Gmail Content Drives Brand Lift in AI Mode

A new study from iPullRank has explored how Google’s Personal Intelligence feature may be influencing brand visibility inside AI Mode search results, with early data suggesting that personal signals such as Gmail activity can significantly increase the likelihood of brands being mentioned.

The research focuses on how Google’s AI systems respond when user accounts are connected to Personal Intelligence features. It compares these results against standard AI Mode outputs without personal data inputs.

While the findings are based on a limited test environment, they offer an early look at how personal context could be shaping brand discovery within AI-driven search experiences.

Study Overview and Methodology

The analysis examined 1,922 AI Mode responses to understand how often brands appeared when Personal Intelligence signals were active compared to when they were not.

Three different Google accounts were used during the testing process. One account acted as a control with no Personal Intelligence features enabled. A second account had Personal Intelligence switched on and was seeded with brand-related signals through Gmail and Google Photos. The third account was a long-term personal account with established usage history and behavioural data.

Across these accounts, the researchers tested eight different categories, including consumer products and services such as coffee machines, running shoes, banks, streaming platforms, hoodies, and SEO-related services. Each category was evaluated using multiple prompt variations to measure how consistently brands appeared in AI Mode responses.

The goal was to determine whether personal data inputs could influence which brands were recommended and how prominently they were displayed.

Clear Lift in Brand Mentions

One of the most notable findings from the report is the increase in brand visibility when Personal Intelligence signals were enabled.

Across the dataset, brand mentions rose from 23.9% in the control scenario to 66.8% when personal data signals were active. This represents a 46 percentage-point increase in overall brand presence within AI Mode responses.

In addition to appearing more frequently, brands also tended to rank higher within the results. The likelihood of a seeded brand appearing in the top three positions increased significantly, rising from 4.5% to 24.9%.

These results suggest that personal context may play an important role not only in whether a brand is mentioned, but also in how prominently it is surfaced within AI-generated responses.

Gmail Identified as the Strongest Signal

A key part of the study focused on identifying which types of personal data had the greatest influence on AI Mode outcomes.

Among the different inputs tested, Gmail content stood out as the most powerful signal for increasing brand visibility.

Brands seeded through Gmail appeared in 53.6% of relevant responses, making email-based data significantly more influential than other sources.

By comparison, brands introduced through Google Photos were cited in only 10.5% of responses, suggesting that visual data may carry less weight than email-based behavioural signals in this context.

This difference highlights the potential importance of structured, text-based personal data in shaping AI-driven recommendations.

Differences Across Industry Categories

The study also found that the impact of Personal Intelligence varied depending on the type of product or service being tested.

Consumer-focused categories such as coffee machines, hoodies, and running shoes showed stronger responsiveness to personal signals. In these cases, brand visibility was easier to influence and more frequently aligned with seeded preferences.

However, more trust-sensitive categories such as banking services and SEO agencies showed much lower levels of influence. In these areas, AI Mode appeared to rely more heavily on broader web sources and established authority signals.

This suggests that Google may apply different weighting rules depending on the level of risk or trust required for a given category.

Web Sources Still Dominate AI Mode Responses

Despite the increase in personalised brand mentions, the study found that AI Mode still relies heavily on traditional web-based sources.

Approximately 49% of citations in the analysis came from external websites, including brand-owned domains and third-party content sources.

Even when personal signals influenced which brands were included in responses, those brands were still often supported or contextualised by web-based references.

This indicates that personalisation does not fully replace traditional search grounding. Instead, it appears to function as an additional layer that sits alongside existing ranking and retrieval systems.

How Personal Intelligence Testing Was Structured

To conduct the analysis, iPullRank built a controlled testing environment using three separate Google accounts with different levels of personal data exposure.

The control account had no Personal Intelligence features enabled and served as a baseline for comparison.

The second account was actively seeded with brand-related signals through Gmail messages and Google Photos content. This allowed researchers to simulate how personal data might influence AI Mode recommendations.

The third account represented a real-world user profile with years of search history, email usage, and behavioural data already embedded into Google’s ecosystem.

Each account was used to run identical prompts across multiple categories, ensuring consistency in testing conditions.

Key Limitations of the Findings

While the results provide useful early insights, the report also acknowledges several important limitations.

Firstly, the study did not have access to Google’s internal ranking systems, retrieval processes, or model-level decision-making. This means the exact mechanisms behind Personal Intelligence cannot be confirmed from the outside.

Secondly, the sample size was relatively small, with testing conducted over a 17-day period across only three accounts. This limits the ability to generalise findings across the wider Google user base.

It is also important to note that Personal Intelligence is an opt-in feature. As a result, these findings apply only to users who have explicitly enabled it and may not reflect default AI Mode behaviour.

What This Means for SEO and Marketing

Despite its limitations, the study highlights an emerging area of interest for marketers and SEO professionals: the growing influence of personal data on AI-driven search results.

The findings suggest that Gmail content, in particular, may play a stronger role in shaping brand visibility than previously assumed. However, this influence does not appear to operate in isolation.

Instead, personal signals seem to work alongside traditional web authority, meaning that both personalised context and broader search relevance contribute to final recommendations.

For marketers, this reinforces the importance of maintaining strong brand presence across both owned channels, such as email, and external platforms, including websites and structured content sources.

It also suggests that AI Mode optimisation may evolve beyond traditional SEO signals, incorporating behavioural and engagement-based data in more meaningful ways over time.

Continued Web Grounding Remains Important

One of the most important takeaways from the analysis is that AI Mode still relies heavily on web-based information.

Even when personal signals were present, nearly half of all citations continued to come from external websites.

This means that traditional SEO fundamentals such as content quality, authority, and structured data remain highly relevant, even in more personalised search environments.

Rather than replacing web sources, personal intelligence appears to enhance and refine how information is selected and prioritised.

Looking Ahead

iPullRank has indicated that further research will explore additional variables, including how long personal signals remain effective over time and whether engagement patterns such as opened versus unopened emails influence outcomes differently.

Future testing will also examine a wider range of industries and product categories, as well as how different prompt styles affect brand visibility within AI Mode.

As Google continues to develop AI-driven search experiences, understanding how personal data interacts with traditional ranking systems is likely to become increasingly important.

For now, the study provides an early indication that Gmail and other personal Google data sources could play a meaningful role in shaping how brands are surfaced in AI-powered search results, even if the full picture is still developing.