Building an AI Center of Excellence: Why a CoE?
At D2M, we implement data and Machine Learning (ML) projects for our clients. An issue we often come across is the issue of governance. In a new field, like ML / Artificial Intelligence (AI), it is common for different business lines to be devising and managing their own ML projects, leading to the proliferation of proof-of-concept, unsystematic projects without a common standard, and a disturbing number of ‘not quite ready for prime time’ projects somehow making it to the production pipelines.
To avoid these problems, enterprises have been implementing centers of excellence (CoE) around ML and AI. But what does a CoE provide?
An AI CoE enables an organization to meet business challenges by incorporating ML services across its business units. The CoE provides enterprise-wide strategy, tools, and infrastructure support, hence offering a common framework for different teams to successfully evaluate, develop, deploy, and support AI technologies.
To support their client’s business needs, an AI CoE will provide the following services:
- Defining an organizational AI strategy
- Use case discovery workshopping
- Use case prioritization
- Model prototyping
- Data analysis and project feasibility assessment
- Data management, including data pipelining
- Tool assessment
- AI technology roadmap
- Integration with Information Technology
- Model implementation and deployment
- Model updating, support, and documentation
- Machine learning training and skill development
- Develop common tools that can be used across projects
These services enable model creation, deployment, and governance across an organization.
The verdict? An AI CoE will launch scalable AI initiatives for clients that support their core business objectives and address their most critical pain-points.
Read our next blog post on the AI/ML series which elaborates on the various challenges an organization will encounter in establishing an AI CoE, and the ways in which they can overcome them.