208-391-4000

before we talk about scope or cost

Your spec was written with AI. Here is what happens next.

A growing share of the documents sent to us are produced with AI, and they read like specifications. AI is capable at a great many things. Hardware development is not yet one of them, which means it also cannot plan a physical product development project. That gap is the reason you are reading this page.

How we actually workWhere to start

What your document is good for

It communicates a possible conceptual approach. That has real value, because it gives us something concrete to react to and it tells us how you are thinking about the product.

That is the whole of it. The approach described is usually not feasible as written, and everything in the document has to be verified by us before it can be relied on. Verifying it line by line costs more than starting from your actual goals, so that is what we do.

We will read it once, for intent. We will not build from it, and we will not treat any number in it as a requirement until we have established it with you.

None of this is a judgment of your product. It is a statement about what a written document can carry. Physical behavior is discovered, not predicted from a conversation.

Why AI cannot plan a hardware program

It has no measured data, because nothing has been built or tested. It has no supplier pricing at your volume. It has not weighed a tradeoff against a real constraint. It will state a target weight, a target cost, and a retail price with exactly the same confidence it states a material name, and only one of those is knowable from a conversation.

The more useful tell is in the documents themselves. Most of them contain a section listing the questions only a physical build can settle. That list is not a footnote. That list is usually the actual project.

Here is the part that matters most. In hardware, the decisions made along the way drive both project cost and product cost up or down, sometimes by a large factor. Material choice, feature set, tolerance, volume, certification path, and architecture all move the number.

An AI document makes dozens of those decisions silently and has no idea which ones are the expensive ones. Working through them deliberately, with you, is the work.

How we actually work

We use a disciplined, phased approach. Each phase exists to retire a specific kind of risk and to produce the information that defines the phase after it. This is the norm in R&D work, and it is the only honest way to run a program where learning is unavoidable.

  • Phase 0. Define the project. Discovery, definition, and risk. Detailed below.
  • Phase 1. Concept generation. Configurations and approaches explored before anything is committed.
  • Phase 2. Engineering and analysis. Feasibility, architecture, and the assumptions that carry the most risk.
  • Phase 3. Iterative design and prototyping. This is where prototypes belong. Building them earlier is expensive and usually answers a question nobody needed answered yet.
  • Phase 4. Iterative design for manufacture. Manufacturing reality layered into the design rather than discovered after tooling.
  • Phase 5. Production preparation and startup. Suppliers, first articles, and ramp.

We do not price an entire program up front, and you should be wary of a firm that offers to. Pricing the whole path on day one requires pretending the learning has already happened. We scope and price one phase at a time, using what the previous phase produced.

Phase 0 is always where we start

A short, focused effort that makes the first step intentional, while decisions are still inexpensive to change. It covers:

  • User and use case definition. Who this is for and what they are actually doing when they use it.
  • A deep dive on the work you have already done. Your research, your documents, your prior builds, and what in them holds up.
  • Identification and stratification of risk. Not a list of everything that could go wrong, but an ordered view of what could stop the program.
  • Low fidelity system and product architecture. The shape of the thing at a level that supports real decisions.
  • Feature set definition. What is in the product, what is out, and what is deferred.
  • Development strategy. The path from here to a product, chosen against your constraints rather than a template.
  • The Phase 1 plan. Scoped and priced off what Phase 0 produced.

We are often able to give an overall development range for the whole project at the end of Phase 0. That range narrows at the end of each phase after it, as decisions get made and risk comes off the table.

The range moves for a reason. The choices we make together push project cost and product cost in both directions, and you are part of every one of them.

Three things worth knowing before we talk

We do not buy ideas

We are a product development firm. We do not purchase concepts, take equity in place of fees, or work for a royalty. You hire us, we do the work, you own the results.

Engineering is one line in the budget

Tooling, certification and testing, first production inventory, packaging, and fulfillment are separate, and together they commonly add up to more than the development that came before them. Plan for the whole path, not the first leg of it.

Licensing is harder than it looks

Many people arrive planning to develop a prototype and license it to a manufacturer. It happens, but far less often than it appears from the outside. Companies license products with demonstrated demand. They rarely license ideas, however well documented.

We have taken on work that started roughly where you are standing. The difference was always a willingness to invest in the definition work first, rather than expecting it to already be done.

Two ways forward from here

You want to explore a Phase 0 effort

Phase 0 engagements often run $10,000 to $20,000 over about two to four weeks. Where yours lands depends on the complexity of the product, the work you have done to date, and how much of it holds up under review.

Tell us that is what you are after and we will set up a conversation about your product rather than about process.

Talk to us about Phase 0

You are not there yet

That is a normal place to be. The reading below is the same material we would walk you through on a call, written by our team and by our CEO for Forbes Business Council. Come back when the investment makes sense.

Start reading

Worth reading first