Question 1

How long does an AI project take from start to launch?

Accepted Answer

A typical SparkFire engagement runs 7 to 13 weeks end to end: one week of discovery and planning, two to four weeks for a proof of concept on your real data, and four to eight weeks to build and launch. Simple automations ship faster; systems touching regulated data or legacy infrastructure take longer.

Question 2

What happens in the discovery phase?

Accepted Answer

In week one we learn your goal, examine the data you already have, and agree on what a good result looks like in numbers. You leave with a written plan and a fixed price for the proof of concept. Discovery is paid, and it is useful on its own even if you stop there.

Question 3

What is a proof of concept and why do you insist on one?

Accepted Answer

A proof of concept is a small working version built on your real data, not a demo dataset, delivered in two to four weeks. It exists so you can see honest accuracy numbers before committing to a full build. You get a clear go or no-go decision, and stopping at no-go is a legitimate outcome.

Question 4

Can we stop after the proof of concept?

Accepted Answer

Yes. The proof of concept is deliberately a decision point, not a commitment. If the results don't clear the bar we agreed on in discovery, stopping is the right call and you keep everything built up to that point.

Question 5

What happens after launch?

Accepted Answer

We keep the system running and improve it. That means monitoring how it performs in production, watching for drift as your data changes, retraining or tuning where needed, and shipping improvements as your needs grow. Ongoing support is a standing arrangement, not an afterthought.

Question 6

How much of our team's time will this take?

Accepted Answer

Most of the demand lands in discovery, where we need access to the people who own the workflow and to representative data. After that we work in one-week sprints with a demo at the end of each, so the recurring commitment is roughly one review session a week.

Question 7

Do you work with our existing engineering team?

Accepted Answer

Yes, and it usually produces a better result. We build alongside your engineers, hand over documented code in repositories under your organization, and make sure your team can operate and extend the system after we step back.

How we deliver AI

Discover & plan

Proof of concept

Build & launch

Run & improve

Questions about how we work

Got an idea? Let's make it real.