It’s about time Machines did all the talking

The most common way you’ll find Indian businesses scaling is by ‘throwing people at the problem’. “Labour is cheap” is the most common reason you’ll hear for it. However, this is not going to always be true. We can’t build systems and companies with mindsets stuck in the past and hope to somehow miraculously usher in the future. That future we’re all building towards will take India to a stage where labour is no longer cheap, and any system built with this assumption as intrinsic to its business, will struggle to survive.

Two things are true about machines, or technology, in general: Technology and machines are going to get cheaper as they turn into commodities, and machines are going to get faster. Neither of the above is true of people. It’s about time we recognized this fact, especially those entrusted with taking India into the next stage, started acting and building as such.

There are certain things machines do well, and we need to be comfortable with taking a step back and letting machines do those things, even if it means making us redundant as far as certain aspects of our day to day work life are concerned. That’s not to say we will no longer be needed; it means we’ll get pushed further down the funnel where our time, talent and intuition are better utilized. Today there are large parts of workflows being done by people that can easily be optimized and automated. And this reality is conspicuously apparent in the financial services industry.

When machines talk to machines, we imply the various processes involved in the exchange of data and information between them. Today exchange of information amongst entities is grossly outdated, archaic, even, in some respects. It happens over phone calls, over emails, in spreadsheets, even scanned documents. This compromises the quality of, and the trust in, the data. It requires human eyes to record, read and cross check data before any of it can be acted upon. This wastes valuable time and eats into productivity. In an ideal world, machines do all the talking. Data is sent from a single source of truth to all entities seeking to consume that data. We now need only to trust the source of data, and not the (potentially fallible) intermediaries through which that data has travelled to reach us.

The use of APIs to exchange data is commonplace in the tech world. It’s quick, it’s clean, it’s ready-to-eat. It allows us to reach levels of scale, speed and accuracy that can’t be achieved should human beings be involved in the process. Data and information can be exchanged and crunched at lightning fast rates, the results of which can then be displayed to the human being entrusted with acting upon those results as far down the funnel as possible. The only scalable way of growing a platform is to have API-level integrations that all co-parties agree to use. One way to think about it is, if a process involves 1 human being and takes 1 day, the only way that process can scale to a 1000 a day is by hiring a 1000 people. Only machines can give exponential, super-linear growth.

The Fintech space is one sector that is swiftly adopting the concept of machine learning. Businesses are underwritten; 3^rd parties like bureaus and data providers are consulted; data is crunched by machine-driven algorithms; loan agreements and other regulatory formalities are completed; and all of this over APIs. No human being, other than the business owner, participates in any part of this process. Needless to say, this would’ve been impossible without machines talking to machines.

To be clear, I’m not taking the stand that all processes should be machine-only processes. That’s debatable, and will have to be dealt with on a case-by-case basis. A good analogy to use is that of Iron Man (or Robocop for those of us that are a little older). Jarvis does all the data collection and crunching for Tony Stark, and it’s finally the man-in-the-machine that decides what to do about it. I don’t intend to usher in a future of run-away bots. However, I do unequivocally believe that all data exchange needs to happen using APIs. It’s about time machines did all the talking.

(The article has been authored by Rohan Angrish, Chief Technology Officer, Capital Float)

Image Source