Pricing AI like natural stone.

Before GOGOGO LLC, I built and ran a natural-stone home-decoration company. Travertine, marble, granite, onyx. We sourced slabs from quarries, polished them, sold them into kitchens and hotel lobbies and bathrooms across İstanbul and the surrounding cities. That company ran for three years, and it taught me how to price.

Most AI startup founders price by cost-plus: figure out what an API call costs, add a margin, charge per call. It's tidy and it's wrong. I price our four products — Goddo, GoPeople, GoVista, GoTrack — the way I priced natural stone. Outcome, not infrastructure. Imagined kitchen, not raw slab.

The natural-stone lesson

When a customer walked into my showroom and pointed at a travertine slab, she wasn't paying for a slab. She was paying for the kitchen she could already see. The polished slab in her hands wasn't the product. The imagined backsplash was. The contractor's quote in three weeks was. The dinner parties she'd throw in two years were.

I priced accordingly. Two customers pointing at the same slab paid different prices, because the kitchens they were imagining were different. The slab cost the same; the value was different. If I'd quoted cost-plus, I'd have left half the margin on the table for the customer who wanted the full luxury build, and overcharged the customer doing a starter bathroom.

The four tiers

Every GOGOGO product is structured as four tiers. The tiers don't track API-call volume directly. They track the use-case sophistication of the imagined kitchen on the customer's side.

Tier 1 — the starter bathroom

Small team, single use case, one workflow they want automated. They have a clear problem and a small budget. Their imagined kitchen is one tile. We price this tier to be a no-brainer — a few hundred per month, no annual commitment, self-serve onboarding. The margin is thin but the volume + product-led growth pays for itself.

Tier 2 — the family kitchen

Mid-sized team, multiple use cases on the same product, integrating into their existing tooling. They want help, not magic. Their imagined kitchen is a full room with appliances. We price this tier to reflect time saved across their team — typically the equivalent of a part-time hire's salary, monthly. They pay it because they can do the math themselves.

Tier 3 — the executive build

Enterprise team, several products bundled, custom integrations, dedicated success rep. Their imagined kitchen is a luxury renovation with bespoke choices. We price this tier annually, with committed-use discounts, and we lead with the strategic narrative — not the per-call cost. "Here's what your operation looks like in eighteen months" beats "here's our API rate."

Tier 4 — the hotel lobby

Large enterprise, all four products, multi-region deploy, strict SLAs, custom legal. The imagined kitchen is a flagship hotel. Pricing here is bespoke. We don't publish numbers. We start with a deep operational assessment, build the integration plan, and the pricing flows from the scope. The contracts run to three-figure-thousand annually, and the customer's ROI math is also in three-figure-thousand annually. The margins are real and so is the value.

What goes wrong when you cost-plus

I see this in three out of five AI startups I talk to. They've calculated what their per-call cost is, slapped a 40% margin on it, and built a pricing page that reads like an AWS invoice. Three problems:

1. The Tier 4 customer feels nickeled. A flagship hotel doesn't want a per-tile invoice. They want a fixed budget for a fixed scope. Per-call pricing makes them feel like they're paying you to use you, which is the wrong emotional structure.
2. The Tier 1 customer freezes. A small team can't predict their monthly bill if it's metered. They'd rather pay too much for predictability than too little for variance. Per-call pricing pushes them away from buying.
3. You leave Tier 2 entirely on the table. The mid-market customer doesn't have an API engineer. They can't read your cost matrix. They want a flat number per month for an outcome. Per-call pricing is invisible to them.

I have watched companies leave half their margin on the table because their pricing page was an honest reflection of their cost structure rather than the customer's value structure.

The two questions I ask before quoting

Whenever I sit across a table from a Tier-3 or Tier-4 prospect, I ask two questions before any pricing comes out of my mouth.

“1. "In eighteen months, what does this look like if it goes well?" 2. "What's the cost of not doing this?"”

The first question maps their imagined kitchen. The second one converts their value into a number they can defend internally. Once I have both, the price quotes itself. I don't have to negotiate. I just name a number that fits between question two's pain and question one's value, and we sign.

What this looks like at GOGOGO today

Goddo's four tiers run from a free starter (Tier 0, actually, with Goddo) up to enterprise white-label deployments. GoPeople's tiers are bracketed by team headcount and message volume, not by API calls — because HR directors think in headcount, not in tokens. GoVista is screens-per-region. GoTrack is cameras-per-store. In each case, the unit the customer buys is the unit their business already measures itself in. Never our infrastructure unit.

“Customers buy the unit they already measure. Your job is to find that unit and price in it, not to make them learn your unit. Natural stone was sold by the slab because kitchens are designed by the slab. Multi-agent AI should be sold by the team or the store or the screen — whatever your customer's operation is already measured in.”

What I'd tell another AI founder

Throw away your cost-plus pricing page. Sit down across from a customer. Ask them what their imagined kitchen looks like in eighteen months. Quote against that, not against your inference invoice. Your customers will thank you for it. Your runway will thank you for it. The contractor who builds the actual kitchen — that's your engineering team — won't be involved in the price conversation at all, which is exactly correct.

If you want to talk pricing, customer ops, or how we structure tiers, I'm easy to reach. okanozalan.com or [email protected].