AI Display Concepts in a Shopper Marketing Pitch

A shopper marketing agency described the same meeting to me three times this year without noticing it was the same one. The team walks into a brand's office with a strong strategy deck, a shopper insight nobody disputes, and one slide for the in-store activation that reads "bespoke floor display, premium finish, hero SKU at eye level". Everything up to that slide is argued. That slide is imagined separately by each of the six people looking at it.

I have spent close to four years in point-of-purchase, on the manufacturer side, watching what arrives from agencies and what happens to it afterwards. This post is about the visual half of a shopper marketing pitch. What the brand is deciding when it looks at a display concept, what you can promise about that image, and what has to happen the day after they say yes.

What is a shopper marketing pitch actually deciding?

Two different things, and they get judged by different people.

The first is the thinking. Which shopper, which barrier, which moment in the store, which measure of success. That part travels perfectly well as words and charts, and agencies are good at it.

The second is the thing itself. The physical piece that will stand in 180 stores with the brand's name on it. A brand marketing director signs off on a budget for an object, and objects get approved by being seen. When the object is described instead, the approval does not really happen in the meeting. It gets deferred, or it gets granted on a mental image that turns out to have been somebody else's.

That second half is where in-store work is unusual. A social campaign can be pitched on a concept and a tone, because the asset is cheap to remake. A floor stand across an estate is a manufacturing commitment with a lead time, and the brand knows it when it looks at the slide.

Why does the display end up described rather than shown?

Because visualizing it used to cost more than the pitch was worth.

To put a credible fixture on a slide, an agency had three options. Commission a 3D render from a studio, which is real money and several days on a pitch you might lose. Ask a friendly manufacturer for a free concept on the promise of the order, which works once or twice per relationship and quietly pulls the design toward what that shop likes to build. Or draw something rough in house and hedge it in the presentation.

So the sensible thing happened. Agencies wrote the display in words, put a competitor photo beside it as a mood reference, and saved the visualization for after the win. Everybody in the category did the same, which is why in-store slides across the industry look interchangeable.

The cost of that habit shows up later. The concept the brand eventually approves is not the one they had in their head during the pitch, and the difference surfaces in the revision rounds, which is where display projects lose their calendar. I went through that arithmetic in AI vs hiring a designer for POP display concepts.

What changes when three concepts go into the room?

The question changes from whether to which one.

A single visual is a referendum on one person's taste. Three directions from the same brief is a conversation, and it does two things at once. It shows the client that their brief had more than one reasonable reading, which is flattering rather than threatening, and it surfaces the disagreement inside their own team while it is still free. The retail director who wants more facings and the brand director who wants more brand space argue in front of you, in the pitch, instead of in round four.

It also changes what your strategy is worth in the room. When the insight slide and the fixture slide are equally concrete, the fixture stops being a decorative appendix to the thinking and becomes evidence of it.

That is the specific thing the free plan exists to unblock, and fifteen credits cover a pitch's worth of directions before you have spent anything.

What can you honestly promise in front of the client?

Say this once, early, in your own words. This is the direction, not the specification.

An AI concept settles the look, and the look is most of what a brand is deciding at pitch stage. Format, material family, finish, where the brand sits, how the product is merchandised. Getting that agreed internally and with the retailer is usually the slowest political step in a campaign, and a concept image gets it done in one meeting.

What the image does not carry is anything a workshop measures. No millimeters, no thickness per part, no joints. Proportions can look completely convincing and still be wrong for the actual pack, because nothing in a raster image knows how deep the box is or how many facings the brand wants per shelf.

That gap is not a problem in a pitch as long as nobody pretends otherwise. The failure mode I see from the manufacturing side is an agency that lets a beautiful render harden into a promise, and then a shop has to explain to the brand that the cantilever does not hold a full facing of product. Frame it correctly and a client who has been told where the line sits trusts everything on the other side of it.

The one promise you can make without hedging is confidentiality. Most pitch work runs on unreleased packaging, so it is worth knowing what each vendor does with what you upload. Nothing rendered or uploaded to AI POP Displays trains AI models, on any plan.

What has to be true of the render to survive the questions?

Four things, and they are the reason a general image model is the wrong instrument for this particular slide.

The format has to be a format. FSDU, endcap, counter unit, glorifier, shelf tray, totem. These are named objects with known proportions, and a general model reads them as style words. Get the base-to-header ratio wrong and the render is a sculpture, which the client may not spot but their manufacturer will, in front of them.

