Artificial intelligence (AI) has become increasingly important as organizations use it to drive innovation, enhance efficiency, improve decision making, and boost digital transformation efforts. But ensuring success requires significant, ongoing efforts.
Consider setting up an AI center of excellence (CoE) to boost the success of your AI initiatives. An AI CoE is a dedicated team of cross-functional individuals that pitch in on a regular basis to develop, implement, and manage AI solutions in an organization. Bringing together a team like this helps align AI initiatives with the overall business model and strategy and ups the benefits of process automation.
But how do you build an AI center of excellence? Use these six tips to make sure you’re well positioned to succeed.
[ Want some inspiration for your automation efforts? Get more than 200 ideas for how to use RPA and IDP. ]
How to set up an AI center of excellence: 6 key steps.
1. Gain organizational buy-in.
Any initiative requires leadership commitment. Your organization’s leaders have to fully back the AI center of excellence for it to succeed. It helps to have an executive sponsor or another business leader guiding the efforts who can also keep the rest of the c-suite and the board in the loop and champion the initiative as it gets implemented.
Make sure to meet your team’s questions and fears with empathy and care. Skittish team members might feel threatened by AI, so communicate the benefits to them in a way that resonates—less manual work, more time for strategic work, or more abilities to advance.
2. Assemble the team.
Akira Kurosawa’s famous film Seven Samurai popularized the film trope of assembling a team of experts to handle a problem. The CoE group is the expert team you’ll need to solve the thorny issues involved in operationalizing AI at your company.
You’ll want to consider the following roles:
- AI strategist/leader: This person will lead the team and be responsible for oversight of the projects and activities the team will undertake. It often helps if you choose some existing technology leaders from your organization to fill these seats.
- Data scientist: These subject matter experts should bring a strong statistical and machine learning background. They’ll build AI models and solutions and identify areas where AI could be applied for further improvement.
- Data engineer: Data engineers will manage building out data infrastructure and ensure data quality.
- Software developer: These engineers will work on building scalable software solutions that integrate with AI models and still maintain strong application performance.
A few roles from other organizational units might be necessary; for example, you might need UX designers to build user interfaces and create data visualizations, business analysts who are well versed in process optimization, and a project manager to keep teams on schedule.
3. Take stock of your current systems.
Once you have buy-in and the team in place, your first order of business needs to be taking a survey of where you are. You’ll want to perform a gap analysis showing where you currently have AI capabilities or projects in place and whether you need to upskill some staff members or purchase software or infrastructure elements. Pay particular attention to your current IT system landscape, as this will affect what you can implement and how fast.
4. Decide which AI projects to take on first.
This step can happen before or after step three (or concurrently). Having a sense of a tangible business problem you want to solve first will often dictate what resources you need, project timelines, and how AI will play into your business strategy and goals. Tackle a specific use case that can generate a tangible benefit for the business. Some ideas include implementing AI-driven document processing for billing departments, using AI for risk modeling in financial lending and insurance, or using AI to analyze and write customer service responses for simple, easy-to-answer requests.
5. Measure progress regularly and iterate.
You’ll want to consistently measure your AI team’s impact on the business. This means setting KPIs and tracking against those measures during development and after deployment. The KPIs you track will depend on your goals for the AI project—for example, an insurance company using AI to process claims might want to track the average time it takes to process a claim or the percentage of claims solved within one day.
Most importantly, you’ll want to have your team iterate on AI. This is why it’s critical to have organizational buy-in early on—you won’t get your AI projects perfect straight out of the gate. But you can start off on the right foot and only get stronger from there.
6. Use AI as part of a broader set of automation technologies.
AI is only one tool in a complete process automation toolkit for automating and optimizing complex business processes. Some tasks are best served by AI; others require robotic process automation or business rules to structure logic. It also helps to use low-code tools that can help you easily build out workflows that deftly pass tasks between automated workers like bots and AI and human employees. This lets you quickly create applications to orchestrate these processes, helping them fit seamlessly into your wider business.
[ Read our related articles: RPA vs. AI vs. low-code and 7 AI/ML Use Cases to Watch. ]
An AI center of excellence keeps you on course.
Unleashing the full potential of AI takes time and commitment—but the effort is worth it. AI can fuel a business transformation that improves efficiency, reduces costs, and drives innovation. Those companies that don’t embrace AI and other automation efforts will fall behind those that do. Build a CoE to ensure your organization stays on course and truly maximizes AI’s potential.
But as mentioned before, AI is only one piece of a holistic automation strategy. Learn more by getting the Gartner Hyperautomation 2022 Trends Report.
