Making Artificial Intelligence Easy to Use
Whenever you speak with recruiters, they’ll usually start going on about the “war for talent” which is just an attempt at making their job sound a lot more difficult than it actually is. There are always people looking to hire and people looking to get hired. Eventually, we’ll just lose the middleman entirely. How hard is it really to browse LinkedIn and message people? That’s why artificial intelligence (AI) algorithms are already starting to “free up recruiters to focus on more value-added activities“. Of course, there will always be roles that are in short demand, and that’s where a good recruiter could be worth their weight in gold. We need “talent” to help design all these AI algorithms that will supplant up to 80% of all service jobs, and said talent typically comes in the form of people who majored in STEM as opposed to underwater basket weaving. The problem is, there are just too few of them at the moment.
While we can work on the pipeline problem by encouraging more high school kids to focus on majors that require a great deal of hard work, we can also take another approach. We can try to dumb down artificial intelligence so much that even an MBA could use it. Let’s look at a few startups making artificial intelligence easy to use.
Founded in 2011, Silicon Valley startup H2O.ai first came across our radar when we published A List of 15 Free AI Software Programs to Download earlier this year. Since then, the firm has taken in a sizable Series C funding round of $40 million from the likes of Wells Fargo and Nvidia which just closed a few weeks ago. This brings total funding to around $73.6 million which is being used to build an “open source machine learning platform that makes it easy to build smart applications“. If their platform is so easy to use, then we would expect to see lots of people using it, which happens to be the case:
The way this open-source tool works is that it creates many variations of artificial intelligence algorithms to solve a problem. Then, it uses artificial intelligence to evaluate which algorithms are most effective. It’s “AI to do AI” which makes us think of Gamalon’s software that we talked about in our article on Computers that Code Themselves Using AI. So what are over 129,000 users doing with this product? Here are just a few examples:
- A startup called Ledger Investing is using the H2O.ai platform to build an entire business around securitizing insurance risk which they claim will be a new trillion-dollar asset class
- Cisco has 60,000 models which it uses to predict purchasing decisions and uses H2O.ai to achieve three to seven times better results
- Paypal is using it to identify hard-to-detect fraud techniques like “collusion fraud” where both buyer and seller collude to commit fraud
- The world’s largest provider of television ads, Nielsen, uses the tool to show you commercials that you may actually watch, and that can demonstrate a positive ROI
- Auto insurance giant Progressive uses it to analyze 14 billion miles worth of driving data from their customers for reducing “time to insights” and predict things faster
While the old-school approach was to have data analysts look backwards to obtain insights from big data, H2O.ai uses big data to look forward, something we talked about in our article on 10 Data Science and Predictive Analytics Startups. Speaking of which, that’s where we first came across our next startup that wants to make AI accessible to just about anyone.
Update 08/20/19: H2O.ai has raised $72.5 million in Series D funding to accelerate the company’s global sales, R&D, and marketing efforts. This brings the company’s total funding to $146.1 million to date.
Founded in 2012, Boston startup DataRobot has taken in a whopping $124.6 million in funding from investors that include New Enterprise Associates and Intel to build an “automated machine learning platform” that will “address the critical shortage of data scientists by changing the speed and economics of predictive analytics“. It’s all about finding the best predictive model for your data, and you can compare models from all sorts of providers – even H2O.ai that we just talked about. Here’s a simple breakdown of the process:
According to a recent article by MIT Technology review, the company has more than 100 customers, though “the concept is a lot less popular with many data scientists, who either feel that their skills cannot be automated or worry that they will be“. That’s because you feed the platform with raw data which it will cleanse and subject to a multitude of algorithms before choosing the best one using – you guessed it – more AI algorithms. DataRobot also seems willing to use their war chest of cash to acquire other players in this space. Earlier this year, they acquired a company we talked about before, Nutonian, which claims to extract insights at a rate of “billions per second” and presently serves clients like Alcoa, Amazon, NASA, and BP among others. In one example, Virgin Australia was able to cut down the time it takes to build predictive models by up to 90 percent, while at the same time boosting accuracy by up to 15 percent. This brings us to our third startup.
Update 08/20/19: For more information on DataRobot, please see our article titled “Automated Machine Learning from DataRobot.”
Update 09/17/19: DataRobot has raised $206 million in Series E funding to continue building out its product line while looking for acquisition opportunities where it makes sense. This brings the company’s total funding to $430.6 million to date.
Founded in 2012, Silicon Valley startup ThoughtSpot has raised $150.7 million to build a platform that “anyone can use to build reports and dashboards in seconds“. The tool can connect with any big data desktop data source and deploys 85 percent faster than legacy technologies. Their ambitious goal is to have 20 million users by 2020, and they’ve already brought on some big-name clients like Capital One, Chevron, and Bed Bath & Beyond. One way to describe the tool is “Google for your business”, and it uses machine learning for understanding the user’s motivations and predicting what information they might find most useful depending on their role. Check out the following example:
In the above screenshot, the tool is analyzing 2.5 billion rows of data to look at iPhone 7 sales across the State of California. In 5.36 seconds the algorithms asked 125 questions about 10 different measures and 50 attributes. The first insight shows us that first-quarter sales for the 25-34 age group were 34% higher in two particular counties when compared to all other counties in the State. We can see how answers are being relayed using natural language in much the same way it lets you ask questions using natural language.
According to an article by Forbes last month, mortgage lender Primary Capital Mortgage used ThoughtSpot to cut the time to get loan status information from 120 to 180 minutes to 2 to 5 minutes a day, saving 600 hours a month in labor. The steep price tag to access the software platform – on average about $210,000 – is easily made up for by the fact that it has saved an estimated 5 million hours of data analyst time or about $250 million in labor costs.
When we look at startups like these three, we can easily see a future where all companies will be using artificial intelligence without even a mention, much in the same way that firms don’t talk about how they use electricity. It’s going to be a given that every firm that wants to compete uses AI. We’ve seen examples before of just how much money AI can save large firms, whether that’s through optimizing the flow of calls in a call center or creating those cool futuristic nanobots we’ve been waiting for. All of these efficiencies go directly to the bottom line which bodes well for future earnings calls, and hints at a future where just about everyone will be using AI to add value in every industry.
Despite what the pundits say, FAANG stocks (Facebook, Apple, Amazon, Netflix, Google) don't give you real exposure to AI. Read about 7 stocks that give you true pure-play exposure to AI in our guide to investing in AI healthcare companies, freely available to Nanalyze subscribers.