← All discussions

Sales & Business Development

HubSpot AI Lead Scoring Hits Accuracy Wall

Segun Lane · 3 Nov 2026

HubSpot rolled out its new AI lead scoring model in early July, promising 90% predictive accuracy for B2B prospects. We've noticed our SDRs chasing leads that the AI rates highly but never engage, while several high intent SMB accounts get a low score and slip through our manual filters. What tweaks can we make to the model or our workflow to keep the genuine high value prospects without drowning the team in false positives?
👍 598 💬 2

2 replies

Lyla Higgins · 3 Nov 2026

I fell into the same trap and let the AI score dictate our SDR outreach, ignoring any manual vetting for scores below 80. What helped us was adding a simple engagement rule in Outreach that bumps any contact who opens three emails or clicks a link into a high priority queue, regardless of the AI number. You could've pushed those manual engagement signals back into HubSpot via the custom property API so the model can relearn, and we saw a 12% lift in conversion after a month.

Leon Graham · 3 Nov 2026

We just drop any lead below 85% and run the rest through a quick LinkedIn sanity check before passing them to SDRs.

Log in or create a free account to join the conversation.