AI 2.0: 3 aspects you should be working on today, part 3

This is the third part of a three-part miniseries on AI 2.0.?

AI 2.0, as I previously defined it, is the natural evolution of AI and data into everyday solutions and business acceptance, also known as the democratization of AI into business reality.?

In this short series, I discuss three concrete ways how this democratization is finally happening and how you should act upon it. They revolve around three key aspects:

1. Democratization of technology: “Embrace pre-built AI services and low code AI” (see part 1)
2. A business approach to AI: “Step away from artificial AI” (see part 2)
3. A shift in organizing AI: “Break down your siloed AI team”

We now dive into the third part.?

Aspect #3: Break down your siloed AI team

Previous parts discussed that smart data technologies are becoming ever more attainable (part 1) and – with the right approach – ever more tangible for business users to actually start exploiting the potential (part 2).?

Are these elements alone sufficient for boosting the widespread use of AI in organizations? Probably not. The organizational component is missing, in other words; how can we organize our companies such that everyday business users get in close contact with AI? If we can achieve that, many more business teams will start adopting AI and integrating it into their companies’ future products & services.

So, how can we really organize ourselves such that all business users get in close contact with smart technology???

The answer is NOT a siloed data, analytics, and AI team. Over the past 10 years nearly all companies have taken a centralized approach at data, analytics & AI. As the technology was complex, the best way was to centralize the few experts into one team. That team often resided in IT or, better still, in a data office that was positioned in between business and IT. Whatever the set-up, while a centralized data, analytics & AI team may have helped to master the technology and take first steps forward, one eventually hits a glass ceiling: the siloed expertise is difficult to bridge to business users. True scaling of AI in the company is hindered. As a result, for many companies (!), the large investments in smart technology didn’t result in tangible benefits.?

The answer IS to integrate AI experts into your business products & services teams. Many businesses are migrating away from a functions-based organizational structure and toward an agile one. This is best characterized by having cross-functional teams that are end-2-end responsible for a certain mission, product or service1. Such cross-functional collaboration greatly improves time-to-market and overall agility of a company. Some refer to this as a transition from a project-centric to product-centric approach’2.

It is crucial to understand that this shift offers a major opportunity to let your AI and data experts (permanently) sit in those cross-functional teams/clusters together with business product owners, business experts, UX specialists, developers, and others. In that way, you optimally exploit the synergies between all those disciplines working together in one team instead of having a separation between a business team that requests smart functionalities3 , and a siloed AI team which delivers. The latter set-up will unintendedly always result in a us-against-them battle, loss of ownership and non-optimal solutions. By decentralizing the data, analytics & AI experts into business products & services teams, those teams will be organically drawn into the possibilities of AI and therefore more naturally adopt its potential.?

Surely, a critical voice will point towards inefficiencies in decentralizing your data & AI experts. How can you still ensure standardization in AI technology, AI methods & tools and AI governance (such as installing a company-wide ML Ops practice, a hot topic these days)? Years of scaled agile literature and practice have shown us that Centers of Excellence are the way to go here4. They bundle people with skills and expertise whose job is to provide leadership and purposely disseminate knowledge within the organization. Centers of Excellence-structures will still provide organizations with the necessary efficiencies of scale in building & deploying smart data & AI technologies.??

In conclusion, as AI and smart data technologies are getting ever more approachable and tangible, the time has come to break down your siloed AI team and have your AI experts join everyday business teams to finally boost adoption and realize the long-claimed potential of artificial intelligence.

?

References:??

1: Organize around products & services: https://www.pmi.org/disciplined-agile/mindset/mindset-principles/organize-around-products-services

2: From Project- to Product-Driven: The Digital Transformation at Ford: https://www.bizops.com/virtual-summit/ford-project-driven-to-product-driven

3: Even more so, ‘requesting smart functionalities’ is hardly ever the case. In a traditional siloed set-up, it’s only the AI team that has the knowledge on such smart technologies. As a result, most AI propositions often have to come from the AI team. By decentralizing the experts into the business products & services teams, those teams will be organically drawn into the possibilities of AI and therefore more naturally adopt its potential.?

4: Centers of Excellence: https://www.pmi.org/disciplined-agile/people/centers-of-excellence