Loom Systems Lands $10 Million in Funding to Predict IT Incidents with AIOps
While Gabby Menachem (CEO) and Dror Mann (VP of Product) where working together at their last startup (a company focused on analyzing social network data), they encountered a problem common to tech-focused departments – not enough manpower to monitor and ensure their business services. The startup used new technologies like microservices and container-based applications to give them the agility, performance, and efficiency they needed to be competitive. But it left one downside, having to monitor and fix technologies that were so new there wasn’t an existing database of problems to expect. In other words, they had to learn what to watch for.
These challenges eventually led Dror and Gabby to connect with an old army friend, Ronny Lehman (who had extensive experience in machine learning and signal processing). Drawing on that experience and their respective backgrounds in Israeli Intelligence, they thought AI and, more specifically, Machine Learning, could be a solution.
Understanding the Work to Find the Solution
Their first step was to observe network engineers spotting, troubleshooting, and resolving IT issues. They noticed a pattern, once an issue occurred the IT staff would dig further into the possible cause eventually leading them to the log files for all of the applications and infrastructure elements touching the impacted service. The staff would then take the information from the log files and search online for possible solutions.
“That’s when the light bulb went off,” said Gabby Menachem. “We saw that log files were the key. If we could use AI to read, understand, and then monitor all log files, we could not only spot the root cause and suggest possible fixes, but predict when an issue was about to occur.”
From Birth to $10 Million in Funding
That light-bulb moment was the birth of Loom Systems. The founders quickly went from idea to product, securing enterprise customers and heavyweight partners like Amdocs, Clal Insurance, and Citrix. The effectiveness of their machine learning engine, dubbed Sophie, and rapid success eventually led to their first funding event in 2017 where they secured $6 million.
“Our success from 2015 to 2017 was despite a market that wasn’t quite ready for AIOps,” commented Gabby. “But the market has seen quite a shift in the past 2 years. Organizations not only grasp what we are doing, but they are seeking it. The fact that we can prove we predict IT incidents before the business is impacted, is resonating.”
A Unique Approach
Loom’s log-first approach differs from that of others in the AIOps space who focus on gathering and correlating events from multiple eventing solutions, then triangulating the root cause. Effectively, they solve a problem after users and customers are impacted. The log-first approach allows Loom to spot slight deviations in normal behavior or ‘tone’ of messages well before thresholds in traditional monitoring tools are reached. Giving Sophie the ability to predict IT issues before customers or users are impacted. Logs also give the granular information necessary to find the exact solution to an IT issue within vendor knowledge sources.
“We pull data from many sources, not just logs, to give us a fuller picture,” said Menachem. “But the logs are our secret weapon because they give us the granular data we need and tell us about issues well before any events are triggered.”
Sergey Gribov of Flint Capital Uses Technical Background to Select a Winner in AIOps
Sergey Gribov uses a unique combination of technical experience, market knowledge, and an understanding of team dynamics to select winning startups. Loom Systems was no exception. Loom addressed a challenge that Sergey was intimately familiar with, having faced it many years before as an IT Operations lead – unpredictable and transient IT infrastructure that leads to outages of critical business services.
“Years ago, I was responsible for the design and operation of my employer’s critical infrastructure system, which grew to more than 500 servers,” explained Gribov. “We were a small startup team, so we designed it from the start with lots of automation built-in. That gave us the ability to automatically address service issues as they happened and, more importantly, before the business was impacted.”
“Loom is doing something similar, but in a much more complex environment. While I had standardized infrastructure, Loom needs to support mixed multi-cloud environments (both public and private) with newer transient microservice architectures,” said Gribov. “Starting with log data gives Loom the ability to predict when issues might occur by tracking small deviations from the norm. Logs also give the granular data needed to detect the root cause of IT issues. Plus, Loom has ability to look at log files from different systems spotting potential problems that can only be found by correlation between different logs – a much harder task for humans than for AI.”
Sergey views technologies just like any other tool – effective for certain jobs and not very effective for others. Problems where the volume of work (in this case mountains of log files) and sheer tedium of the task involved (reading through dry log files) make it not only undesirable, but also impossible for humans to address alone are, from Gribov’s perspective, the perfect application for artificial intelligence.
“ITOps is the perfect space for artificial intelligence to thrive, which is why I view AIOps as a market that will grow exponentially in the years to come,” said Gribov. “And again, because Loom isn’t solely focused on eventing data they can catch potential IT issues well before other eventing focused AIOps solutions.”
But it wasn’t just the technology that excited Gribov about Loom Systems. “After meeting the team, I saw that despite their distributed model (offices lead by each of the 3 founders in different geographies – San Francisco, New York City, and Tel Aviv) they had a unique synergy that allowed them to work as a united team. At Flint Capital, we have a similar distributed approach, and find that if you can get the cultural model and communication right, the value impact can be exponential.”
“In the end, I’m confident that we have invested in the right company in the right space,” commented Gribov. “We expect big things to come from Loom Systems in the years to come and are proud to be a part of their growth.”