Masterclass AMA Session. AI journey checklist: Scaling with confidence under the AI Act

1. Collibra can integrate with the AI OPS systems to record the tests (bias, security tests, ...) and the outcomes, allowing anybody in the AI Council to use that information for risk assessment. Second we can register the risks in the risk register and link it to the use case, the data and/or the model. And relate it again to controls and remediation actions if desired. This allows full traceability and control on those risks. 
2. AI governance is still in motion but NIST has a framework (aka playbook) and the EU AI Act stipulates obligations, so there are guidelines being created. We support those OOTB - allowing you to use them as guidance or checks.
3. This is where governance comes handy: by ensuring visibility on the different use cases and their lifecycle, by allowing the decisions to be clearly documented and audited, by promoting literacy on data and AI - it is possible to provide guardrails and guidance to allow to manage the risk, the cost and the value of any AI use case. 

@nadiaperez Yes we do. For Collibra, the type of AI is not that different. We will change that in next release, depending on the type of model, but for the use case and the connection to data, we want to be flexible. We support classic ML, genAI and Agents are also coming. Information governance is organised by some customers in Collibra, thanks to our operating model and the business glossary. Assets that represent knowledge or information can be created on the conceptual, logical and physical (eg systems) level.