it had a few major problems:
The process was time-consuming
Follow-up happened on an inconsistent cadence
It required constant manual decision-making
And historically, conversion rates were low
It was critical to our process but difficult to execute consistently as MQL volume increased.
Rather than continuing to rely on a manual process, we implemented HubSpot’s Prospecting Agent to handle MQL follow-up within our pipeline. Our goal was to improve MQL-to-SQL conversion while reducing the time spent manually reviewing leads. First, we configured Prospecting Agent to focus specifically on MQL follow-up. Then we updated our marketing automation to support the process. After contacts completed our MQL nurturing workflows, they became eligible for outreach from the Prospecting Agent.
From there, the process looks like this:
1. Leads are nurtured through marketing automation
2. When they complete the nurture stage, they are automatically surfaced to the Prospecting Agent
3. The agent drafts personalized outreach emails based on the contact’s activity
4. Our team reviews and sends them
This allowed us to move from manual lead review to AI-assisted prospecting at scale.
We kept the initial test small. In less than two weeks:
Fewer than 10 emails were sent
1 meeting was booked
Resulting in 1 MQL → SQL conversion
HubSpot reports that companies using their AI prospecting agent see up to 2× higher response rates, engage twice as many leads, and reduce research and personalization time by as much as 95%. Source: HubSpot's AI Prospecting Agent
Previously, reviewing MQLs required digging through activity timelines and deciding what message to send.
Prospecting Agent now surfaces leads and drafts outreach for us, which means we spend minutes reviewing emails instead of hours researching contacts.
One unexpected benefit is the quality of the emails being drafted.
Because the agent references recent activity and engagement, the outreach feels:
Timely
Relevant
Personalized
This is often difficult to achieve consistently when writing emails manually at scale.
Even with limited data so far, the early conversion is promising.
More importantly, the system is now consistent.
Instead of hoping someone on the team finds time to review MQLs, we now have a process that ensures every qualified lead has the opportunity for follow-up.
For teams considering where to start with AI inside HubSpot, MQL follow-up is a great place to begin. Here’s the framework we followed:
1. Identify a manual process that affects pipeline growth
Look for workflows that require repetitive lead review or manual outreach.
2. Narrow the scope
Instead of trying to automate everything, focus on one stage of the funnel. For us, that was MQL → SQL conversion.
3. Connect your marketing automation
Ensure leads become eligible for prospecting after completing key nurturing workflows.
4. Let the AI draft, but keep human oversight
Review and approve messages before sending.
This hybrid model keeps outreach efficient while maintaining quality.
AI tools often promise efficiency, but the best implementations solve real operational bottlenecks. Prospecting Agent is particularly powerful because it helps sales and marketing teams solve a common gap:
Great leads are generated but never followed up with consistently.
By automating the discovery and drafting process, teams can focus on conversations and relationships, rather than administrative work. When HubSpot introduced its new wave of AI tools at INBOUND, it was easy to feel overwhelmed by the possibilities. But the best approach to adopting AI isn’t implementing everything at once. It’s starting with one clear problem and one measurable outcome. For us, that meant improving MQL to SQL conversion.
And so far, Prospecting Agent is already proving to be a powerful way to do exactly that.