Klarity AI Blueprint
Why we need a blueprint
Accelerate the integration of AI in your industry with Klarity AI Blueprint :
Klarity AI Blueprint facilitates the industrial implementation of AI by proposing design tool chains dedicated to each of the most common industrial needs and problems.
Choose from our library of AI Blueprint and find the one that best suits your needs : The “planification AI Blueprint” can help you deploy an AI that will maximize the level of automation and quality of your industrial process. If your need is more related to minimizing the amount of anomalies and errors in your production chain then the “anomaly detection AI Blueprint” shall be perfect to integrate the most efficient AI for this purpose.
You can also combine different AI Blueprint and easily integrate multiple fine tuned AI to improve the efficiency of your different processes, boosting your company’s overall productivity at the same time.
While any AI Blueprint can be easily and fully customized to fit your need and specificities, it is also possible for Safenai to design a tailor-made AI Blueprint for you.
How blueprint are split for different typology of use cases
Klarity AI Blueprint enables different industrial sectors to be addressed with minimum effort :
- End-to-end process to deploy an AI Function for a sector : for instance, visual inspection.
- Associated tool chain to : enrich dataset, build a robust AI, quantify uncertainty, monitor in operation…
- Can be customised to exactly fit the needs.

How blueprint interact with AI Function, AI Component and AI Models

(source : DeckProductV3.pptx)
How is constitue a blueprint
for each blueprint we will present the following elements :
- Which use case are covered by this blueprint ?
- The associated intended purpose with specific question
- The different activity integrated inside the AI function associated to the blueprint
- The list of artefacts per stage proposed by the blueprint
- The workflow associated to the blueprint
Content of the blueprint :
- Tutorial to build demonstration :
- Content of the tutorial
- Based on notebook, exporting artefact to karity, using kc-kraft
- different step of the demonstration : (TODO DISCUSS AND IMPROVE)
- Naive implementation of use case
- A trustworthy implementation based on confiance.ai
- An operation upgrade during operation
- Content of the tutorial
- A ML Ops implementation to interact with