With the advent of the Internet, the rise of chief technology officers and chief information officers has emerged. As mobile applications grow over the next decade, some companies have added Chief Digital Officers. A similar phenomenon is currently unfolding, whether an organization appoints a top AI officer or a CAIO.
While most business leaders can increase their team's AI expertise, the benefits of hiring a CAIO can vary, says Birju Shah, clinical assistant professor at Kellogg School and head of AI at Uber for three years.
“Not every business needs a chief AI officer,” says Shah. “All the time until the Fortune 500, most companies will need to train or change current executives to gain AI capabilities, but that doesn't necessarily mean creating a Chief AI Officer position.”
Here, Shah offers some advice on when an organization needs CAIO, how to build roles, and how big and small can determine how a company can optimize AI investments in the long term.
How do you know if CAIO is required?
With all the attention on artificial intelligence, the competition for the top line top AI officers is fierce and expensive. Currently, the position holds a median North salary of $350,000. Top companies threw around seven figures of signature bonuses to attract top talent. And that's before the long-term cost of investing in the infrastructure needed to implement cutting-edge AI strategies.
Given the potential costs of such roles, organizations should use a three-string threshold to determine whether CAIO is needed or whether the placement of other resources is sufficient.
First, the threshold for businesses considering hiring CAIOs is over 1 million customers. “If you're below a million customer scale, it's easier and cheaper to have people handle it. If you're above a million customer scale, things become more subtle.”
Secondly, can your company offer the same product to all customers or is it moving towards personalization? If your company is betting on personalized products and services, AI is a basic investment.
“Netflix is the gold standard for personalization, using machine learning data for consumer streaming statistics,” says Shah. “This is based on the behavior of the user. It shows that your favorite director, your favorite actor, you watch it over and over.
With AI, Netflix will be able to build custom and personalized content at increasingly individual levels. “This is something the studio dreamed of, but it has never been possible.”
Third, on a practical level, organizations need to implement resources and expertise to implement AI.
“This is the biggest threshold companies are missing,” says Shah. “You need someone to do math in your company. You need people to have a bioinformatics background, a diagnostic background. Most companies don't have those skill sets in-house.”
For businesses that meet the 3-string threshold to require CAIO, ensuring a long-term focus starts with finding the right person.
“Chief AI Officer is a very rare ability set,” says Shah. “We need competitive intelligence. We need to know our infrastructure, network equipment, real estate. Besides, we need to know the issues and limitations of AI.”
Future CAIOs should be familiar with where data is for every business unit and should be able to collect the most important data for that purpose, Shah says. This requires that companies be well networked internally to identify and implement use cases that are useful financially and operationally.
What should CAIO do?
Too often in large companies, new CAIOs bring in third-party “Tiger Teams” and give business line leaders the opportunity to use AI, but they don't determine whether AI tools are better than current processes. He points to a recent MIT report, indicating that around 95% of the company's generative AI pilot programs had little or no profit in revenue.
Therefore, it is important to enter the process with clear ideas on how to integrate CAIO within the company. Shah sees two common versions of this implementation.
The first version is what Shah describes as the “platform” version. Here, CAIO works horizontally across the organization and supports all use cases. In this view, CAIO customers are internal executives who approach them to learn the problems that need to be solved. in It will not only be enlarged during the experiment, but also during the experiment.
“You need to have a healthy understanding of your company's workflow, duration,” says Shah. “Those who design strategies must talk to everyone who owns these workflows. For example, at Pharma Company, they need to lead interactions with R&D heads, sales managers and physicians to unleash potential new drug discoveries and talk with people who help adopt those drugs.”
Shah was this type of AI leader while he was with Uber. Uber Rides, the company's riding division, has come to him with ideas for using AI in dynamic pricing, reducing cancellations, or improving service to customers. He ran trip data via AI for progressive improvements.
“Most of what I did with Uber Rides was able to be copied for Uber Eats and Uber autonomous driving use cases,” says Shah.
The second type of implementation is what Shah calls a “partnership” strategy.
“In private equity, for example, an approach to turning a company around is a better partner, not an effort,” says Shah.
In this setup, organizations choose a leader in the low-performing business line and become CAIO. This person will be given a budget to partner with AI service providers such as Microsoft, Openai, and Palantir, and get sales growth and low-performance results like SuperCharge, and use AI tools to bring sales transactions closer faster.
CAIO creates a playbook and is then given to other leaders in the company, whose positions turn into different business units every two years, he says.
“In Chicago, for example, Shore Capital is committed to providing a platform for portfolio companies to come under the umbrella of their partnerships,” says Shah. “Shore Capital will then provide partners and playbooks to improve the outcomes, including profits, revenue growth, sales growth, cost reductions and more.”
How can small and medium-sized businesses implement AI?
Just because a company isn't big enough to require CAIO, it shouldn't be avoided from creating AI strategies. However, Shah says this approach must be more targeted for small businesses as they are trying to capitalize on the possibilities of this new technology.
When AI costs rise, large companies can generally negotiate with vendors to support them at scale. Small businesses may need to be creative because they don't have the same leverage.
“Small businesses must recognize that they have no competition on the vendor side or have no contracts with Openai, but they can compete on the customer side,” says Shah.
In these businesses, implementing AI can mean seeking customers as partners when developing valuable new tools that benefit both groups.
“If a small business is doing well, they call the customer and say, “Can we do AI with you? It may not work right away, but can we do it together?”
Because of close links with clients, SMEs can often extract more unique personalization requirements from the company and provide more value. Small businesses that exceed their weight with AI may be able to gain a reputation for publishing about it.
This edge in operation is often found in highly specialized features. For example, human resource outsourcing companies that supported small and medium-sized businesses in California were able to demonstrate that they were using AI to handle legal claims more efficiently. Apps built by the company for this feature have recently been sold over 10 times the value of the company.
“Large companies don't have the luxury of dealing with AI in a common way. “Instead, they focus on annoying issues and when humans can't solve it, that's when they try AI.
