AI Conversational Components Q&A: A Fullstack Developer’s Perspective

Published by:

Eran Soroka

We sat down for a chat with Alex Tavor, CoCoHub’s stellar fullstack developer, to gain a deeper perspective on conversational components (CoCos)

Q: Alex, tell us: What exactly is a conversational component?

“It is an isolated, reusable chunk of conversation with a chatbot, wrapped in a standardized API.”

Q: Who benefits the most out of adopting development using conversational components? Who is a CoCoHub “developer” or “vendor”?

“CoCoHub developers would be people who develop a lot of chatbots, in particular those who have flexible production procedures. Studios working with varied customers. In short, anyone who’s developing a lot of bots, and has a lot of different requirements for bots’ features.

A vendor could be someone who has chunks of conversation they want to reuse, and/or profit from. Developer – anyone who’s making or maintaining a chatbot.”

“Take a chunk of needed functionality
made by someone else, and integrate it effortlessly
into your code.”

Q: Sitting down with a developer who has absolutely nothing to do with conversational components and is working “old-school”, how would you explain CoCoHub to him/her?

“Conversational components are to chatbots what package managers are to code. They let you take a chunk of needed functionality made by someone else, and integrate it effortlessly into your code.”

Q: How would you explain the benefits of conversational components to a potential “vendor”?

“To name a few, it allows someone who’s made a bot to monetize parts of it, on top of whatever profits the original bot made – it transforms old code into capital. If you’re a studio making bots, it allows you to show off your bot-making prowess to potential customers and provides exposure. If you’re making a lot of bots, it allows you to organize your work in a way that’s easier to reuse in new bots.”

“Simplifying adding / extending
functionalities of bots”

Q: How do you see the first steps of adoption of this method of
developing chatbots?

“Probably as part of a larger framework, like Dialogflow. CoCoHub components can be effortlessly integrated into almost any workflow, and using conversational components simplifies adding / extending functionalities of bots. So first steps would be a developer saying “Ooh, what’s that, a well-tested way to get a user’s name that I can add with a single line of code? Sweet, let’s use that rather than spend the next week writing it myself”.”

Q: Do you see objections to this method of development?
What will naysayers have to say?

“By using components someone else made, you’re relinquishing some control. What if you utilize conversational components everywhere, and then suddenly CoCoHub goes under? What if A-I’s having server trouble? What if there’s uncustomizable content that you don’t see right now, and it’s bad / broken / unsuitable? What if a vendor suddenly cancels a component you’re using, or changes the price radically? Using someone else’s work, especially work that’s online and live, adds another technological and content failure point to your system.

It requires your customers to accept you’re using another piece of third-party functionality in what they buy. What if the vendor is malicious, or dishonest? Using third-party functionality that every conversation passes through opens the door to all kinds of potential trouble the third-party functionality provider can cause. A namer component provided by a dishonest vendor can compromise the privacy of your clients, for instance. In other words, conversational components add more potential trust failure points to your system.

These are the usual risks of using third-party functionalities in your pipeline, though. Vendors stand to gain much more from being good brands than from being sneaky or negligent. And, of course, we approve each vendor and component only after extensive review. With time I believe vendors will specialize in different types of components, become known for being good at specific things, and guard these reputations, much like content providers do in other industries.

“Focused on content, reducing the need for expensive specialists and slow processes.”

Q: Why do you “believe in” conversational components?

“The demand for chatbots will grow dramatically. Conversation is a much more natural interface for a lot of interactions. With chatbots becoming more human-like, using them becomes more attractive to companies. So the demand for this type of interface will grow. Mordor Intelligence seem to agree, giving 34.75 % compound annual growth rate projection for 2020-2025. This growth will increase the demand for comfortable chatbot development tools, as more people will be making more bots. Luckily, this is exactly what we provide.

Cocos provide standardization and an easy way of using content developed on various platforms in the same bot. The continuing competition between the various platforms makes a neutral standard provider like CoCoHub useful to developers. Among other things, it allows developers to leverage the relative strengths of different platforms in a single bot.

Components are a natural next step to take as the industry matures. The same thing happened in other industries – whether physical manufacturing or programming languages, eventually industries converge on components when they become developed enough.

Chatbots are, by now, not a technological problem, they’re a content problem. Cocos allow chatbot developers to focus on content, reducing the need for expensive specialists and slow processes.
This is needed and will be needed more as the industry matures.”