What Is Bellto? An AI Agent for POP Campaigns, From Brief to Shop Drawings

Bellto is an AI agent for POP campaigns that takes a brand's launch from the brief to shop drawings. That is the whole product in one sentence, and the rest of this post is what each half of it means, because both halves are doing real work.

I have spent close to four years in point-of-purchase, and Bellto is what I am building now. This is the reference page for it, written for the people who will use it rather than for a launch announcement.

What is Bellto, in practice?

A POP campaign today lives between an email, a PowerPoint and a supplier. A brand decides it is launching something, describes it to an agency or straight to a manufacturer, and waits one or two weeks for a first proposal to come back. That proposal is usually a picture and a paragraph. If the brand wants to compare two manufacturers, it gets two different readings of the same brief rather than two prices for the same thing.

Bellto sits in that gap as an agent rather than as a canvas. It asks for what it is missing, proposes the mix, and designs each piece with you. Six steps run from the campaign brief to the delivery package.

The output at the end is not an image. It is the set of files a workshop opens in its own CAD to quote from.

Who is Bellto for?

The reader I have in mind is a trade, shopper or retail marketing manager at a brand that does POP several times a year and does not have a workshop in house. The campaign has a retailer, a date and a number, and the constraint is not creativity, it is the loop between deciding what to make and getting something a manufacturer can price.

Agencies are the second half of that audience. A shopper or trade marketing agency running many campaigns at once has the same problem multiplied, and needs every one of them to come out manufacturable the first time rather than after a round of rework at the shop.

POP manufacturers can sign up as well, and they should. What arrives from Bellto already carries dimensions, a cutlist and a defined product, which is exactly the part of the job a shop does for free today just to be able to put a number on a photo and a paragraph.

This is not a hypothetical shift. In calls with US POP manufacturers in August 2026, one told me that 4 of roughly 48 proposals in the last six months had started from a client's AI reference, "and more and more". Another said 5 of about 30 orders in the last month did. Brands are already arriving with a picture. Bellto is about what has to arrive with it.

How does a campaign run through it?

Six steps, in order.

1. You tell it what you are launching. Brand, product, channel, number of stores, budget and date, in a conversation. Before proposing anything, it asks for what is missing. Take an illustrative launch, a premium cutlery range going into 180 stores in November on a 300,000 € budget. The agent comes back with the arithmetic (1,667 € per store) and asks for whatever else it needs before it will propose a mix.

2. It proposes the mix and the materials. The budget gets divided per store and the agent tells you which formats fit inside it. Floor stand, glorifier, shelf strip, stopper, in the material your brand calls for.

3. It designs every piece with you. Each piece in the mix is designed at real scale, with proportions derived from your product dimensions and facings instead of drawn by eye.

4. You change anything, in seconds. More facings, a different height, another material. The model is parametric, so one dimension change recomputes the whole piece and re-runs the manufacturing checks without waiting on anyone.

5. Approved pieces come out manufacturable. Five formats each, covered in the next section. Every model goes through engineering review before it is delivered.

6. It tells you who to ask. Bellto suggests manufacturers that fit by material, format and volume. The quote comes from the manufacturer.

If a campaign like that is on your desk right now, the Bellto waitlist is where it starts.

What do you get for every piece you approve?

Five formats, and each one has a job at the workshop.

• A 3D model. The assembly, in the geometry everything else is derived from.
• A part-by-part cutlist. Material, thickness, size and joint for every part.
• Dimensioned drawings. Front, side and top, with hidden lines separated onto their own layer so the shop can read what is behind what.
• A STEP of the assembly. The interchange file every CAD package opens.
• One cut DXF per manufactured part. Closed contours, ready for the cutting table.

That combination is the point. A render approves a look. It carries no millimeters, so a manufacturer receiving one has to redraw the thing before it can price it, and two manufacturers will redraw it two different ways. A package with dimensions, joints and cut files in it can be quoted without redrawing, which is also what makes comparing two quotes meaningful.

What has actually been built?

Two pieces have been built and measured, and the numbers below are theirs.

The first is a four-panel acrylic counter glorifier with an exact envelope of 200 × 200 × 250 mm. Four parts, four cut DXF files, and 0.0 mm³ of overlap measured as volume across the six pairs of parts. The STEP was reimported and verified with no solids lost. The drawings carry hidden lines on their own layer.

The second is at floor stand scale, a four-shelf unit of 1000 × 650 × 1820 mm with a header and casters. It comes to 241 parts and 250 cutlist rows, 89 manufactured and 152 bought in, each with its material, profile and joint. Going from 16 to 20 facings recomputed the entire unit from the parametric geometry alone in one minute and eighteen seconds, with the manufacturing checks green afterwards. On the smaller piece, a whole-piece recompute after a dimension change lands in under a second.

Compare that to what the same change costs today, which is an email to the manufacturer and days of waiting for a revised proposal that may or may not have re-run any checks at all.

Those are the numbers behind the five-format package, and the reason Bellto is built around a package rather than a picture.

How is this different from a render tool?

A concept render and a manufacturable package answer different questions, and a campaign needs both.

The render answers "will the client approve this look", and it needs format proportions, materials that behave and believable product load to be worth showing. That is the job AI POP Displays does, and it is where the concept image in a POP campaign comes from. The package answers "can this be built at this budget, and what exactly is being quoted". Nothing in the render carries the millimeters that question needs.

Structural CAD sits on the other side of the same gap. It draws the board, the fold and the cut, and it is used by the people who cut material. Image generators draw the look. As I went through in where AI fits in a POP display manufacturer's workflow, those are two halves of one workflow. Bellto is built on the point where they meet, and for the brand side of the desk rather than the shop. The wider map of what exists in this category is in the best AI tools for POP display design, and the closest single comparison is the Vizcom head-to-head, where Bellto sits as the third column.

How do I get access to Bellto?

Through the waitlist. It is open to any brand or agency, and to POP manufacturers, and it asks for a work email plus, optionally, the campaign on your desk.

We are building it now, and we are contacting the first ones soon to run the first real campaigns end to end. It opens by invitation, in small groups. Pricing goes first to the people on the list. There is no card.

One naming note, since search will hand you other things. Bellto, the AI agent for POP campaigns, lives at aipopdisplays.com/bellto. Any other company using that name is unrelated to this.

If you are launching something into stores this season and the manufacturing package is the part that worries you, join the Bellto waitlist and tell us what the campaign is.