AI Governance is Required for Enterprise AI Applications
AI governance is what makes an AI application viable, and what makes it Enterprise. Before on-boarding a new AI application at the Enterprise make sure you have AI governance to ensure trustworthy AI.
Generative AI applications are currently disruptive to enterprises, as many organizations are scrambling to set up guardrails and policies to catch up with this new technology. This technology has massive appeal to business lines, and many are prototyping with Generative AI for their business use cases. Additionally, Generative AI vendors have made it easier for businesses to access AI, especially for those without data scientist talent.
You can now prototype AI applications in business lines quickly, without the need to include IT, who was previously needed for their coding and software engineering skills. And with the switch of analytics platforms from on-premise databases to cloud data lakehouses, the need for IT becomes seemingly less valuable, as the data teams in business can just set up their own cloud instances…this further allows business to develop at lightning speed, while leaving IT further behind.
But IT can’t be left behind, as AI governance is needed to make an AI application viable at the Enterprise. In partnership with the CDO and the data organization, IT needs to have AI governance in place in order to govern the ‘pillars’ of AI applications at the Enterprise: people, process, tools, and policies. AI governance is needed to ensure trustworthy, explainable, responsible, and compliant AI for the Enterprise.
The problem at the Enterprise for many organizations is that IT has not caught up to this new Generative AI technology, and IT departments at many organizations are currently scrambling to set up guardrails and policies to govern this technology. The problem is that you can’t just piggyback on existing IT governance for BI and software development.
You can’t piggyback AI on top of BI, as they have the following differences:
different pipelines,
different processes,
different lifecycles,
different talent, and
different governance.
Ideally, you would have a data and AI strategy to go along with data and AI governance, but this article is focused on the need for AI governance. I will tie this all together with my colleagues who are experts in data and AI strategy, to see how this intersects with data and AI governance (and AI roadmap) in a future article. I may also assemble an expert panel on this (stay tuned).
As a colleague has stated in one of his articles on Generative AI; “Generative AI is the easiest technology to build prototypes with and the hardest to build products on.”
In the meantime, for Generative AI projects at the Enterprise, organizations need to look at LLM private instances for training (customization), serving, maintenance, and monitoring for enterprise use cases (LLMs in production), while maintaining a firewall between internal corporate servers and external ones. This is still a work in progress at many enterprise organizations, as IT departments (in partnership with their counterparts in the CDO and data organization) catches up on how to govern this for the Enterprise, to make these AI applications trustworthy, explainable, responsible, and compliant.
If you're looking for support from me, here are a few options:
Enterprise Data Science Consultancy: With my consult team comprised of a Senior Data Scientist, Senior ML Engineer, Senior Data Engineer, and Senior Cloud Engineer, we will help you architect and build your Enterprise Data Science platform, and transfer knowledge to your IT team to maintain and optimize it. We will also overlay an MLOps framework to manage the AI solutions you build on this platform. If you don’t have an MLOps team, we will help you build one. Please get in touch about this consultancy here
Coaching and Mentorship: I offer coaching and mentorship; book a coaching session here