GWI's human intelligence is now in Gemini

TL;DR: Spark MCP 2.0 is live in Gemini Enterprise as a custom connector. Teams can ask about consumers in plain language inside Gemini and get answers from GWI's data, with the source attached. Human intelligence from GWI  is now accessible in all four major AI assistants: Claude, ChatGPT, Copilot and Gemini.

If your organization runs on Gemini Enterprise, GWI can now sit inside it.  Thanks to the new Spark MCP 2.0 integration, Gemini joins the list of AI assistants able to access and interrogate GWI’s data. Through simply adding this custom connector, non-specialists in your teams can ask about your consumers in Gemini and get answers grounded in human intelligence based on GWI's own research, with the source attached.

Spark MCP 2.0 is now live in Gemini Enterprise

With Gemini added, GWI now works in Claude, ChatGPT, Copilot and Gemini, which covers the major enterprise LLMs. Whichever of those your organization has standardized on, your teams can reach the same human data from inside it.

 The latest GWI MCP (Spark MCP 2.0) is one of the simplest routes into accessing GWI's data for any organisation. Data specialists and even non-experts can ask a question the way you'd ask a colleague, and the answer comes from trusted, recent and verified human research rather than the open web.  Responses come back fast, but even more importantly they come back with reliable data that organisations and decision makers can trust.  

The new MCP also brings in additional guardrails and workflows to help tease out the most accurate, and relevant, insights the person needs in that moment.

One in four AI chatbot web visits now goes to Gemini

Wider research* indicates that Gemini is the second most-visited AI chatbot, with 25% of AI chatbot web visits in September 2026, more than Claude (9.6%) and Copilot (1.5%) combined.

So, for Google-first organizations, that's where a large share of everyday AI work already happens. A planner sizing an audience, an analyst checking a behavior or a marketer pressure-testing a brief can now get a read on real consumers in the same window, without exporting data or opening another platform.

Sourced, consistent answers without learning GWI

An answer about your audience is only useful if it holds up and the person asking can trust it. Spark MCP 2.0 is built around four things:

• Real human data, with the source. Answers come from GWI's own research, built on 17 years of research across 100+ markets and 45 billion annual data points. Each answer carries its source, so you can see where a number came from.
• Consistent answers. Underlying tools and guardrails set the market and timeframe first, and saved audiences stay fixed from one question to the next, so your numbers don't drift as the conversation goes on.
• No learning curve. You ask in plain language. There's no GWI syntax to learn and no training to sit through.
• Guided workflows. People who aren't research specialists can query GWI data without learning the taxonomy or datasets, with guidance that takes them from a first question to a complete answer.

In practice, someone can type "What matters most to Gen Z in the UK when choosing a streaming service?" and then ask a follow-up about the same audience, without resetting the market or rebuilding the group. For the teams you support, that means the same question gets the same footing every time it's asked, whoever asks it.

What you need: Gemini Enterprise and GWI MCP access

To get the most out of Gemini and to ensure teams have fully trusted, verified GWI data to make those important decisions, organizations need two things.  Firstly, organizations need to use Gemini Enterprise and then also have a GWI plan which allows individuals and teams to tap into the vast human intelligence available in the data.

*Source: Similarweb via Visual Capitalist, September 2026. Web visits only; excludes apps and embedded use.

‍