How we price AI product builds

Most agencies make you book a call to find out what something costs. That works for them because the call is their best chance to anchor the price upward. It works against you because you wasted thirty minutes finding out we are out of your range.

We do not do that.

This is how V19 prices, what the tiers actually mean, and the three things that move the number.

The published tiers

Three engagement shapes for AI product work, all on the services page. The summary version:

• Sprint - from $750. One focused week. Fix one thing - an audit, a prototype, an unblock.
• MVP - from $4,500. Three to four weeks. Idea to launched product. You keep the code.
• Fractional CTO - from $3,500 per month. Embedded build and advise, ongoing iteration.

The full per-service breakdown sits behind the accordions on the services page. Each one has Starter, Growth, and Premium bands with USD and INR numbers.

These are starting points, not final prices. They exist so you can do the back-of-envelope math before you book a call.

What actually moves the price

Three things, in order of impact.

1. The shape of the problem, not the size.

A four-week MVP that ships one user flow against a clean stack is a known quantity. A four-week MVP that requires us to reverse-engineer a legacy system, integrate with three regulated APIs, and design around a hardware constraint is not, even if the surface feature set looks the same. The first is Starter pricing. The second is Premium pricing, sometimes with a discovery contract before the build.

We will tell you which one yours is on the discovery call, and we will not pretend the second is the first.

2. Whether the answer is "use AI" or "build AI."

Most projects do not need AI built from scratch. They need AI used well - a Vapi voice agent, an off-the-shelf RAG library, a Claude API integration, an n8n automation. These ship fast and price low.

A small percentage of projects need AI built - a custom retrieval pipeline against domain data, a multi-agent orchestration layer, an eval harness for a regulated decision system, a fine-tune. These cost three to five times more and take longer. We are honest about which category you are in.

3. How much certainty you need at launch.

A prototype that proves a concept costs a fraction of a production system that holds up under traffic, evals, observability, security review, and the seven boring infrastructure layers that turn a demo into a product. We will build either, but we will not pretend one is the other.

If you tell us "I need to show this to investors next month," we will quote a prototype. If you tell us "I need to put paying users on this in 60 days," we will quote a production build. The two numbers are different on purpose.

What is included that other agencies bill extra

• Discovery and scoping (no charge until you have a written proposal in hand).
• A weekly demo every Friday. No extra "stakeholder reporting" line item.
• 30 to 90 days of post-launch bug fixes and small iteration on most projects.
• A written runbook at handover. The repo, the infra, the credentials - all transferred.
• Senior engineering on every commit. There is no junior tier here to "save you money" by being slower.

What is not included

• Third-party software licences and AI API costs. We help you forecast them and you pay them directly.
• Design assets beyond what the product needs. If you want a brand bible, that is a different engagement.
• Marketing, growth, performance ads. We build the product. We will introduce you to people who do those if useful.
• Ongoing maintenance past the included window, unless you move to retainer.

Build with equity

For founders pre-seed where cash is tight but the vision is sharp, we will sometimes take a mix of cash and equity (typically five to fifteen percent depending on stage and scope). This is selective - we have to believe in the project, not just the founder. If you want to explore this, mention it on the discovery call.

What to do if our published prices are above your budget

Two honest answers.

One: Pre-build, before money is the constraint, fund the discovery. A $750 Sprint to validate scope and ship a prototype will tell you whether the full build is even worth it. Many of our best client engagements started this way.

Two: If even the Sprint price is out of reach, we are probably not the right vendor for you right now, and that is okay. We will not lower the price to win the deal. Underpriced work makes for resentful teams, and we have seen what that produces. We would rather you find a vendor who is genuinely in your price range than have us deliver an unhappy version of V19.

That second one is the hardest sentence in this post. It is also the one that lets us keep the published prices honest.

If the numbers above match what you are working with, book a call. The discovery is free, the proposal lands in 48 hours.