Follow Datanami:
June 11, 2021

AI: It’s Not Just For the Big FAANG Dogs Anymore

It’s been said that AI has a 1% problem, that only the biggest tech firms—the Facebooks, Amazons, Apples, Netflixes, and Googles, or FAANGs, of the world–have the resources required to pull it off.  But thanks to the rapid advance of data science technology, the democratization of compute in the cloud, and the availability of data, that is becoming less true by the day.

We’re in the midst of a great democratization of big data and AI that’s benefiting companies of all sizes and maturity levels. Companies across many industries are responding to this shift by increasing in their investments in AI.

According to a Gartner survey from October, nearly 80% of companies had AI projects underway in 2020, and 75% will continue or start new projects through mid 2021. “Enterprise investment in AI has continued unabated despite the crisis,” Gartner analyst Frances Karamouzis said.

IDC predicted last summer that AI spending would double over the next four years, growing from $50.1 billion in 2020 to $110 billion in 2024. “Companies will adopt AI — not just because they can, but because they must,” IDC analyst Ritu Jyoti said.

Deloitte Insights proclaimed that “we are entering a new chapter in the adoption of the current generation of AI technologies.” “Capabilities are advancing, it is becoming easier to develop and implement AI applications, and companies are seeing tangible benefits from adoption,” the company stated in its annual State of AI report last summer.

“It is true that not everyone has adopted AI technologies yet–there are still barriers, and many are working to scale the benefits,” Deloitte continued. “However, it appears that AI’s ‘early adopter’ phase is ending; the market is now moving into the ‘early majority’ chapter of this maturing set of technologies.”

That is good news for companies that are downstream of the giant tech companies that, in large part, set us on the path to AI over the past two decades by creating the big data and machine learning technologies that have turbo-charged statistical computing.

Dulling the FAANG’s Advantage

The way Dataiku founder and CEO Florian Douetteau sees it, the “FAANG mafia” has built-in advantages, in particular when it comes to having gargantuan piles of data and the wherewithal to hire thousands of engineers to ensure that data is ready to be fed into AI systems.

(Michael Traitov/Shutterstock)

But thanks to the advance of enterprise AI software and cloud computing, the advantages of the FAANG mafia (plus Microsoft, making it MFAANG or FAANGM or GAFMAN or something) are no longer insurmountable.

“You don’t necessarily need a very large amount of data to get things done. You don’t necessarily need really complex machine learning to achieve results,” Douetteau says. “We believe very firmly that the success of AI in the enterprise depends mainly on the number of people available internally, and not the size of the data you have.”

Dataiku focuses its development efforts on creating enterprise AI solutions that can be used by data and business analysts, as opposed to highly trained data scientists. The company’s intended users are analysts who are well-versed with Excel, and want to take their data skills to the next level, which the company does in its product by bringing a visual and collaborative approach to the data workflow, from preparing the data to optimizing the machine learning models.

Data scientists may still be needed in some situations, particularly as coaches or mentors to guide less experienced analysts, and for building very complex predictive models. But Douetteau says data scientists today are only needed 10% to 20% of the time. In other words, analysts can make up 80% to 90% of the AI team. For smaller businesses, AI success can be found with a ratio of six analysts to one data scientists, he says.

“Sometimes, there’s this misconception that data science and AI is about moonshots, very large projects, as in you build a brand-new data-driven product or approach in your organization that changes everything,” Douetteau tells Datanami. “But not everything is like the Netflix recommender system…The reality is there’s a lot of things you can start by having a small amount of data. There are lots of initial step you can do in terms of getting simple analysis or model of your data that can definitely have a business impact.”

AI for SMBs (and SMEs)

It may seem counterintutive, but smaller companies actually hold an advantage today–if they are digital natives. If they run their own IT systems, then it’s up to them to make data accessible, which can be difficult. But if they run their operations in the cloud, then their data is already accessible from AWS S3 buckets or Microsoft ADLS.


In some cases, this gives SMBs an advantage over much larger enterprises that are storing data in multiple repositories, sometimes dozens of different far-flung file systems and databases.

What’s more, small companies that utilize SaaS applications can leverage the fact that these SaaS providers may already have used their customers’ collective data to build compelling AI capabilities into their applications, giving these customers a FANG-like advantage without having to hire a small army of data engineers and data scientists.

Ryohei Fujimaki, the founder and CEO of dotData, says he is increasingly seeing small and midsize enterprises (SMEs) finding success with his product, which provides AutoML capabilities with a heavy focus on feature engineering.

“Previously, building models, building features, was extremely difficult” and typically required a data scientist. “But today, particularly for SMEs, this type of automation tool can help those aspects a lot…Our type of automation is definitely helping them to ramp up the speed of their AI journey.”

Fujimaki noted how one of dotData’s smaller customers was able to build AI solutions without a huge investment. The company,, develops a subscription management service that is provided to other businesses as a SaaS offering. It wanted to add a predictive capability to identify payments that were likely to fail.

“For them, the biggest barrier was…skill,” Fujimaki tells Datanami. “They are a cloud-native company, so the data is stored in AWS. On the AI side, they didn’t have data scientists, so they needed AutoML functionality.”’s product manager was able to use dotData to comb through their data and identify the right features that would go into the predictive model. Even though he didn’t posses preexisting talents in data science, the pilot was a success, and’s leadership recognized the value that it brought.

“The most important skill that [customers] have to have is the input side and output side,” Fujimaki says. “The input side means that…the data must be ready for an AI product. Without data access, there is no AI product.” The second critical element that must be present is having a defined output, or a result from the AI that will impact the business.

Even if a SME has developed a great machine learning model, if they do not connect it back to the business, the project will be a failure, Fujimaki says. “That’s why they first need to make sure they have the data and the business outcome,” he says.

The FAANGs (plus Microsoft) have developed a solid lead in AI that may never be exceeded. But the rising AI tide is lifting all boats, from the biggest mega-yachts to the smallest skiffs. Thanks to advances in machine learning software and cloud computing, the pieces are now there for even the smallest player to bring AI to bear on their businesses.

Related Items:

Exposing AI’s 1% Problem

Make Your Own AI

Three Tricks to Amplify Small Data for Deep Learning