How CDW Cut Demo-Building From Weeks to Minutes with Gemini Enterprise

The Gemini Enterprise (GE) Demo Generator is a low-code web tool that builds fully working, custom demo environments for Gemini Enterprise. It's currently published as an open-source sample rather than a packaged product.  A sales associate can provide the generator with context about a company's business challenge, and the tool provisions a real dataset, an AI agent equipped to act on that data and a live operations dashboard, all built around that specific business rather than a generic script.

Google frames the value in two ways:

1. Speed: A demo comes together in minutes instead of weeks.
2. Realism: Everything runs on actual infrastructure instead of a scripted walkthrough.

None of what it produces is staged. The tool provisions actual BigQuery datasets, Google Maps grounding and Firestore databases, so the agent is set up to run against real infrastructure, not a scripted walkthrough.

Gemini Foundry, CDW's own demo generator tool, is built on the early version of Google’s Demo Generator. CDW designed the tool so that our sales and pre-sales teams can use it directly without engineering support. A member of the sales and pre-sales teams can provide key details using a short form and walk away with a working demo; nothing to configure and nothing to run themselves.

The generator form walks through six steps:

1. Customer information
2. Defining the goal
3. A summary of the agent and its data
4. Deployment
5. Running the live demo
6. Cleanup

Researching a Customer

The demo generator starts with one piece of information: the customer's domain. A member of sales or pre-sales types in a company's website, then the tool researches it, generates a short business profile, flags a handful of relevant business challenges, and suggests specific workflows an agent could automate, each one tagged for how well it fits the platform.

CDW ran a test on our own domain through the demo generator tool. The result was an accurate profile and summary of business challenges including multi-vendor deal registration and rebates, quote generation at scale across a quarter million accounts, and the shift from hardware sales toward SaaS. Once an accurate profile is generated, the user can pick a suggested workflow or build one from scratch.

Browsing By Template

If a domain search does not generate an accurate overview of the customer, you can use the Template Hub. The hub is organized by industry (retail, finance, manufacturing, healthcare, etc.) and by job function (sales, marketing, operations, IT, etc.).

Each template has a library of built in business scenarios to choose from, and that library goes beyond IT and customer service. It could be a retail finance manager adjusting pricing against competitor data, a human resources manager predicting staffing needs from foot traffic and event schedules, a marketing lead generating localized ad copy across dozens of markets at once.

Once a user selects the scenario that closely represents their use case, they can proceed into the workflow.

Keeping Customer Data Anonymous

The generator tool can anonymize all the real company data it pulls in, swapping in a fictional name and other details with context that is accurate enough to make the point without revealing any real sensitive information.

From Click to Live Demo

Once a user locks in a scenario, the demo build takes about two minutes. The tool synthesizes a data model, generates the underlying data set, drafts system instructions for Gemini, and puts together the deployment scripts.

The generator tool then hands off to an automated deployment step that provisions the Google Cloud resources and gets the agent running, without a terminal or manual scripts.

Every generated environment also cleans up after itself. By default, a deployed demo tears itself down 72 hours after it's created unless someone extends it or cleans it up manually first, so unused demos don't sit around racking up cloud costs.

Building on Google's early version of this generator tool meant CDW wasn't starting from scratch, and that head start is apparent in what the tool does for CDW's own teams today.

Before Gemini Foundry, building a custom demo was a fully manual process, taking anywhere from a day to a couple of weeks depending on complexity and resources. Today, the same process takes about 13 to 15 minutes, and it doesn't require a technical person to run it. Anyone on the team can build their own demo now, which frees up the technical team from having to build these by hand for every request. Google has described a comparable shift for its own version, cutting demo prep from a matter of weeks down to minutes.

The bigger shift here goes beyond one tool: sales conversations can start with proof built into the discussion, instead of just a promised result. This is already an everyday reality for CDW's team, and it's a decent preview of where this is headed. With Google investing in more tools like this, it's a safe bet that AI will keep chipping away at the parts of sales that historically take up excessive time and resources.