Ep20 - AI in HubSpot: Founders' Cut

For our 20th episode of HubHype, Erwin is joined by a very special guest, Thalox co-founder Falk von Rönn, for one of the most wide-ranging conversations the podcast has had yet.

They cover how Thalox tested and measured AI tools across the team, real AEO results from a customer case, the newly launched Thalox Normalizer, HubSpot's evolving campaign ROI attribution, and a big debate on the future of SaaS.

Watch the episode here.

How do you actually measure the impact of AI tools?

We recently went through a structured three-month test to evaluate AI tools such as Claude, ChatGPT, Copilot, and others before making a company-wide decision. Rather than going by feel, we tracked every AI use case in a spreadsheet, logging time spent with and without AI for each task.

The result?

→ A 2.6X improvement in time savings across the team using AI tools.

The final decision to go with Claude came down to one key factor: the ability to build our own agents using Claude's functionality, something that is critical for a consulting team that wants to encode and reuse their knowledge at scale.

Erwin also highlighted a pattern he's seeing with customers: even when HubSpot users already have AI features available, like meeting summaries and note-takers, they're simply not using them.

His recommendation: build a library of standard prompts for your team, especially for sales.

What does AEO look like in practice?

AEO (Answer Engine Optimization) is one of the most talked-about topics in HubSpot right now, and we have already seen real results with a customer.

The approach was straightforward: take existing blog posts and knowledge base content and optimize them for AEO to improve visibility in AI-generated answers.

The outcome surprised even Erwin:

→ ~300% increase in mentions and citations - and they weren't starting from zero.

What you can see inside HubSpot's AEO tool today: the prompts people are entering into AI engines, which sources are being cited, and specific recommendations on how to improve your visibility.

As Falk points out, no other major CRM is offering this natively. To do something similar with Salesforce, you'd need third-party tools. HubSpot is bringing these topics into the platform immediately, and that's a real competitive advantage.

One important note: like SEO, AEO takes time. You won't see results overnight.

What is the Thalox Normalizer?

One of the most common data problems we see across customers is messy properties. A single field like "Country" might contain US, USA, United States, and U.S.A., all meaning the same thing, but treated as different values by the system.

This is where the newly launched Thalox Normalizer comes in.

The tool captures messy property values across your HubSpot account, uses AI to identify which values share the same meaning, and normalizes them into a single clean value, without overwriting existing data.

Instead, a new property is created in HubSpot with the normalized output, which can then be used for lead scoring, segmentation, or reporting.

Key things to know:

• Works on top of HubSpot: easy to install, ready to go
• Uses AI to identify and group values with the same meaning
• Creates a new property rather than overwriting existing data
• Can be run repeatedly over time as new data comes in
• Allows manual overrides if you disagree with a suggested normalization

Can HubSpot finally help you measure ROI properly?

ROI attribution has always been one of the hardest things to get right in marketing, and Erwin and Falk have been in business long enough to know that almost no company does it well.

But HubSpot is making progress. Watch more here:

Is SaaS dying?

To close out the episode, Falk turned the tables with a question: given how capable AI agents are becoming, is SaaS on its way out?

Both Erwin and Falk agreed: SaaS isn't dying, but it has to change.

Want to hear the full discussion? Watch the episode below.