How Gemini AI Helped Promevo Replace Its Expensive, Complex E-Commerce Platform

Promevo employees pride themselves on being multi-faceted. Our company is not as large as some of our competitors, but we make up for it in responsiveness and curiosity. In other words, we love to find creative ways to solve problems.

As a result, we’re super comfortable with Gemini AI, Google’s suite of generative artificial intelligence tools. Every department and employee is working with Gemini (and Gemini Enterprise) to find ways to improve our services and client satisfaction.

It could be as simple as building agents that summarize support tickets, or check lines of gPanel code.

For me, Gemini AI helped replace an expensive, bulky e-commerce platform that required constant hand-holding and manual intervention.

The end result? A sleek digital catalog built around quotes rather than transactions, with minimal operating costs and human intervention—and record traffic and satisfaction from our clients.

The Problem with B2B E-Commerce

Promevo is known for our Google Cloud services and development, as well as gPanel and elite Google Workspace support. But we’re also one of the largest ChromeOS device sellers in the world.

Much of that Chrome success came from our e-commerce shop, which enabled our clients to complete purchases on their own.

But there were three major issues:

1. Our ChromeOS clients are mostly educational institutions and governmental organizations who almost always need quotes for approval before completing a purchase.
2. The cost of these transactions can be tricky. Many of these organizations are tax exempt, which requires verification. Plus, calculating shipping for hundreds of Chromebooks and Meet Kits is a manual process.
3. The modern economy is wild. Prices change daily. Tariffs, supply chain disruption, and rapid AI advancements in hardware—it's difficult to keep up.

There are also operational complexities in connecting all the various downstream systems involved (including the e-commerce platform, the CRM, our billing platform, order management, distributor systems, and more).

No matter what, every transaction requires a human touch. The challenge was making it easy for clients and convenient for our team to manage.

The Solution

As a result, we decided to build a quote-generation catalog. In other words, an e-commerce platform without the ability to complete a transaction. It needed to surface our products for easy browsing, and give clients the ability to create a custom quote.

Once they had their items and quantities added, it needed to be simple to submit that quote to our ChromeOS team. From there, the ChromeOS team would execute the transaction as usual.

This simplified solution would address clients' need for quotes. It would allow the ChromeOS team to capture the best pricing across distributors. And it would reduce systems complexity in trying to connect the complex e-commerce platform to all our downstream systems.

Enter Gemini AI & Gemini Enterprise

As the head of marketing at Promevo, this project is my responsibility. I know a bit of code and can manage a content management system (CMS) but by no means am I a developer.

Like you, I’ve seen the hype around using AI to navigate coding projects (a.k.a. vibe coding).

I could hire a vendor to do this. I could even bring in another e-commerce platform with quotes built in.

Instead, I challenged Gemini to help me build something specific to Promevo's needs. I’l detail the steps I took below, or you can watch this brief video.

Phase One: Creative Iteration

The tools at my disposal:

• HubSpot (our CMS) has a function called HubDB that enables us to publish web and email assets from a database. It’s like building web pages from a spreadsheet.
• Our product table is also in HubSpot (which is also our CRM), so I knew we could easily transmit active products into the HubDB.
• As a Premier Google Cloud partner, we have access to Gemini AI and Gemini Enterprise

My first step was iteration. I used some of the Prompting Best Practices I’d picked up from our AI experts to come up with how to go about this, or even if my HubDB idea had legitimacy.

Once Gemini and I had the concept in mind, my first step was to create a Gem using HubDB and HubSpot CMS documentation. (Gems are Gemini instances built around a specific context, including files and instructions, so you can drill deep into a topic without needing to explain the context in every prompt.)

From there, the HubDB Gem guided me through the following steps:

• Exporting from our e-commerce platform, so the data could fit HubDB standards
• Updating the HubDB settings and processes for importing and publishing the database
• Understanding how to add new columns and populate data on an ongoing basis
  
  

Phase Two: Building

With a foundation in place, I pulled in Gemini Enterprise to help sort and edit the e-commerce export. This helped me do things like edit values to match HubSpot’s expected properties and extract data from product descriptions.

For example, it can identify that a Chromebook screen size is 32 inches by pulling that from an online description - then populate the “screen size” column with 32 inches instead of 32” or 32inches, the latter two of which wouldn’t fit the pre-set properties in the HubDB.

The next challenge was moving thousands of images hosted on the e-commerce platform into the HubSpot file table. This could be done manually but would take weeks (very painful, tedious weeks, if I'm being honest).

Turns out there are APIs for this exact thing (specifically the HubSpot Files API). Gemini wrote a script and walked me through the process step by step. The next day, all our files had switched over.

With a clean HubDB database that contained all our needed properties, including product images, the Gem was then tasked with helping me build five critical assets:

• A catalog page surfacing the entirety of our product table, including search and basic filtering
• A virtual cart function to store and transmit the user's quote
• A product template for each of those products
• A quote summary and submission page with a HubSpot form
• A confirmation page confirming the user's request after submitting the form

I added a document about Promevo brand standards to the Gem, as well as gave it some context around e-commerce best practices, and it did the heavy lifting from there.

From there, it was all iteration. Change this color, move this item, add a box here; coding by chatting, in other words.

Some examples of things it helped me do in minutes that would take days coding by hand:

• A conditional product gallery (if a product has more than one image, it will create a gallery of the additional images; if not, no gallery appears)
• A visible “View Quote” button that increments as the user adds more item to their “cart”
• Automatic pagination on the catalog page that adapts to the user's view, search queries, and available items

All things told, this project took about 40 hours to execute. If I weren’t so picky, it could’ve been far less.

The Results

The new catalog launched December 31, 2025. In January, https://shop.promevo.com had the highest traffic in the history of the site, with more requests than we’ve ever seen in a single month.

We're no longer beholden to a price published on a listing. Instead, our ChromeOS team captures the best available price at the moment a client asks for it.

Adding items is as simple as creating a row in a spreadsheet and filling out columns. Removing items is just as simple, and the entire site responds in real-time.

Of course, Gemini produced tutorials on everything for the team.

Our team still has to calculate tax and shipping (or lack thereof) and generate a sign-able quote for the client. But those things were always needed and we can now focus on creating efficiencies in those processes.

Phase Three & Beyond

Gemini continues to help us make the quote catalog better. We’ve already added a bunch of functions, including a special offers module, a featured item module for pages across our site, and advanced filtering.

We're working on an Agent that will update product details for us, as well as modules for our most popular products, semantic search, AI chat, and much more.

As cringy as “vibe coding” may sound, this platform saves the company a significant amount of money because we no longer need an expensive e-commerce platform nor multiple people to maintain it. It also has clear ROI in time savings, both in the actual construction of the Catalog but in ongoing maintenance.

You can check it out yourself at https://shop.promevo.com.

Think you have a project Gemini can help with? Reach out to Promevo's experts today and we can help you get started.