Your AI-enabled device submitted to FDA in 3 months, guaranteed.

A fixed fee, quoted on the fit call once we know the device. We do not publish a number because classification, product code, pathway, and whether you need a reader study move it a long way.

What drives it: a full design file, a cybersecurity package including the penetration test, the clinical protocol and statistics, and the submission itself, written by regulatory consultants and engineers together. Teams that clear cheaply usually pay for it in months instead.

We are busy and routinely turn down projects, so we do not discount for volume or promises of future work.

Yes. Expect roughly double the price of a 510(k).

The documentation is similar. The difference is far more back and forth with FDA, and more risk: with no predicate to argue from, FDA has to agree the device warrants a new classification, then negotiate the special controls that come with it. The review clock is 150 days, not 90.

Check whether you still need one. The predicate landscape moves fast, and a competitor taking a De Novo creates a predicate you can file against, turning your De Novo into a 510(k). We look at what cleared recently before assuming the harder path.

Timeline. Your device is submitted within three months of kickoff or you get a partial refund. The clock starts once you have provided the inputs and named your people. It pauses for unpaid invoices, the Thanksgiving and Christmas holidays, anything outside anyone’s control such as a government shutdown, and any time we are blocked by your team.

Clearance. If FDA issues a not substantially equivalent decision, you get a partial refund. The one exception is what you would expect: it does not apply if the NSE came from the device missing the acceptance criteria we agreed in writing during the strategy phase. We guarantee our work, not your model.

Both are in the contract, with the refund amounts stated.

No. We have done it in three months many times, and that is what the guarantee is written against. Some engagements run four months, and rarely longer.

It works because the engineering and clinical tracks run in parallel from week one, the strategy is locked before the clock starts, and we write about 95% of the documentation rather than coaching your team through it.

What pushes it longer is usually unfinished software, slow data, reviews sitting unread, or FDA feedback changing the study. The clock pauses while we are blocked by your team, which is how the guarantee survives those cases.

After submission, FDA’s review clock is 90 days, and most 510(k)s take four to six months once you count answering FDA’s questions.

If your device does not have a clear predicate and a clear clinical evidence strategy, chances are we will want to do a regulatory strategy engagement first.

The three-month clock assumes the pathway is settled: the product code, the predicate, the claims, and the acceptance criteria you will be judged against. Week one locks those down and schedules everything against them, which only works if the answers are already defensible.

Strategy is a smaller, separate piece of work, with an FDA Pre-Sub where FDA’s view is uncertain, and the three months starts when it ends. Better to spend a few weeks settling the pathway than find out in month five that the predicate does not stretch.

Worth real thought. It is one of the few decisions that changes both your risk and your commercial options.

Splitting a platform from its algorithms banks an early clearance on the low-risk piece, and a data delay then blocks one submission instead of all of them. A well-specified API between them keeps future algorithms from forcing a re-clearance of the host.

Bundling adds about four weeks of review but sells as one product. A third option is to clear the algorithm alone, output a report the hospital’s existing viewer can display, and skip the viewer entirely.

Splitting rarely changes the total calendar. It changes what gets blocked when something slips.

The question we most often answer with bad news.

Train and test on separate sites. Geographic separation is strongest; temporal separation is a weaker fallback. FDA generally wants test data from at least three sites and has grown stricter about this.

Non-US data is not automatically disqualifying, but in most cases FDA will want at least half the data to come from the US.

If the data does not support a submission yet, we will tell you to wait. We have talked clients out of starting for this reason.

This is what a PCCP is for: it lets you ship defined changes without a new clearance, as long as the original submission specifies the scope of those changes and how you will validate them.

It works well for retraining within a defined envelope, and poorly for expanding the indications for use, which FDA treats as a new submission regardless.

Whether you need one is decided in the strategy phase, because it has to go into the submission, not get added later.

We write about 95% of the documentation. What stays with you is real but bounded: containerizing the algorithm, running the verification and validation tests, and joining weekly sessions so the code and the documents match.

Week one is the most demanding, because that is when we lock the intended use, the claims, and the architecture.

We can work alongside you on a retainer, supporting your team rather than taking over.

What we cannot do is offer the timeline or clearance guarantee. Those depend on us owning the work from the start, and on decisions we would not have made. Better to say so than attach a guarantee we cannot stand behind.

A fair challenge, and one put to us directly: a list of forty documents does not tell you what is being built or when.

The deliverable is the design file and the filed submission, and the outcome is clearance. The document inventory is how we track it, not what you are buying. Each phase has a named end state tied to a milestone, and the date is fixed from week one.

In most cases it does not. The fee is fixed for the scope in the contract, and nothing changes without a written change to the SOW you agree to first. No surprise line items.

The exception is being blocked. If we are blocked more than four weeks because your team cannot address an issue, we may start a retainer to keep the team available so the project can resume quickly. There is an added fee for that period, and the team will help you get unblocked.

We can scope it, and plan the usability testing for your submission.

The use-related risk analysis and the task analysis come first. They identify the critical tasks and decide whether a summative evaluation is needed at all. If one is, it has a hard floor of 15 participants that cannot be negotiated down.

You need a full QMS before you can commercially distribute, not before you can submit.

You are not starting from zero. The design and development portions are set up as part of this engagement. Our separate QMS service stands up the rest in about two months, and it is a one-time build: later devices add their design file to the system you already have.

One warning: fully automated, AI-driven quality systems have produced FDA warning letters. Keep a human in the loop.

Separate from our fee, paid to FDA just before submission. They rise most years in October.

Small Business Designation cuts them by about 75% if your gross receipts including affiliates are under $100M. It is simple to apply for and takes about two months, so start early. We have seen submissions held up waiting on it.

Also outside our fee: buying data, clinical annotators, and running the study.

Yes, although it changes the nature of the engagement.

We can draft just the software and cybersecurity documentation while your consultants handle the overall submission, or work alongside your team on a retainer. Either way we cannot offer the timeline or clearance guarantees, because those depend on us owning the critical path.

We do not offer hourly consulting.

When the device is hardware-first and the software is incidental. We do not take on implantables, catheters, surgical instruments, orthopedic hardware, or the electrical and mechanical work on capital equipment.

When CE Mark or other non-US submissions are the main goal. We are a US FDA firm and partner for Europe rather than pretend otherwise.

When you want a light review or hourly consulting. We add the most value in a major role, so for a second set of eyes we are not the right fit.