The material has to behave. Acrylic, corrugated board, PET, metal and wood each do particular things with light, edges and thickness. An adjective in a prompt gets the mood of a material rather than the material.

The product load has to be their product. This is the detail that wins or loses credibility fastest, because the one thing everyone in that room knows by heart is their own pack. Facings, shelf counts and placement have to come from their real packshots instead of being composed by the model into something photogenic.

The store has to be their store. The question a brand asks is what it looks like in the aisle, so the fixture needs to sit in a supermarket aisle, a pharmacy counter or a beauty boutique rather than on a studio backdrop.

In AI POP Displays those four are structured inputs rather than prompt adjectives, and references travel labeled by role, so a packshot gets used as product and a sketch gets used as structure. That is the layer I spent close to four years learning, built once so it does not get retyped per brief. AI POP Displays runs on Google's Nano Banana Pro. If your next pitch has a fixture slide in it, that brief is the one to run through it rather than a test prompt. For how the wider category divides up, the tools roundup covers what each one is built for.

What happens the day after the brand says yes?

The pitch ends, the campaign starts, and the concept goes out to manufacturers to be priced. This is where agency projects leak time.

Send one approved concept to three shops and you do not get three prices for one display. You get three different displays, each with a price. One assumes 5 mm acrylic and a bonded joint, another 3 mm and a mechanical fixing. One sizes the shelves to the packs, another to the picture. Every reading is reasonable and every one moves the number, so the agency compares quotes that are not comparable, and the differences surface at the sample stage where changing anything is expensive. I wrote that gap up in full in from AI render to quotable display.

The other half of the day-after problem is scope. A pitch wins a campaign rather than a fixture. A floor stand, a glorifier at the counter, a shelf strip and a stopper, spread unevenly across an estate on a fixed budget. Which stores get which piece is arithmetic nobody enjoys doing in a spreadsheet, and it is the subject of planning a POP campaign with an AI agent.

What we're building next: Bellto

Bellto is an AI agent for POP campaigns that takes a launch from the brief to shop drawings, built for brand and shopper marketing teams and the agencies that run campaigns for them. It is the part that starts where the pitch ends.

You tell it what you are launching, meaning brand, product, channel, stores, budget and date, and it asks for what is missing before proposing anything. It proposes the campaign mix and the materials, floor stand, glorifier, shelf strip, stopper, with the budget split per store. Then it designs every piece with you at real scale in parametric 3D, with proportions derived from the product dimensions and the facings, so a dimension change recomputes the whole piece and re-runs the manufacturing checks in seconds. On the worked piece published on the landing page, a four-panel acrylic counter glorifier of 200 × 200 × 250 mm, that came to 4 parts and 4 cut DXF files, with 0.0 mm³ of overlap measured across its 6 pairs of parts and a STEP that reimported with no solids lost. The whole piece recomputes in under a second after a dimension change.

Every approved piece comes out in five formats. A 3D model, a part-by-part cutlist, dimensioned drawings, a STEP of the assembly and a cut DXF per part. That is a package a workshop can quote without redrawing it, and every model goes through engineering review before delivery. Bellto then suggests manufacturers that fit by material, format and volume. The quote comes from the manufacturer, and now every manufacturer is quoting the same piece.

For an agency that is the difference between winning a campaign and then spending six weeks translating it. We are building Bellto now, the waitlist is open to any brand and agency, and we are contacting the first ones soon to run the first real campaigns end to end. It opens by invitation, in small groups, and the campaign you describe when you sign up sets your place, so a concrete launch with a retailer, a store count and a date moves faster than general interest. Pricing goes first to the people on the list, and there is no card. POP manufacturers can sign up too.

How I would run the next pitch

Take the brief you already have and produce three directions from it instead of one, with the format, the material, the product load and the store correct in all three. Put them beside the insight slide. Tell the room once that these are directions rather than specifications, and let the client's own team argue about facings while it costs nothing.

Then decide, before the meeting, what the day after looks like. If the answer today is emailing a picture to three manufacturers and hoping the numbers line up, that is the part to change next.

If there is a pitch on your calendar this month, the free plan covers its fixture slides, and if what you are about to win is a whole campaign rather than one display, the Bellto waitlist is where that work goes.