Most people use about 10% of HubSpot. They set it up, add their contacts, and use it as an expensive address book. The AI features go untouched. The automation tools sit idle. The pipeline management capabilities — the ones that could actually change how they sell — never get configured.
This is not a HubSpot problem. It is a setup problem. Here is the configuration that makes HubSpot's AI features actually work — and turns a CRM into an active sales system.
Start With Clean Data
No AI feature in HubSpot — or any CRM — works well on dirty data. Before you configure anything, clean your contact database. Remove duplicates. Fill in missing company information. Update contact titles and emails that have gone stale. Segment your contacts into meaningful categories that reflect how you actually work with different types of people.
This is unglamorous work and most people skip it. Do not skip it. Dirty data produces unreliable AI outputs, which produces distrust in the system, which produces abandonment. Two hours spent cleaning your database before you build anything on top of it pays back in the quality of everything that follows.
Configure Your Pipeline With Precision
HubSpot's AI deal scoring and pipeline risk alerts are only as useful as the pipeline stages they are monitoring. Vague stages produce vague insights. Here is how to configure stages that make the AI useful:
Define Each Stage With an Exit Criterion
Every stage needs a specific, verifiable condition for a deal to move forward. Not "in discussion" — "discovery call completed and pain confirmed." Not "proposal stage" — "proposal sent and prospect has confirmed receipt." This specificity is what makes pipeline reports meaningful and AI scoring accurate.
Set Close Date Expectations Per Stage
Configure expected close date ranges for deals in each stage. A deal in "Proposal Sent" that has no expected close date is invisible to HubSpot's forecasting AI. Set realistic ranges and keep them updated — this is the data that drives forecast accuracy.
Configure Activity-Based Risk Alerts
Set the number of days of inactivity that triggers a risk flag at each stage. A deal in early stages might tolerate 7 days of inactivity before flagging. A deal in "Contract Sent" should flag after 2 days. Tune these thresholds based on your actual sales cycle, not defaults.
Connect Your Email and Calendar
HubSpot's AI features for deal scoring and relationship intelligence depend on activity data — emails, calls, meetings. If that activity is not being logged automatically, the AI is working blind.
Connect your email provider to HubSpot so every email to and from contacts is logged automatically. Connect your calendar so meetings are captured. Install the HubSpot mobile app so you can log call notes immediately after conversations. These three connections turn HubSpot from a manual data entry burden into a system that captures context as you work.
The AI Features Worth Using Every Day
AI email drafting. From any contact or deal record, HubSpot can draft a contextually relevant follow-up email using everything it knows about the deal. Click the AI button, review the draft, adjust if needed, and send. This alone removes the blank-page problem from follow-up and makes consistent follow-through achievable at volume.
Deal score monitoring. Check your deal scores in the pipeline view every Monday. Deals with declining scores are at risk — identify them early and determine what action is needed before they go cold. Deals with high scores are your best near-term opportunities — make sure they are getting your attention.
Predictive lead scoring. HubSpot's predictive lead scoring assigns a likelihood-to-close score to every contact based on their behavior and characteristics. Use this to prioritize your outreach — the contacts most likely to convert deserve more of your time than the ones who are not ready.
Every Monday morning: open HubSpot pipeline view, filter by risk-flagged deals, review AI deal scores for your top 10 opportunities, identify the three deals that need immediate action, and set follow-up tasks before you do anything else. This ritual takes 20 minutes and consistently produces better pipeline management than any amount of intuition-based selling.
Using Claude Alongside HubSpot
HubSpot manages your pipeline data. Claude helps you think through what to do with it. For complex deals, pull the deal notes out of HubSpot and work through the strategy in Claude — what is the obstacle, what would move this forward, what should the next touchpoint say? For difficult emails, draft in Claude first and then send through HubSpot so the activity logs correctly.
The combination of a well-configured HubSpot as your data and workflow layer with Claude as your thinking and drafting layer covers most of what a salesperson or small sales team needs without adding additional tools.
The Payoff
A properly configured HubSpot setup with AI features active does two things that matter: it makes your pipeline visible in a way that gut-feel never can, and it eliminates the manual overhead that has been quietly stealing your selling time. Deals stop falling through the cracks. Follow-up happens consistently. Your forecast gets more accurate. And the hours you were spending on CRM maintenance go back into conversations.
That is not a small return for configuration work that takes an afternoon.
Take the Free AI Blueprint Quiz →