In this special guest feature, Thameem Khan, VP of Product Strategy responsible for Platform vision at Boomi, has been watching a new wave of AI companies emerge that focus on providing value to the enterprise with predictions based on business goals. Any other purpose, he argues, turns AI into an expensive science project. Thameem is very passionate about predictions using data and loves connecting the dots and clouds.
AI for business has gone from being a theoretical science project to become a critical competitive advantage for forward-looking companies. A growing number of organizations are turning to AI to help them transform data into intelligence that delivers sustained business impact. But some companies have yet to recognize the power of today’s rapidly maturing AI and risk being left behind. And some AI still doesn’t deliver what organizations want most: business impact.
The Early Days of AI
I have been following the AI space for some years now, and like many of you reading this, my introduction to AI was through the famous Andrew Ng’s ML course on Coursera. After the course, I got hooked on and started getting deeper into AI. It was becoming apparent that the main reason for AI to gain mainstream attention was the data and computing that did not exist prior. As soon as I started exploring real-world problems, I realized that AI did not have the most sophisticated stack and required knowledge of many technologies, languages, and frameworks. I suddenly found myself lost in trying to understand these technologies and not solving the problem that I had intended to solve. That is when I realized that I needed to take a break and give the AI space some time to mature.
AI Has Matured Quickly
In the last couple of months, I started looking into the AI space more profoundly and was amazed at the progress that the AI community has made. As with any technology, once the adoption of tech reaches a critical mass, you start seeing great minds come together to start building tools that abstract the complexities. These tools help the end-user concentrate on solving the problem rather than worrying about the nitty-gritty. I believe AI has reached this stage. The AI stack has matured to a point where all you need today is data, and the frameworks will generate models and sort them based on accuracy for you to choose. And all of this without writing a single line of code.
AI’s Critical Mission: Delivering ROI
The next real question we must answer is, “How is AI helping companies?” AI was blown out of proportion because of all the hype, and FOMO kicked in with enterprises. Suddenly, we saw an influx of investment with new AI teams spinning up and huge demand for data scientists and engineers. Many enterprises have AI teams working on real data and building models for accuracy without thinking about or even knowing the end goal. We forgot to ask an essential question in between all this confusion. To quote Aible’s Arijit Sengupta: “Is AI delivering ROI?” If it’s not, then it’s just an interesting solution to a problem nobody cares about.
The Power of Collaboration
We can better explore the real power of AI when we involve the business users, set goals, and maintain a constant feedback loop. In most enterprises, a data scientist, who may or may not have a good understanding of the business, makes the models and decisions. In many cases, AI may not even be the solution that is needed. Many implementations also suffer from stale models that do not evolve with changes in the data. Just like humans, AI needs to evolve its predictions with the ever-changing business. The success of AI is in providing value to the enterprise by providing predictions based on business goals. And enterprises should have a clear plan of action to implement these predictions. Only then can AI provide ROI, or else it becomes an expensive science project that your company is funding.
AI is Disrupting Traditional Business Models
AI is changing and will continue to change how we are and will be doing business. Startups collect a plethora of data and use AI to learn more about their users, then recommending or matching them with a tailored, personalized experience for the end-user. An excellent example of a successful AI company is Lemonade. On the surface, it is another insurance company, but in reality, it uses AI to automate many of its processes. Their use of AI technology has helped them achieve close to $100 million in revenue in four years and grow to 1,700 customers per employee vs. competitors who have 150 to 450 customers per employee.
AI’s Next Frontier: Mainstream Impact
So, has the AI tech stack/tools evolved enough? Has the AI hype died down, and is it time to think about AI in terms of ROI? Should enterprises start thinking about how to use AI to build a bigger moat for the next decade? In my opinion, the answer to all the above questions is Yes.
We have companies like YData and Superb AI providing data labeling, Pachyderm providing data lineage, Tecton with feature stores, and CI/CD with Seldon and Algorithmia making up the tech stack. On the other hand, companies like Aible are shaping AI thinking in terms of ROI and are providing a hosted turnkey solution. And companies like Lemonade, Affirm, UpStart, Fiverr, and Pinterest are already proving that AI gives them the edge to be disrupters in their space.
It’s time enterprises start leveraging AI as a mainstream technology rather than a science project in the basement with mad scientists running the show.
Sign up for the free insideBIGDATA newsletter.
Join us on Twitter: @InsideBigData1 – https://twitter.com/InsideBigData1