Introduction

Not all enterprises have reached the same Artificial Intelligence (AI) maturity level. The group consisting of large enterprises are the early adopters of the technology. They have mature AI strategies and show significant synergies with concrete business goals between the various business units. Additionally, they are the best placed to take advantage of the ongoing innovation. After all, AI is all about continuous learning.

State of AI in the Enterprise

In the latest and fourth edition of “The State of AI in the enterprise,” Deloitte provides insights on the state of AI and intelligent automation in business. The report results from a survey from almost 2,900 business executives from 11 global economies engaged in AI strategies and investments within their organizations. The focus is on various behaviors, from the overarching AI strategy and leadership to the technology, data approaches, and the available tools to help the workforce operationalize AI.

AI profiles

Four buckets divide the AI market as surveyed by the research. Each bucket describes the state of an organization towards the goal of a successful AI transformation.

The buckets are:

1. Starters – 29%:
   • The companies in this group are newcomers to the AI transformation. Therefore, they lack experience and knowledge and have a lot of catching up. As a result, the companies at this stage are the least likely to have many successful deployments.
2. Pathseekers – 26%:
   • This group shows that they have the know-how to deploy successfully. The next step is to scale up their efforts towards different AI deployments.
3. Transformers – 28%:
   • Transformers are In the process of being fully transformed. As a result, the business has a working strategy with repeatable practices and delivers the most substantial AI outcomes.
4. Underachievers – 17%:
   • Companies have buildup enough knowledge and have some deployments. However, they still need to increase and improve their transformation practice.

More than half (54%) of the surveyed indicated that they have several different environments ongoing. They see two things as a necessity to help fuel their business—first, the need to have multiple use cases to drive the Ai effort. Second, making significant investments to support those efforts.

The remaining groups (46%) are just getting started with shadow AI environments to prove the value of AI. Unfortunately, many projects won’t see the light of day as they will never get to the production stage for various reasons. However, learning from failures is a part of the AI learning process. Therefore, it is essential to have a tight collaboration between the line of business, the data scientists, and the IT teams at this stage. It starts with understanding the kind of data you have, where it is housed, and what type of data can be leveraged to drive the business.

Practices for success

According to the Deloitte research, four (4) leading practices define the success of an AI transformation.

1. Strategy:
   • The organization needs to translate the business goals into an AI strategy.
   • The strategy must be explained, shared, and followed.
2. Operations:
   • An operational model is required to support the strategy.
   • The model must be flexible enough to accommodate a dynamic environment.
3. Culture and Change Management:
   • An AI transformation can be intimidating and culturally different from a traditional transformation.
   • Fail early, fail often!
4. Ecosystem:
   • Most organizations shouldn’t tackle an AI transformation by themselves.
   • Ecosystems are critical as they connect clients with the right level of technology, solutions, and practitioners. Deliver solutions to the client all the way through implementation, not just advisory but the implementation and the operationalization of that solution.
   • About 83% of the surveyed have two (2) or more ecosystems. So, they are looking at bringing different types of technologies and services to deliver the benefits of AI. Those are cost reduction or complexity reduction, the transformation of a business, or some innovation.

Reasons for adopting AI

There are two (2) main factors why an organization initiates an AI transformation.

The first factor is an unexpected change in the market landscape that forces the organization to adapt quickly through innovation. An excellent example is the outbreak of COVID. Pharmaceuticals did not have the appropriate methodologies to address COVID and relied on AI to accelerate to find answers.

The second factor is competition. An AI transformation becomes a competitive differentiator. The organization with the best price, product availability, and in-person experience will win.

Challenges to the adoption of AI

There are underlying elements needed to adopt AI and add value to the business.

There is no value to be extracted without large amounts of quality datasets and no AI solution. Organizations are swimming in data, but they don’t know what they have as the data is in multiple silos. It could be in legacy data warehouses, data lakes, or it might be in the cloud owned by different departments.

The transformation does require significant investment in infrastructure and continuous alignment with the business goals. What business outcomes do you want to obtain, and what insights will you drive through the data? A risk is that the business might not realize what AI can achieve.

A common challenge is the lack of a talent pool, as they are different types from traditional data analysis. The data scientists, data engineers, and DevOps/MLOps teams work together towards the business goals. It is not uncommon for an organization to rely on an ecosystem of partners to help bootstrap its AI effort. For example, identifying the skillset shortage and addressing the skill gaps. Additionally, guidance on upskilling existing employees and a roadmap to onboard new talent.

There is a particular need for an AI roadmap for organizations in the early stages of their transformation journey. It starts with a starter project to experiment and learn as much as possible. The next step is to bring over most of the tooling and apply it to another business problem. Finally, repeating the process will build up expertise and knowledge that improve the maturity of the AI transformation scale across different projects.

Summary

Not everybody is at the same stage in their AI journey. However, a growing number of organizations are building up enough AI knowledge to increase their success rate significantly. Many challenges still lie ahead, such as data management, business goals, and acquiring the right talent pool, but there are numerous ways to get help. For example, engaging one or more partnerships or ecosystems is an excellent way to provide guidance and bootstrap an AI transformation.

I am a co-host of “Utilizing AI”, a podcast focusing on AI in the Enterprise. We discuss the above topics and many others at length in the podcasts. To learn more about the podcasts, follow this link.