Chatbots and Their Impact on How We Interact With Technology

Chatbots and Their Impact on How We Interact With Technology

Technology is advancing at an unbelievable speed, and every few years, there comes a new technology or technical concept that promises to impact important parts of computing in a significantly positive way.

Some of these technologies that came in the past have either truly transformed the way computing works or the way we interact with technology. Others just turned out to be a hype with lots of great promises and little substantial impact.

Chatbot is one of the newest set of technologies that re-introduces the promise of a huge positive impact on our lives by making technology more like us.

Famous words that have been used to brand chatbots and associated technologies are these:

“Bots are new apps.”

Many people, including myself, started watching this new cluster of technology last year and started taking them seriously a few months ago.

Well, you have to take a particular set of technology seriously when the likes of Google, Microsoft, Facebook, Amazon and Apple all say in one voice that “Bots are the future of computing,” and they do this in the same year. I cannot remember another time when all of these tech giants agreed on a common topic before.

This write-up is my effort to share some personal thoughts on an intriguing concept. I like to share my thoughts on bots by putting some context around usage, design considerations, their potential future place in everyday computing and our interaction with technology.

So “bots” – what are they?

For those not totally familiar with the concept, let me provide a quick overview of what “bots” are.

In a simple sense, chatbots (or bots) are a new kind of access and delivery channel. They are considered to be part of what is called “conversational User Interface”, which means they are a kind of user interaction unit that can engage with a user in a conversation, either via an interface where users can chat with the system or an interface that is voice enabled, which means a user can engage with the system in a natural, voice-enabled conversation. Just like how we speak with each other.

And, this is the big deal – the fact that we can engage with a technology system in a more human-like conversation style. This is promised to be an actual game changer.

Of course, there’s no new concept here. Products like Siri, Alexa and Cortana have been doing this for a few years now. In my view, the big deal is that now computing, platforms, talent pool and tools are making it possible for a vast number of people and organizations to build such systems. You don’t have to be Apple, Amazon, Microsoft or Google to build a bot anymore.

There are many different kinds of functional categories that bots fit into. There are those bots that are chat enabled and work within your favorite chat platforms, such as Facebook messenger and Skype.

There are those that are voice enabled and listen to you in your natural voice (even if you have an accent, like I do) and talk to you in a human-like voice.

Further, there are bots that are only informational and provide you information based on a question that you asked. There are also bots that can take actions based on your commands.

We also have bots that are rule driven. Rules are pre-defined and when a particular rule is met, a bot can take an action or can provide you relevant information.

Finally, we have smart bots. These are bots with a personality. These bots are driven by big data, advanced analytics and artificial/augmented intelligence. They learn a lot about you as they interact with you to help you make right decisions. They pro-actively take actions and even talk to other bots if needed.

An important thing to recognize here is given the way bots are designed today, they act as both input and output channels. Most of the processing, computation and analysis is still done on existing compute infrastructure, and that is a great thing because we do not need a new bot-specific compute layer. A single compute system can serve your enterprise applications, websites, mobile applications and bot at the same time.

A use case – Bot for doctors?

Bots – more particularly analytics, machine learning and AI-driven bots – can have a huge number of practical implementations. Value that bots can add in research, clinical trial setups, medical record analysis, recruiting, disease symptom logs, interaction with governments and many other areas can be significant if we use technology carefully.

I like to present as a case study a situation here that I experienced first-hand recently. This is a based on a personal experience.

Last week, I visited my doctor for a regular checkup. I also wanted to get my medical records from her since she does not visit the town I stay in anymore.

So, upon my request, her office gave me a stack of printed sheets that had my medical records and visit logs that she created at the time of my visits. I have been seeing her for the last three years, and I must have seen her five times in this period, yet my visit logs were close to 40-50 pages.

When I read the documentation while waiting for her assistant to call me in, I was taken aback with the amount of details and text that she had put in for each visit. On an average, this was close to 10-12 pages of details about each visit that I had.

When I saw her, I had to ask – When does she type all this information? And does she do this for each of her patients?

She explained that she does this for each patient (so I wasn’t the special one). It is mostly done as soon as a patient leaves, and that is why I wait 20-30 minutes to see her because she had to write these logs for her previous patient and then needed some time to review my data from previous visits.

