Client story Kravet LLC gets more contextually relevant search results with Gemini Enterprise

Business challenge

Kravet already uses a custom-built AI agent to help their employees find information like product FAQs and HR policies on the company’s intranet. However, there were several issues with this custom tool:

• High maintenance costs: The custom agent was built outside Kravet’s existing Google Cloud technology stack and required ongoing annual maintenance and extra security.
• Information silos: The existing custom agent only looked at a predefined set of content from the intranet site. It didn’t include employees’ individual work files from Google Workspace, like their email or documents in Drive.
• Slower customer service: When employees work on large projects, they lose time gathering information that lives in many places—email, personal Drive, shared Drive, and external sources. This delay causes a gap in customer interactions, forcing staff to call or email customers back later with the necessary information, slowing down customer service and project starts.

The executive team at Kravet didn’t think these issues could be resolved with a simple IT fix. They saw work inefficiencies and friction points that a smart AI project could solve. Kravet wanted a way to rapidly deliver relevant information to staff so they can answer customer questions in real-time.

Solution

Kravet decided to trial the new Google Cloud agentic platform, Gemini Enterprise, to see if it offered a better, more efficient solution than their custom tool. Kravet worked with Insight, a multiple-time Google Cloud Partner of the Year, to roll out a proof of concept (POC) to directly compare Gemini Enterprise’s capabilities with their existing custom agent.

Defining a clear plan and quick deployment

Insight structured the Gemini Enterprise POC for Kravet into a defined trial with a precise set of deliverables. The collaboration was excellent from the start. “The Insight technical and change management team leads were ready on day one, and their project management approach was very crisp,” says Jesse Lazarus, CTO at Kravet. “This focus made the setup process very simple.”

The most important assurance Kravet got from Insight was clear expectations, which Insight set well for this project. “Insight excels at these types of engagements because they set very clear project expectations,” says Lazarus. “They run very smooth projects, as a result.”

Because Gemini Enterprise is an out-of-the-box solution with native connectors for Google Workspace, it only took the Insight and Kravet teams a few hours to set it up.

Focusing on context and security

Beyond keyword search capabilities, the critical factor for Kravet in the Gemini Enterprise trial was how it would integrate with Google Workspace applications. Most importantly, Kravet needed it to honor existing user rights and permissions.

• User-specific access controls: Kravet needed to ensure users wouldn’t see content in agent search results they didn’t have permission to view. This was a very important strategic goal because it meant Kravet wouldn’t need to replicate data for their AI tool, keeping it in its logical, secure place.
• Out-of-the-box security: Gemini Enterprise’s design inherently honors all pre-existing security settings in Google Workspace. Kravet immediately realized that Gemini Enterprise security works as intended.

Standing collaborative user testing group

Kravet used a standing, cross-functional team of about 40 people from various areas of the business for the Gemini Enterprise trial. All of them already have some experience with AI tools. Insight’s change management team helped prepare these trial users with:

• Hands-on sessions that explained how Gemini Enterprise compares to standard Gemini tools.
• Best practices for prompting the agent.
• Office hours when users could drop in with questions or blockers.

Impact

Qualitative improvements

The trial confirmed for Kravet that Gemini Enterprise can deliver much better search results by combining all of a user’s data sources. The trial led to:

• Better contextual answers: Users found that the AI responses were much more contextually relevant and well-rounded than what they got from their custom intranet agent. The new agent knit together content from Drive, Sites, and users’ personal mailboxes to synthesize succinct responses.
• Faster project catch-up: The platform offered clear, immediate time savings. After time off, one project manager used Gemini Enterprise to review and summarize a dormant project. The agent quickly recapped the latest email exchanges, shared files, file activity, and next steps. “Gemini Enterprise saved the project manager hours reorienting himself and pulling all the project information together,” says Lazarus.
• More relevant deliverables: The agent’s ability to access all the files that specific users can access in Google Workspace made complex tasks easier. For example, it lets the HR team compare resumes to job descriptions to score candidates, pulling in background information from Drive and inboxes—without being asked—to show how the job description was built, resulting in a much better comparison.

Quantifiable benefits

• Cut deployment time from weeks to hours. “After designing, coding, training, and deploying our own custom agent versus what we had to do to enable and deploy Gemini Enterprise for users, it was night and day,” says Lazarus.
• Eliminated time and effort to build robust security, because Gemini Enterprise’s inherent design automatically honors pre-existing Google Workspace permissions.

Future outlook

The Gemini Enterprise proof of concept gave Kravet more confidence in their long-term goal to migrate all their AI activities to Google Cloud. The trial showed the path forward. The fast setup, robust security, and contextual search features proved that Kravet’s future AI development should happen inside the Google Cloud ecosystem. Kravet also saw value in Gemini Enterprise for its access to deep research tools like NotebookLM and the Idea Generation agent.

The engagement also introduced Insight’s strong AI capabilities to Kravet. Kravet really values Insight’s flexibility and responsiveness and will consider them for additional AI projects. “In the future, there’s a path with Gemini Enterprise to have a conversation with your data to automate actions and create real-time efficiencies,” says Lazarus. “Insight can help get us there.”