BeGlobal creates tailored commercial proposals in under a minute

Context

BeGlobal is a Dutch company specializing in branded corporate gifts. When large organizations want to send personalized presents to hundreds or thousands of employees, BeGlobal helps them select and customize products from a catalog of around 100,000 items.

Every commercial proposal is a curated experience. BeGlobal’s team needs to:

• Understand the client’s occasion, budget, and brand.
• Find suitable products in a large product database and CRM.
• Build a visually appealing PowerPoint presentation with product images, descriptions, and on-brand visuals.

PromptGorillas, an AI automation and education agency, partnered with BeGlobal to transform this process. As Lars van Gils, Founder of PromptGorillas, explained, “They had a very large repository of products and one of the biggest troubles was making those offers to companies.”

PromptGorillas used n8n as the backend orchestration layer, combined with a custom frontend, Nano Banana for image generation, Supabase for product data, and Google Slides for the final presentations.

Challenge

Before automation, BeGlobal could only create and send about 50 offers a year.

"They had a huge bottleneck that they could only create and send 50 offers a year. And the main reason they could only send 50 offers and not more was because it took too long to create an offer."

The core issues were:

1. Manual product selection in a huge catalog. Sales reps had to search through a database of roughly 100,000 products to find matches based on availability, color, material, budget, and theme.
2. Manual presentation creation. Once products were selected, the team created a PowerPoint proposal by hand: inserting product images, writing descriptions, and trying to match the theme of the occasion (Christmas, Easter, etc.).
3. Limited volume and growth potential. Because each proposal could take several hours, BeGlobal’s sales capacity was capped. They were missing out on pipeline and revenue simply because they couldn’t produce offers fast enough.
4. No reuse of historical client data. For existing customers, past offers and CRM data were valuable context, but not integrated into the proposal creation process. Everything was rebuilt from scratch for each new request.

BeGlobal needed a way to turn this highly manual, creative workflow into a fast, reliable, and scalable system without sacrificing the quality of the proposals.

Solution

PromptGorillas designed and implemented a fully automated, AI-powered offer creation pipeline, with n8n orchestrating the backend workflows.

1. Conversational intake via a chat agent

The process starts in a custom frontend where a chat agent collects all the information needed:

• Type of product(s)
• Budget per unit
• Occasion or theme (Easter, Christmas, etc.)
• Client name and domain
• Whether it is a new or existing client

The agent summarizes the inputs and validates them with the user before proceeding.

2. Smart product search across 100,000 items

Once the input is confirmed, a second n8n workflow kicks in:

• It checks if the client already exists in BeGlobal’s systems.
• If it does, it taps into CRM data and historical offers.
• It queries Supabase, which stores product data for roughly 100,000 products.
• An AI Agent node in n8n helps generate the right queries against the database, supported by code nodes that construct precise SQL based on the user’s constraints.

Lars described this technical core: “The query is done using a GPT node, which knows how to query the database, and using code nodes to specifically create the right query. It sends it to the database, it transforms the data, sends it back to the frontend using a webhook, and it shows all the data that it’s collected.”

3. Automated themed visuals with Nano Banana

To make the offers visually compelling and on-theme:

• n8n takes the selected product images from the frontend.
• It routes them through Nano Banana to generate contextual, occasion-specific visuals.

What used to take hours of manual image sourcing and editing is now fully automated.

4. Automatic Google Slides proposal generation

The final step is creating a client-ready presentation:

• A Google Slides template lives in BeGlobal’s Google Drive.
• n8n copies and duplicates the appropriate slide layout for each selected product.
• It fills in product details, prices, and the newly generated theme images.

As Lars described: “We start with a single Google Slides template that defines where products should appear. Based on the products selected in the frontend, we duplicate the relevant slides the required number of times and automatically place each product in the correct position.”

The entire end-to-end workflow, from chat intake to finished presentation, is executed by n8n in the backend, triggered and controlled via the custom frontend. The tech stack includes:

• Frontend: built with Cursor
• Backend orchestration: n8n (multiple workflows, agents, GPT nodes, code nodes, webhooks)
• Database: Supabase
• Image generation: Nano Banana
• Presentation layer: Google Slides in Google Drive

Stephen highlighted the impact on execution time: “Right now the whole workflow take around 52 seconds to create a whole presentation. From end to end.”

Results

The impact for BeGlobal is both operational and commercial.

1. From hours to minutes per proposal

Where offer creation used to take several hours end to end, it now takes just a couple of minutes of human time, with the heavy lifting automated and the workflow itself executing in under a minute.

Lars summarized the time savings: “The time to make an offer has really gone down from a couple of hours to a couple of minutes.”

This frees the sales team from repetitive tasks and lets them focus on higher-value activities like prospecting and relationship-building.

2. 10x increase in offer capacity

With the bottleneck removed, BeGlobal can now handle far more demand.

Lars estimates: I think they’re now creating nearly ten times as many offers as before.

More high-quality proposals in the market mean a significantly higher revenue potential, without needing to hire a large number of additional staff.

3. Stronger revenue pipeline and faster response times

Even though long-term revenue metrics are still being collected, the direction is clear:

• More offers generated for the same or less effort.
• Faster turnaround time for proposals.
• Ability to serve a “way bigger pipeline”

4. A future-proof AI automation architecture

The solution is built to evolve:

• n8n’s agent framework, MCP, and modular workflows allow PromptGorillas to plug in new AI tools as they emerge.
• When Nano Banana launched, they integrated it mid-project: “We started the project, and when Nano Banana was released and turned out to be a perfect fit for this use case, we integrated it right away to generate product images.”
• The stack stays model-agnostic and adaptable, enabling BeGlobal to benefit from future AI improvements without a full rebuild.