Web analytics has undergone significant transformation in recent years. Laws such as Law 25 in Québec and the General Data Protection Regulation (GDPR) in Europe have reshaped how companies collect and use online data. While both aim to protect users, their scope differs: the GDPR applies internationally to any company processing European data, whereas Law 25 applies only to organizations operating in Québec. Both frameworks emphasize informed consent, responsibility, and transparency, but the GDPR imposes stricter penalties and broader enforcement (Commission européenne, 2023).
These regulations, combined with browser-level privacy shifts, have limited the tracking capabilities that once fueled digital marketing. In 2024, Google initially planned to eliminate third-party cookies through its Privacy Sandbox initiative, but later opted for a hybrid model allowing users to choose whether to permit data tracking. This change reflects the complexity of balancing privacy, personalization, and performance, and highlights the need for collaboration between industries and users in redefining analytics strategies.
Therefore, can generative AI help marketers better understand their customers when traditional metrics fall short?
Less Data, Bigger Challenges
In the past, companies could easily monitor how visitors navigated and interacted with their websites. Now, explicit consent is required before tracking behaviour, resulting in incomplete datasets. Platforms such as Google Analytics 4 (GA4) rely on machine-learning-based data modelling to estimate missing events or conversions, methods that can be useful but not always accurate (Google Analytics Help, 2024).
Consequently, marketers must turn to complementary sources of insight, such as customer comments, social media reviews, surveys, or chat transcripts. These qualitative signals reveal not only what users do but why they do it, offering emotional and contextual understanding that numeric dashboards alone cannot provide.
How AI Could Help
Generative AI is uniquely positioned to analyze such unstructured data. Beyond counting clicks or sessions, AI models can read, summarize, and classify thousands of text entries to uncover underlying patterns or sentiments.
For example, AI can conduct sentiment analysis of customer reviews to detect whether users express frustration or satisfaction and identify the themes driving these feelings. Large-language models (LLMs) can also help generate personas or journey maps that synthesize behavioral and emotional dimensions of customer experience (McKinsey & Company, 2024).
More advanced use cases involve cross-referencing structured and unstructured data. Imagine correlating GA4 engagement metrics with topics found in reviews: a spike in “checkout” complaints aligned with a drop in conversions could guide targeted UX improvements. Such AI-assisted hypothesis generation could accelerate marketing research and experimentation.
Adobe Analytics, powered by Adobe Sensei, represents another leap forward. In 2023, Adobe introduced Sensei GenAI services, enabling marketers to obtain conversational insights and automatically generated recommendations within dashboards (Adobe Inc., 2023). According to Adobe’s Analytics Features Powered by Sensei report, these capabilities can detect anomalies, identify contributing factors, and summarize performance in natural language, helping analysts understand why a metric changed, not just what changed (Adobe Analytics Datasheet, 2024).
Case Study: HubSpot
HubSpot, a leading inbound marketing platform, offers a concrete example of how generative AI complements analytics rather than replacing them.
In 2024, HubSpot integrated OpenAI’s GPT-4 into its CRM to power tools such as the AI Assistant and ChatSpot, which can summarize marketing reports, extract insights from campaign performance, and generate content recommendations based on CRM data (TechCrunch, 2024).
The platform’s State of Marketing 2024 report revealed that 64 % of marketers already use AI tools to interpret analytics or optimize campaigns (HubSpot, 2024). For a subscription-based company like HubSpot itself, understanding why customers churn or stay is crucial. By combining GA4 data with customer-service transcripts, its own teams (and its users) can detect sentiment shifts and behavior patterns signaling retention risks.
This case illustrates how AI can bridge quantitative performance metrics (such as click-through rates and session time) with qualitative customer feedback, helping organizations move from raw data to strategic insights.
Advantages and Limits
| Advantages of AI in Web Analytics | Limits and Risks |
| Integrates qualitative and quantitative data to enrich understanding of customer behavior. | Risk of bias or misinterpretation in AI-generated analyses. |
| Automates data summarization and reporting, saving analysts’ time. | AI models often operate as “black boxes,” reducing transparency. |
| Enables analysis without invasive tracking, aligning with privacy laws. | The use of user-generated content may raise concerns about consent and ownership. |
| Detects patterns and anomalies humans might overlook. | Requires data quality and human oversight to validate conclusions. |
While these technologies can enhance marketing intelligence, they also raise concerns about accuracy, fairness, and accountability. Generative AI can hallucinate patterns or overlook cultural nuances in text, underscoring the need for human expertise in interpretation.
A Critical View
In my view, the real opportunity of AI in analytics lies not in automation but in interpretation. Machines can surface patterns at scale, but only humans can interpret their meaning and assess their ethical implications. The marketers who will thrive in this new landscape are those who treat AI as a collaborator, a partner that expands insight, yet still relies on human context, empathy, and responsibility.
Conclusion
The evolution of privacy regulations and the decline of cookie-based tracking are reshaping the foundations of digital marketing. Generative AI presents an opportunity to reframe web analytics around context, emotion, and language rather than focusing solely on clicks and sessions.
However, its success depends on the ethical collection of first-party data, transparent model design, and collaboration between AI systems and human analysts. As current examples from Google, Adobe, Contentsquare, and HubSpot suggest, the future of web analytics will not be purely algorithmic but hybrid, where human creativity and machine intelligence co-create a deeper understanding.
Yet, this progress raises a deeper question: if AI can recreate patterns from limited data, are we truly protecting privacy, or merely reinventing surveillance in a more sophisticated form?
References
- Adobe. (2024). Introducing Sensei GenAI for Adobe Analytics. Retrieved from https://www.adobe.com
- Commission européenne. (2023). Le Règlement général sur la protection des données (GDPR).
Contentsquare. (2024). AI-powered Experience Analytics Platform. Retrieved from https://contentsquare.com - Google. (2024). Privacy Sandbox Update. Retrieved from https://blog.google
- Google Analytics Help. (2024). Data Modeling in GA4.
HubSpot. (2024). State of Marketing Report 2024. Retrieved from https://www.hubspot.com/state-of-marketing - McKinsey & Company. (2024). The AI-Powered Marketing Organization.
TechCrunch. (2024). HubSpot Integrates OpenAI to Launch ChatSpot and AI Assistant.