Now imagine if she had a voice-enabled bot or maybe an Alexa that she could talk to right after I leave and dictate my visit log in her natural voice. She could have activated the bot by using an activation phrase (“Hey, my smart bot, make a log for patient XYZ.”), and the bot starts listening to the information about my visit.

Because these reports are broken down into multiple segments, a conversational style that bots offer will be extremely helpful for such a situation.

Further, right before my doctor gets ready to see me, she can prep by again activating the bot and asking it to tell her about my previous visits.
Using this bot will not only save her a lot of time and effort but will also give her more time to spend with me and will reduce my wait time from 30 minutes to 10 minutes (hopefully).

Moreover, these bots can be used during my visit to take ad-hoc notes, asking for newer treatments (e.g. a bot can search for newer clinical trials for the doctor by way of a voice command), adjusting treatments and even sending prescriptions to pharmacies, all at once during the visit.

One last point before concluding this example. What would happen if, as a patient, I have a counterpart bot at my home where I record my daily status, feelings and symptoms and this data is transported, in a secured way to a system that my doctor’s bot can access?

Now, my doctor does not have to wait for me to come in after three months and ask me how I felt during that time. Doing so anyways gives her only a snapshot of how I felt in the last 24-48 hours because of memory limitations that we all have. It does not provide a multi, data point-driven trajectory.

With the help of these two bots, now she knows how I felt over a period of time. This may give her an opportunity to call me in sooner, suggest me to not show up for a scheduled visit and maybe she could also adjust my treatment plan based on the data that my bot is exchanging with the system that her bot can access.

This is a simple story but hopefully, it puts some perspective around how chatbots and conversational systems can make a real impact on real-world problems.

What’s needed to build impactful bots?

Here are four things to keep in mind when designing bots for various purposes.

1. Start with “Why?”

This probably is the most important thing to consider before exploring bots. You must always start with the question, “Why do I need bots?” or “Why do I think they can make a difference?” instead of asking, “What can I do with bots” or “How do I start using bots?” It is easy to get carried away with the coolness of technology, but keep in mind that even though bots can have a massive positive impact on many situations, they are not a great fit for every situation. There are a huge number of use cases where a visual user interface will leave bots miles behind.

2. Analytics and data

Smart bots require intelligence, and intelligence comes from constant data points, access to historical data, useful algorithms, machine learning and big data. Wouldn’t it be great if my doctor’s bot could tell her on its own that bi-weekly B12 shot that she has prescribed me is not working for me and has also not worked for 78% other patients in her care with a profile similar to me?

3. Strong integration

It is extremely important to ensure that bots can integrate with existing systems and can leverage existing computing infrastructure. My doctor’s bot must be able to integrate with her patient database, discs and other systems that she uses. It should be able to integrate with the software she uses to send electronic prescriptions to my pharmacy. They should also be able to integrate with public APIs and with other bots. Most of the bot architecture available today can do this as long as the target system supports an open and secure integration architecture.

4. Privacy and security

Privacy and security must still be top priorities for designing bots. It becomes more important when we start thinking about using bots in medical, clinical and healthcare settings. How to activate a bot, how to integrate them with other systems, how to have a bot integrate with other bots and how to make sure that a smart bot, when activating itself in an automated fashion, is not giving information to someone who does not or should not have access to this information. All of these security design considerations must be thought of carefully.


So are bots really new apps? Perhaps not yet, but they certainly have their own value preposition, which if used for the right kinds of problems or opportunities, can add far more value than an app in a classic sense. However, for a huge number of other situations, bots may not be as effective as an app could be.

Even though bots are not yet in a position to replace all kinds of user interfaces and apps, aren’t you happy that they are here to help us achieve things that one could only hope for a few years ago?

I think bots demonstrate a huge amount of promise today. Advancements in big data, machine learning and augmented intelligence will one day in the near future get us very close to the original premise of bots – that “technology can be more like human beings.”

Article written by Manoj Vig
Image credit by Getty Images, Moment, Yuichiro Chino
Want more? For Job Seekers | For Employers | For Contributors