Affect Performance Team

|Analytics|Nov 4, 2026

GWI Introduces Agentic Consumer Intelligence for AI-Powered Audience Research

GWI Introduces Agentic Consumer Intelligence: AI Agents for Consumer Audience Research

On October 12, 2026, GWI introduced its new Agentic Consumer Intelligence architecture, which enables consumer audience research directly through ChatGPT, Claude, Microsoft Copilot, and enterprise AI systems.

The new tools automate research tasks: they identify target audiences, select the necessary data, perform analysis, and produce findings with source references.

The solution is built on GWI's proprietary research database, which contains information about consumer demographics, interests, media consumption, purchasing behavior, and preferences.

What Data Does GWI Use?

GWI, formerly GlobalWebIndex, has conducted consumer behavior research since 2009.

The company runs online surveys in which respondents report their interests, habits, use of digital platforms, attitudes toward brands, and purchasing preferences.

Its research covers dozens of countries, while a standardized methodology allows audiences to be compared across markets.

According to GWI, the company conducts more than 5.5 million surveys annually across its syndicated and custom research programs.

This process generates billions of individual data points describing different consumer characteristics.

For example, GWI's research database can be used to analyze:

Demographics: age, gender, education, and other characteristics.
Purchasing behavior: product categories purchased, preferences, and purchase intentions.
Brand attitudes: awareness, product usage, loyalty, and preferences.
Media consumption: use of social networks, streaming platforms, search engines, and other channels.
Interests and lifestyles: hobbies, sports, travel, and consumer habits.
Psychographics: personal attitudes, values, and views on social and consumer issues.

According to the company's own estimate, GWI's research datasets contain more than 45 billion individual data points per year.

This figure describes the scale of information collected and processed, not the number of people surveyed or the number of distinct data sources.

For marketing research, GWI uses aggregated metrics calculated from respondents' answers using the relevant sampling methodology. The company describes this approach in its research methodology.

What's New in GWI Agentic Consumer Intelligence?

In January 2026, GWI introduced Agent Spark, an AI analyst capable of answering consumer audience questions using the company's research data.

For example, a user could ask it to identify the most popular social platforms among American sportswear shoppers or compare consumer preferences across age groups.

Agent Spark analyzed GWI data and generated a natural-language response.

The October update expands on this concept.

The system can now independently organize multi-step research, select the necessary audiences, and apply a consistent analytical methodology.

The new architecture includes three core components.

Guardrails: Controlling Audiences and Data Sources

Guardrails allow researchers to define in advance the audiences, segments, and datasets that should be used in an analysis.

For example, a marketing agency can create a segment of U.S. consumers ages 25 to 44 who regularly purchase sportswear.

Once the segment is configured, employees can ask questions about that audience using consistent definitions and restrictions.

This is particularly important for large organizations where different specialists conduct research on the same target audience.

Guardrails help preserve comparability between results and prevent arbitrary changes to the segments being studied.

Guided Workflow: Automating Research Tasks

Guided Workflow allows an AI agent to independently determine the sequence of analytical operations.

For example, a user might submit the following request:

“Compare electric vehicle buyers in the United States and Germany. Identify their main purchase motivations, most popular information sources, and differences in consumer preferences.”

The system interprets the request, selects the necessary data and audiences, performs the comparison, and produces a structured result.

The user does not have to develop every analytical query or build several reports in sequence.

Intelligent Tools: Standardizing Analytical Operations

Intelligent Tools ensure consistent application of filters, classifications, and metric definitions.

For example, when comparing audiences across countries, the system should use comparable characteristics and account for differences in the relevant research samples.

This helps reduce inconsistencies that can arise when different employees independently conduct similar research.

GWI Integrations With ChatGPT, Claude, and Microsoft Copilot

The new architecture makes it possible to use GWI's research tools directly within existing AI applications.

The company offers several integration methods through its Agentic Portfolio.

Model Context Protocol (MCP)

MCP enables GWI research tools to connect to compatible AI assistants, including ChatGPT, Claude, and Microsoft Copilot.

For example, a marketer can open Claude and ask it to identify the characteristics of a new cosmetics brand's target audience.

The AI assistant queries GWI, retrieves the research findings, and generates an answer.

Importantly, the quantitative metrics come from GWI's database rather than being generated independently by the language model.

Spark API

Spark API is designed to integrate GWI's research capabilities into proprietary applications and enterprise systems.

For example, an advertising agency can integrate GWI into its own media planning tool.

When developing a strategy, the system can request target audience characteristics, media consumption patterns, and consumer preferences.

The resulting information can be used to select advertising channels and prepare a media plan.

Agentic API

Agentic API extends this model by allowing users to submit a comprehensive research objective to GWI rather than an individual analytical query.

In this case, the GWI agent independently determines the sequence of actions, performs the required queries, and returns a completed analysis with supporting data.

GWI therefore offers both direct access to research information and automation of its interpretation. The distinction between these approaches is described in its Agentic API documentation.

GWI Ad-Selling Workflow: Automating Advertising Proposal Development

Alongside its new agentic architecture, GWI introduced Ad-Selling Workflow.

The tool is intended for media companies, advertising platforms, and commercial teams that sell advertising inventory.

Its primary purpose is to automate the preparation of sales proposals using research on a prospective advertiser's target audience.

For example, a media company sales manager may plan to pitch an advertising placement to an automotive brand.

The system can research potential car buyers and identify their demographics, interests, and media preferences.

GWI then helps develop a rationale explaining why the audience of a particular advertising platform matches the brand's potential customers.

As a result, the sales team receives a research-backed foundation for its presentation to the advertiser.

Ad-Selling Workflow is primarily designed to reduce manual work in proposal preparation and make research accessible to employees who do not have specialized analytical skills. GWI provides additional details about this ad-selling use case.

How GWI Differs From General-Purpose AI Audience Research Tools

The main difference is where the information comes from.

ChatGPT and other general-purpose language models can generate descriptions of target audiences based on their learned knowledge and publicly available publications.

However, such answers do not always make it possible to verify the origin of quantitative metrics, sample characteristics, or the research date.

GWI uses its own consumer surveys.

As a result, research findings can be accompanied by information about the audience, sample size, research period, and data source.

For example, instead of making a general statement about Instagram's popularity among cosmetics buyers, the system can provide a specific metric for a defined audience in a selected country.

At the same time, research quality still depends on sample representativeness, accurate audience definitions, and the statistical methods used.

The presence of an AI agent does not eliminate these methodological limitations.

GWI should also not be confused with Website Visitor Identification systems or Customer Data Platforms.

The platform is designed to research consumer segments, not identify individual website visitors or analyze the purchase histories of specific customers.

Availability of the New Tools

GWI provides several ways to work with its research data.

Agent Spark is available directly within the GWI platform.
MCP integrations with compatible AI assistants are available to users on eligible paid plans.
Spark API and Agentic API are intended for developers of enterprise applications and analytical systems.

The availability of individual tools and commercial terms depends on the selected product and subscription plan.

At the time of the October announcement, the company had not published independent comparative studies that would make it possible to quantify the accuracy and effectiveness of the new agentic architecture.

Conclusion

GWI's October update expands the use of artificial intelligence in consumer audience research.

While Agent Spark provided a conversational interface for interacting with the research database, the new agentic architecture makes it possible to automate sequences of analytical operations, standardize audience definitions, and use findings directly in enterprise workflows.

The launch of Ad-Selling Workflow demonstrates a practical application of this technology: automating the preparation of advertising proposals based on data about actual consumer behavior.