Data Science

Why Everyday AI Can Outshine Moonshots

Back in 2013, Florian Douetteau had already been working in data science and AI for more than a decade, but he didn’t feel the field had suitably advanced in making machine learning enterprise-ready. He decided to take AI democratization into his own hands, founding data science and machine learning platform company Dataiku, where he serves as CEO.

Florian Douetteau

Nearly 10 years later, Dataiku is helping to operationalize AI across a range of business use cases, from fraud detection and customer churn prevention to predictive maintenance and supply chain optimization. If the final destination is weaving AI capabilities so thoroughly into the fabric of day-to-day work that people forget it’s there, enterprises are typically somewhere in the middle of the journey, Douetteau says. To get there, they should look inward.

In this “AI From the Front Lines” interview, Douetteau and Romain Fouache, Dataiku’s chief revenue officer, speak with Beena Ammanath, executive director of the Deloitte AI Institute, about their vision of AI in the enterprise, the importance of building systemization and trust for AI, and how execution will be more important than innovation in democratizing the technologies.

“It’s not a technology issue—we can build platforms able to continually process and enhance data and build new AI on top to optimize business processes,” Douetteau says. “But to make that happen, companies need to transform themselves and the work their professionals do.”

Ammanath: Your company’s mission is “everyday AI” for the enterprise. How do you see AI transforming companies?

Romain Fouache

Fouache: A lot of companies think about AI and advanced analytics in terms of moonshots, and those big use cases can help drive a desire for change. But where I see the most impact is in the more everyday use cases, where AI becomes so commonplace that we think about it the way we think about using a computer: It’s no longer remarkable but rather a standardized, universal tool to advance productivity. There are companies where hundreds of people in HR, logistics, and finance have learned to leverage advanced and predictive analytics to augment each and every one of their activities. That’s the big transformation.

There are so many definitions of AI floating around. How do you define it in the enterprise context?

Douetteau: AI is not about science fiction or robots. It’s about activating data to automate business processes, particularly in those areas where manual, slow, costly, and potentially erroneous decisions can instead be made faster and more effectively. AI is also about using data to imagine your future rather than look backward.

Fouache: We need to think about the business outcomes and not feel constrained by any given definition of AI. Getting the biggest impact from the data is not only about the most advanced machine learning technologies but also often about extremely simple operations like matching data from different sources or computing simple trends. I am OK with putting all of this under the AI umbrella. For example, finance teams can often find success in accelerating their internal audits by moving critical data out of spreadsheets and applying some relatively basic machine learning techniques to automate risk assessment and more efficiently direct the team’s efforts.

Beena Ammanath

How can enterprises build or maintain competitive advantage as AI becomes more democratized?

Douetteau: Access to data has been increasing for quite some time. The real differentiator is how fast an enterprise can embed AI into its business processes. The reality is that organizations should be prepared to change and optimize their business processes so that AI isn’t just an innovation but a core aspect of their operations. From there, it’s all about execution.

Yes, in Deloitte’s fifth annual “State of AI in the Enterprise” report, our research indicates that organizations that have undergone significant changes to workflows or added new roles are more likely to achieve AI outcomes to a high degree. It should become part of the core DNA. What does that look like from a people perspective?

Fouache: The goal should be to embed AI into every business process to the extent that nobody talks about data scientists anymore because everybody is doing data science without even thinking about it. Time and time again, there can be fear that people might not be capable of change, but people have adapted the way they work throughout history to embrace new technologies and drive organizational outcomes. We’ll likely see a major shift as AI evolves from something that can only be done by a handful of individuals to something available to everyone.

In our report, we developed an analysis model defining four profiles of organizations based on the frequency of full-scale AI deployments and the outcomes achieved through AI initiatives. We determined that 27% of respondents still qualify as “starters” in this area. What is your advice for companies early in their AI journey?

Fouache: Early on, the most important thing to consider is how specific applications of AI can position your organization for future success. When doing so, it’s important to get the business involved and choose applications that are going to be relevant to the board. For example, maybe a company is using a static, rules-based approach to pricing. In some cases, a simple AI model can outperform those traditional approaches, providing a clear win that impacts the bottom line. These real—as in applied, not theoretical—proofs of concept can be a valuable way to educate stakeholders about the uses of AI and data.

Many companies are in the stage where AI already has some presence in most business functions. How can those companies most effectively scale their use of AI to create competitive differentiation for the future?

Douetteau: It’s important to build systemization and trust. To systemize AI, a self-service mindset is key, whereby business stakeholders can work within large, centralized systems that serve as a data core—because the more people who can contribute to AI, the faster they can build and innovate it. To create and maintain trust in AI as a long-term asset, companies should focus on their capacity for governing and understanding AI with transparency. They can then develop a good sense of AI’s costs and value and establish the proper governance for data privacy and regulatory compliance, as well as ensure ethical processes are in place.

What are some other emerging tech trends that are important to this evolution?

Douetteau: Innovation is not driven only by existing business applications but also by accessing vast stores of data floating around in the cloud. Another important trend is embedding AI into the business. It’s not about creating another dashboard—people are already suffering from information overload—but about embedding it into processes and automating it so that AI is felt everywhere without creating more to manage.

Fouache: Another big shift has been the availability of large-scale, elastic computation capabilities. Things that used to take six months to requisition can be spun up in six seconds.

What are the biggest AI-enabled business opportunities you’ve seen come to fruition?

Douetteau: There are examples across many industries. Some financial services organizations are using AI and data automation to be more forward-looking and efficient. In the manufacturing space, companies are getting better at managing supply to match demand and getting creative with business models by using data and advanced analytics across their supplier ecosystem. The most successful projects are marked by strong alignment with the strategy of the company and an urgency to change.

Fouache: The examples I love are not the moonshots. They’re the use cases of century-old companies in pharma, engineering, energy, and elsewhere that maintain their relevance by making everything they do better with AI. That includes improving how marketing dollars are spent, developing smarter pricing strategies, improving the design of a jet engine, or simply speeding up positive customer service interactions. Such companies have managed to educate and energize thousands of employees to embrace AI and help customers improve outcomes. It’s exciting to see thousands of people passionate about changing the way they work and companies that could have been driven to obsolescence maintaining their competitive advantage thanks to AI.

This is the 13th article in the series “AI From the Front Lines,” which goes beyond the hype to reveal the opportunities and challenges enterprises can realize through AI and data analytics, featuring the real-world experiences of enterprise executives in conversation with Beena Ammanath, executive director of the Deloitte AI Institute.

Copyright ©2022 Dow Jones & Company, Inc. All Rights Reserved. 87990cbe856818d5eddac44c7b1cdeb8

In this “AI From the Front Lines” interview, Douetteau and Romain Fouache, Dataiku’s chief revenue officer, speak with Beena Ammanath, executive director of the Deloitte AI Institute, about their vision of AI in the enterprise, the importance of building systemization and trust for AI, and how execution will be more important than innovation in democratizing the technologies.

Source: https://deloitte.wsj.com/articles/why-everyday-ai-can-outshine-moonshots-01669665578

Donovan Larsen

Donovan is a columnist and associate editor at the Dark News. He has written on everything from the politics to diversity issues in the workplace.

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button