Skip to content
Fast Company Executive Board

The Fast Company Executive Board is a private, fee-based network of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience.

AI and data democratization

With zettabytes of data available, why haven't business tools caught up with consumer applications of AI?

[Images: greenbutterfly/Adobe Stock]

Profile photo of Kathy Leake

With available data for most business users being orders of magnitude greater than it was 20 years ago, the problem now among most companies is access for business users without overloading analytics teams with requests. For reference, at the beginning of 2020, the amount of data was estimated to be 44 zettabytes—one zettabyte has 21 zeros!

Data democratization is a concept that has taken hold, but what does it actually mean? The simplest definition of “data democratization” is making data accessible and available to all employees and stakeholders without bottlenecks and gates.

I was recently on a call with a chief analytics officer with a wide breadth of experience in BI and analytics. He told me he remembers when Tableau came out. With over 20 years of experience, he has seen many tools and experimented widely with them. It was his opinion that at a base level, these tools are about data delivery. We call them business intelligence tools, which they are, but what they do is deliver insights via data.

He lamented the differences between a tool and a solution, but the main thing I recall from our conversation was how he felt that modern BI tools are behind the curve in using AIML to predict trends for a deeper business impact.

We discussed how business users have had to accept sub-optimal experiences with BI relative to everyday consumer applications, and that the field is ripe for innovation.

Take a free weather app for instance. With a fairly good degree of certainty, the weather app on your phone can predict the highs, lows, and regional weather patterns for your area. Heck, you can see where a storm is going to be in the future, visually. I digress.

The broader question is, why haven’t business tools caught up with consumer applications of AI? With zettabytes of data available, one might think this would be a natural application of AI. What is the reason companies have not caught up? Why haven’t the tools in the market caught up? The difference comes down to trust, fear, and nuance.

Every business situation that requires data is nuanced based on role and individual purview; this presents a problem, although not one that is insurmountable. When a well-built AIML system spits out an answer, it can make an existing plan or forecast look silly or wrong. Alternatively, if the AIML doesn’t have context, the technology can look ill-developed or ill-informed. This can cast doubt over the technology and bring critique and mistrust.

Where are the enterprise tools that can take an “in-house” AIML model that a well-informed data science team built and allow for easy consumption of the outputs? Also, where are the tools that can bring an “out of the box” AIML solution to companies that need them? This is the next step and where innovation and progress can be made.

While I agree that there could be some fear of the technology for the reasons described above, I think the overall sentiment toward AI in BI tools is uncertainty. Most tools “work” and promise the best insights, but at best deliver incremental benefits. Current tools are “working,” and they may even have their own models for AI and ML, but where are the enterprise tools that can deliver, disrupt, and help organizations move toward true self-service analytics and data democratization?

Laying the foundation for the democratization of data takes vision, patience, and intentionality according to Adobe’s Eric Matisoff. Is data democratization on the horizon and how will that happen? Incrementally and culturally and with enterprise tools that utilize AIML in a way that is scalable and available.

Matisoff points out that checks and balances are vitally important, as is good data governance. Governance will build trust in the outputs, and checks and balances should assuage critics’ concerns. Incrementally speaking, future company management should take a long-term view, set goals, and be encouraged with incremental changes because this will take time. Culturally speaking, leadership should invest in a culture of data transparency as well as leaders who share the same sentiment. What is the point of implementing a new tool unless there is a culture to back it up?

At the end of the day, data democratization is going to be a process, and it is going to take companies that are forward-thinking and open to celebrating the small milestones along the way. It is also going to take tools that are innovative and scalable, and a buyer who will not put up with anything less than excellence and who has high standards for the tools they use.


Kathy Leake is the award-winning Founder and CEO of Crux Intelligence, a next-gen AI-powered business intelligence solution.