Skip to content

Everybody’s an Engineer Now

Chatbots are shifting tech from screens to natural language, favoring startups and challenging traditional firms.

4 min read
Everybody’s an Engineer Now
engineers, builders, architects, construction site, construction, building, architecture, construction worker, infrastructure, manual labor, hard hats, reflective vests
Before Growth is a weekly column about startups and their builders prior to product–market fit.
Originally published in Mam Startup.

Tracking what’s happening across the ocean, it’s hard not to notice the craze associated with chatbots, i.e., applications communicating with their users through conversations, not screen interfaces. Their numbers are growing. And likely, this market will expand even more in the coming years.

Products like Apple’s Siri, Google Now, Amazon Alexa, or Microsoft’s Cortana have been available on the market for some time now. However, about six months ago, we saw a significant leap in this technology. Every major brand now needs to release its own chatbot. Startups are following the example of bigger players, naming their bots personal assistants and attempting to automate more service sectors with their help.

Yet other companies—like Facebook with its Messenger Platform—are going meta, building platforms for building chatbots. The last time it was this heated was in 2008, when the App Store for iOS was created.

It seems we’re witnessing the birth of an entirely new ecosystem that could be with us for years to come. Is this justified? One of the goals of my article is to convince readers that we indeed are dealing with something significant. How do I know this? While creating Ada, a personal assistant that helps people rent apartments, I invested a huge part of the last year into chatbots. So, I have access to information directly from the front lines.

Technically speaking, a chatbot is a dialogue interface powered by algorithms from the field of artificial intelligence, capable of understanding text input by the user in natural language—like Polish—and proficiently responding. More practically speaking, interactions between a user and a chatbot, instead of taking place on a screen, are conducted simply through conversations.

Even in this short definition, two characteristics appear which set the tone for the ongoing revolution. The first is the reshaping of the user interaction mode with the application around entering text commands, instead of clicking on the screen. The second is basing this mode of interaction on natural language, which non-technical individuals can handle, instead of predefined technical commands like “!help” or “brew install ruby,” which may be familiar to those who’ve ever been told to do something in the console.

Let’s consider a simple example based on real estate. The most important function of a personal assistant in the rental market is searching for a new apartment. Therefore, the assistant takes an order from the user. An order might be a message like: “I’m looking for a studio apartment for 2000 zł near the main station—but if you find something cheaper, it could be somewhere else.”

Those who consider chatbots as a passing trend fail to recognize that the above order is fundamentally a computer program dictated by someone who doesn’t necessarily know how to program. The capabilities and functionality of such a mini-program are dictated by the grammar and meaning of the dictated sentence, which are not limited by a screen interface defined in advance by programmers and designers. Of course, how the command is executed depends on the quality of the chatbot, but the paradigm shift is fundamental.

Until now, only programmers were able to write a program that browses listings and accepts or rejects apartments not close to the station, if they meet an additional, secondary criterion. Now, anyone can do it. If you think this isn’t a big deal, allow me to add another piece to the puzzle.

The Holy Grail for most startups is to find and tame some disruptive innovation. Such innovations were perhaps most fully described in Clayton M. Christensen’s book “The Innovator’s Dilemma”. Christensen lists among the most important features characterizing disruptive innovations a greater flexibility of a given technology, whose costs are initially weaker performance and paradoxically higher implementation costs.

All of these features also describe chatbots. So far, they are suitable for creating a narrow set of specific solutions. It’s hard to imagine a mature company like Uber, whose product is entirely based on conversations using natural language. Additionally, the maturity of mobile technology markets, as well as easy access to tools and programmers, means that for now it’s cheaper to build a regular mobile application. However, it’s worth noting that these are disadvantages that will fade in accordance with Moore’s law as more and more companies invest more resources in better chatbots. What will not change is the universality and flexibility of such a mode of interaction—the fact that the possibilities of chatbot mini-programs are limited only by user creativity and what building blocks for queries the product creators give them.

If the laws described by Christensen hold true this time, the same features will lead conservative companies to be unable to respond to startups, which will early adopt the new paradigm. They will build knowledge resources about AI, learn to build products better and cheaper under these new conditions, and also hire the right people and organize their processes so that these individuals can better utilize new technologies.

Taking all of this into account, the question about chatbots for me is quite simple: do you want to be at the forefront of this new wave, or defend against its onslaught? I already know the answer.

Thanks for being a part of Before Growth! The newsletter has grown mainly through word of mouth. Would you mind sharing it with a couple of friends who might like it too?

Kamil Nicieja

I guess you can call me an optimist. I build products for fun—I’m a startup founder, author, and software engineer who’s worked with tech companies around the world.

Related posts

Chaining Prompts

Chains go beyond a single LLM call and involve sequences of calls.

Chaining Prompts

Intelligence as an API

AI models are getting hard to beat when it comes to getting simple answers right at scale.

Intelligence as an API

Pairing with AI

Raise your floor with large language models.

Pairing with AI