Building an AI platform that prevents a CTO's worst nightmare
Human intelligence is an amazing attribute and unique among all other life forms. We discover a problem, solve it skillfully, then build a machine to take over the problem-solving while we move on to discover and solve new problems. An example of this process that has always fascinated me, and is often not given enough credit is the invention of the automated telephone exchange which replaced the manual switchboards. This one simple innovation was key to introducing the data revolution to the world at large.
A lot of skill was required to operate switchboards efficiently, and if not for this discovery, a large number of people would be wasting their lives doing repetitive and unproductive work, and communication would still be costly and only affordable by a few. Instead, today we have millions of devices that do this for us while we have moved on to build a digital world via the internet where we can get any data from anywhere in milliseconds at almost no cost. The folks who might have been working at these switchboard exchanges were among those who went on to become entrepreneurs and engineers who ushered in this revolution.
In this digital world, all enterprises decisively rely on data services to run their businesses.
The modern IT infrastructure on which these data services run is often complex, built with hundreds of different technologies, running on thousands of servers, all glued together with millions of lines of code. These gigantic systems also need to be extremely agile to keep up with growth and to stay ahead of the competition.
Everyday thousands of lines of new code are written, new components are added and from time to time changes in configuration and architecture are made. Each day, every change/addition made has rippling consequences over time. Mistakes are made and things fail. With time they pile up, eventually resulting in a catastrophic breakdown of the entire system. That is when it becomes very difficult to trace the cause of the current breakdown within a very long list of past changes. All CTOs have been through intense firefighting sessions where a key service is down and they had to spent hours or even days figuring out what to fix. To help us diagnose the problem, today we have several superb monitoring tools which collect the health and performance data of each and every component of the infrastructure and application. The huge volume of data generated by these tools still need to be painfully analysed by the tech team to understand what happened.
ArchSaber is one such intelligence platform which performs diagnosis in real time by continuously analysing the huge amount of performance data, and even makes instant recommendations on what can be done to solve the current situation, cutting down several hours of downtime and efforts of your top engineering talents. With every diagnosis it can better understand your problems and forewarn you about incidents that seem imminent, so you can prevent them from happening all together. Read more about what they do here.
ArchSaber was founded in June 2016 by Apoorv Garg, Arpit Jain and Ashish, good friends and batchmates from IIT Delhi’s Computer Science class of 2009. At the time the idea was seeded, Apoorv was working in Google (Bangalore), Arpit in AlphaGrep (Mumbai) and Ashish in Zomato (Gurgaon).
The entrepreneurial journey
India has a reputation of being a rich ecosystem for consumer product startups, and not so much for deep tech B2B startups. But we found this to be inaccurate. When we started up in 2016, all three co-founders had just two years of industry experience. Nevertheless, we were able to get the right guidance and support as and when needed.
One of the major milestones in our journey was getting selected among the final six from over 450 startups that applied to the NetApp Excellerator. We learnt a lot from the programme, but especially remember the takeaways from a two-day bootcamp held by veteran startup coach Nick O'Connor (Director of Alchemist X). Over the two days we learnt what it takes for a three-person team to win the race in comparison to an enterprise with 50 or more employees. In short, it meant:
- Double the hard work: 14-hour workdays instead of eight hours, smaller weekends and fewer holidays
- Triple the intensity: Focus, focus and even more focus. We learnt to identify a valuablebut easy to solve use case within the larger problem and solve it at any cost.
- Quadruple the smartness: It’s not your intelligence but the wisdom of not makingthe obvious wrong decision with limited resources and time that you have. You either have it or borrow it from your mentors but you must develop the knack of getting things
- right in the first go. Hack as much as possible to get things done quickly and easily.
One big eye-opener for us was the realisation that product adoption is a more difficult problem to solve than making the product itself. Even after doing research on making the product and getting a market fit for it, you still have to surpass this final hurdle.
The reason it is so difficult is that even when your product is solving a real problem and has a large number of target users, a majority of these users will be trend adopters and not trend setters. Identifying the trend setters or early adopters in the vast sea of your target users and finding channels to reaching out to them can be really challenging.
In B2B, finding early adopters is even more challenging as the margin for risk-taking is much less and so is the access to the decision-makers of enterprises. Adoption is done by visionaries who can foresee the value in an incomplete solution and have the potential to help you find product opportunities within the enterprise and refer you further. What follows is recognition and trust from the industry. We found the following traits that distinguish these B2B early adopters from others:
- Mostly founders or key stakeholders, who are actively seeking any competitive edge they can derive from upcoming technology.
- Interested in the longer roadmap or vision of your product and discover new use cases that it can solve for them
- Willing to use a product that isn’t complete
This is where industry connects and mentorship that programmes like the NetApp Excellarator offer can be very crucial to achieving early adoption as well as accelerating the growth stage.
Disclaimer: The views and opinions expressed in this article are those of the author and do not necessarily reflect the views of YourStory.