Founded in 2003, Splunk is a leading provider of software to investigate, monitor and analyze data. The initial focus of the company was on security, which proved to be spot-on. But since 2015, there has been a transformation. Splunk has pulled off 13 acquisitions (with the latest deals being for Streamlio, SignalFx and Omnition) and the revenues have jumped by 300%. Along the way, Splunk has moved into new categories like app development and IoT (Internet-of-Things).
So to learn more about the company’s success, I had a chance to meet up with the CEO, Doug Merritt. We talked about a wide range of topics like the cloud, the importance data and the company strategy.
But there was something he said that caught me off guard: “AI does not exist today.”
I paused for a bit and then I thought: Isn’t AI one of the most important trends? Aren’t companies like Google, Microsoft and Faceook investing huge amounts in this technology?
Of course all this is true.
But then again, Merritt’s statement does point out that the concept of AI is slippery. The inherent problem is that it encompasses many types of technologies like ML (Machine Learning), Deep Learning, Natural Language Processing (NLP) and so on. Yet all of these are narrow forms of AI and really do not work with each other.
But the original vision of AI–which goes back to the 1950s–is about systems that can truly learn about anything, across any domain. It’s what we see in movies where a machine is indistinguishable from a human.
But the reality is that we are no where near this. “It could be 50 to a 100 years to get to AI,” said Merritt. “There are many issues and challenges to work out, such as with computational power and energy. Keep in mind that the human brain only uses 50 watts a day. It’s also a very complex distributed system that has a high filter for intuition.”
If Not AI, Then What?
For now, Splunk is focused on ML when it comes to making its technology smarter. This is actually one of the oldest forms of AI that leverages traditional concepts of statistics and probabilities (like Bayes Theorem).
Note that at this week’s annual conference, Splunk announced plenty of new offerings that use ML:
- Splunk Natural Language Platform (NLP): This makes it possible to perform analytics with verbal requests, such as by using a mobile device. The responses can then be used in a dashboard or saved in a search result.
- Splunk VictorOps: It uses ML to provide for intelligent alerts that are routed to the right teams, so as to resolve issues faster.
- Splunk’s IT Solutions: This platform leverages ML to better take advantage of an organization’s data.
- Machine Learning Toolkit (5.0): The software makes it easy to create custom visualizations and assistants using ML.
“Existing ML tools are static and can be difficult to use,” said Tim Tully, who is the Chief Technology Officer of Splunk. “As for our software, the goal is to make it so you do not have to be a data scientist. You can simply drag-and-drop ML, such as for sentiment analysis using Twitter or anomaly detection.”
What To Do?
For managers looking at digital transformation–which often involves ML–the process can be overwhelming. This is why Merritt advises a methodical approach.
“A good place to start is where your business already has some digital capabilities,” he said. “It doesn’t matter if it is marketing or sales. What you want is an area where you will not have gaps with data. Then look at possible use cases. You don’t want to solve too much too soon, say with a data lake.”
Then, once you have some early successes, you can get bolder. And eventually, the focus can be on sophisticated capabilities like ML. “Amazon did not start using ML until 2005 or so,” said Merritt. “The first ten years was about building the core infrastructure for e-commerce.”
But the bottom line is that companies need to start now on leveraging data and digital technologies. “It’s a major risk to disrupt your business,” said Merritt. “But the odds are definitely much better than doing nothing.”
Tom (@ttaulli) is the author of the book, Artificial Intelligence Basics: A Non-Technical Introduction.
read more at http://www.forbes.com/entrepreneurs/ by Tom Taulli, Contributor