Stage 6 — Deployment
A model sitting on your laptop helps nobody. The last stage is handing it over — putting it somewhere a farmer, a doctor or a shopkeeper can actually use it.
1What deployment is
Tap a stage to stand on it, or press play and watch the work go round. Drag sideways to turn.
It answers: How do we get this into real hands?
2What deployment looks like
- The leaf-disease model becomes an app a farmer points at a plant.
- A hospital model becomes a screen the doctor already looks at.
- A spam model becomes part of the email service, invisible to you.
In every case someone had to install it, connect it to the real data coming in, and teach people to use it. That work is deployment.
A bus in the depot
3It does not stop here
Once real people use the project, you find out things no test could tell you. Farmers photograph leaves in the rain. Users type questions nobody expected. Those discoveries send you back to problem scoping, and the cycle turns again.
4Practice
▶Show the answer
Deployment is the stage in which the evaluated AI solution is integrated into a real-world environment and put into operation, so that it can deliver value and impact to its users and stakeholders. It matters because until a model is deployed it produces no benefit for anybody, however well it performed in testing.
5Check yourself
Which stage of the AI Project Cycle comes last?
A school's AI attendance system has been tested and works well. It is still on the developer's laptop. Which stage is left?