11 February 2025

This AI Analyzes Your Website Like a $10k Consultant [Tutorial]

1Frameworks

7Insights

Frameworks in this episode

Marketing7 steps

Reason-Browse-Refine Conversion Audit

Turn AI browsing observations into prioritized, implementation-ready CRO tests

Read →

Insights & moments

The myth-busts, hot takes, explainers, and tools worth keeping.

Hot Take· 1

Hot Take16:00

AI Browsing Will Dramatically Expand Website Testing

The host predicts that browsing agents will increase the number and speed of feedback loops businesses can run across their web assets. Early adopters are expected to compound their advantage by learning and iterating faster than competitors who ignore the technology.

Browsing agents reduce the cost of collecting website feedback.
More feedback loops can accelerate optimization.
Early experimentation may create a widening competitive advantage.

“the implications to really scale the amount of testing and feedback loops any business has to try to get more value out of their website…”
Host · 16:00

#ai-agents#experimentation#competitive-advantage#growth

Explainer· 2

Explainer07:00

What Operator Found on HubSpot's CRM Landing Page

The agent praised the page's messaging, prominent CTA, visuals, social proof, information, and general user experience. Its improvement opportunities included clarifying the CTA benefit, strengthening the message hierarchy, reducing scrolling friction, and adding testimonials or case studies.

Add benefit or reassurance to the main CTA.
Surface key benefits and features earlier.
Reduce the scrolling needed to reach important information.
Strengthen trust with testimonials and case studies.

“Get free CRM, no credit card required as an example”
Host · 07:00

“reduce the amount of scrolling required to access important information and features”
Host · 07:30

#hubspot#landing-pages#cro#b2b-saas

Explainer14:30

Browsing Agents Can Review More Than Landing Pages

The host extends the use case to competitor pages, YouTube channel assets, video descriptions, and authenticated software experiences. For products, a person can take control to log in and then let the agent assess setup difficulty or a specific workflow.

Analyze strong external pages to learn why they work.
Review public channel pages, descriptions, and transcripts.
Audit authenticated setup and task flows after human login.
Use agents to evaluate how easy or difficult actions feel.

“If you find something, a page that you think is really good, have operator look at it and tell you why it's really good, even…”
Host · 15:00

“it can read transcripts, but it can't go through and watch the video”
Host · 15:00

#browser-agents#user-experience#competitive-analysis#onboarding

Tool· 2

Tool06:00

Operator Exposes Browser Control, Notifications, and Session Video

Operator runs inside its own browser and allows a person to take over when an action requires intervention. It can issue notifications for login or status needs, while its sharing feature provides a recording of the navigation performed by the agent.

Operator browses through an observable browser session.
A person can take control when intervention is needed.
Notifications can surface login or update requests.
The completed navigation can be shared as a video recording.

“what it will actually do is it will give you the video recording of the entire navigation that it did.”
Host · 06:30

#operator#browser-agents#session-recording

Tool15:30

Use Qwen 2.5 to Build a Custom Audit Agent

OpenAI Operator is presented as the ready-to-use route, while the open-source Qwen 2.5 model is suggested for building a custom browsing system. The custom route is positioned as an option when a business wants to run the workflow at substantially greater scale.

Operator offers an immediately accessible implementation path.
Qwen 2.5 provides an open-source alternative.
A custom agent can support larger-scale audit operations.

“if you wanted to maybe build a custom agent do this at like real scale, you could use Quinn 2.5 to do that.”
Host · 16:00

#qwen#open-source#browser-agents#automation

Takeaway· 2

Takeaway03:30

Use Slow Reasoning Models for High-Value, Reusable Work

The host distinguishes advanced reasoning models from faster transactional models. Their slower responses are justified when the output is important and reusable, such as a foundational prompt that can support many future audits.

Reserve deeper reasoning for high-impact work.
Create reusable assets rather than repeating expensive reasoning.
Use faster models for more transactional tasks.

“what you wanted to do is you want to use a one Pro on work that's very important and that you can reuse.”
Host · 04:00

#reasoning-models#prompting#ai-workflows

Takeaway11:00

AI Can Teach the Rationale Behind CRO Recommendations

The refined output did more than prescribe page changes: it explained why each recommendation might work. This makes advanced models useful to people without deep marketing expertise, while still supplying practical alternatives such as button copy and sub-labels.

Ask models to explain the rationale behind recommendations.
Use contextual explanations to build marketing understanding.
Request several implementation patterns and copy variations.

“you could know nothing about marketing, right? Like, you could know nothing, and it's going to tell you why it's making these recommendations.”
Host · 11:00

#cro#marketing-education#copywriting#ai