AI vs BI

Last night at the AI panel a few wise folks brought up the topic of BI.

Is AI = BI? Is BI = AI? Is one a subset of the other or are they completely different? Let’s unpack some of this.

AI manifests itself in the form of software. Software is algorithms coded using a programming language. From that perspective, AI = BI. They are all just Turing Machines.

Digging deeper, however, there is a difference between the set of algorithms being typically applied to AI products vs BI products. One can use a model like Linear Regression in both AI and BI but the variety and depth of models and algorithms being applied to AI products is much richer than BI. Most BI products are a combination of SQL queries and joins. I’d say BI is not as sophisticated as AI (although it’s not her fault if the creators don’t fully exploit it).

The data quantity standards are lower for BI. This is a complement, not a criticism. Since BI doesn’t advertise complex statistical models (Machine Learning, Deep Learning), it’s humble and is willing to do simple analytics on whatever data is available. AI, on the other hand, simply isn’t applicable if there’s not enough data. You can really can’t call it AI if your dataset has 37 rows of data. Please don’t do that.

BI is almost always tied to dashboards. There’s almost always a visual element associated with BI products. With AI, while some applications like Predictive Marketing & Sales have a rich visual layer in the form of dashboards, a lot of AI quietly works behind the scenes and you may never even notice it.

New interfaces that don’t rely on a black screen (think Alexa and family) are being associated with AI and not BI. But that’s because of the next point.

Perception — this is likely the biggest difference. BI is jaded. BI is tired. It’s been around many years and has its share of successes and failures. It simply doesn’t have the wow factor anymore. We, as a species, are driven by what’s new and shiny. BI isn’t new and shiny. There are no BI startups making news. None of the big tech companies are talking up BI. They are all talking up AI. There’s “social proof” that AI is the thing to do. You wouldn’t even be reading this if the title was “Learn more about BI today”.

In sales & marketing, AI and BI have the potential of working together. Data from an AI-driven product is consumed by different types of BI users at the organization. At one company we’re working with, some users will be consuming SalesTing data into Microsoft PowerBI. At another one, it’s Tableau. AI and BI have already become friends.

Organizational structures — most companies already have a BI team and I don’t see them renaming it to an AI team anytime soon unless they are a company selling AI products. BI teams are typically an exclusive club. They do the heavy analytics and inform management who then make decisions that impact all employees. With AI, many products will simply have AI built-in and each employee will use without ever realizing (or caring) what it is. AI will directly touch more humans than BI.

Conclusion — I’m going to take a stand. AI is a superset of BI.

Deep in my heart, however, they are all Turing machines. That’s one truth that will stand the test of time.

Show your support

Clapping shows how much you appreciated Piyush Saggi’s story.