Comparing Chatbots To Mobile Apps

Taking A Look At The Effect These Two Mediums Are Having On Each Other

Cobus Greyling
6 min readApr 2, 2020



Apps did not supersede websites, similarly chatbots won’t supersede apps. But they will change your behavior…

Example Of Conversational Digression

I want to take a look at how chatbots will change our practices and more specifically how other mediums will be impacted.

These mediums include not only apps, but also the web.

But First, What Constitutes A Chatbot…

You could argue a chatbot is made up of two elements; A Conversational UI enabled by Conversational AI.

It must be noted that there are Conversational User Interfaces with not much Conversational AI in the back, which does not make for a very human-like conversation.

There are common human conversational elements which should be part of a chatbot conversation. These include digression (like the example here), detection of intent, information gathering (entities), disambiguation and irrelevance detection.


During the initial stages of chatbot development cycle, rapid progress is made with seemingly low effort. This is due to the simple reason that there is no graphic user interface (GUI). With mobile applications there is a much larger barrier to entry, or initial barriers to progress.

With a mobile application most of the work is upfront with much lower subsequent maintenance and updates.

In turn, with chatbots most of the work only start after the launch; getting an initial viable conversation interface (MVP) out into the market is seemingly low effort.

Chatbot & Mobile App Development Effort

This is not wrong of in and of itself, but is worth taking note that with chatbot development there is a false summit, and the real work lies ahead in most cases.

Chatbots Need An Ecosystem

The conversational interface / chatbot needs to have an environment to live in. It needs to be hosted in an environment, and an App is only one of these ecosystems.


The Real Power of Chatobot Lies With Two Principles:

  • User data is entered unstructured. The chatbot structures the data for processing after user input. Subsequently output data is again unstructured for conversational-like output.
  • Chatbots are versatile; being able to exist in a host of environments. Especially suited for conversational applications like WhatsApp, RCS etc. Hence mobile applications can be augmented with a conversational interface.

In this sense a chatbot allows for a company to surface their products and services in more environments, adding a conversational component to a mobile app.

Required Investment

A chabot can be built and integrated to current conversational environments (Messenger, WhatsApp, SMS/Text etc.) at very low initial cost. There is also not the impediment of approval processes like apps (there are exceptions).

Startups see a chatbot as a quick win for customer engagement with little to no cost impediments; apart from time.

The use of messaging apps and time spent by users within those environments make for ease of audience access and quick personalization.


There is a sense of app fatigue with users. This, paired with gaining critical mass with apps can be hard due to market saturation, make chatbots attractive.

According to Statista:

While apps appear to be slowly falling out of favor with startups, other trends are emerging. Looking at the terms with the largest increase in mentions in startup company descriptions between 2010 and 2016 possibly allows us to glimpse into the future and gives us an idea of what the next big thing might be. According to CB Insights’ analysis, these terms are “virtual reality”, “machine learning” and “natural language”.

Is the App Economy Past Its Peak?

Getting a chatbot going is lesser technical endeavor than a mobile app, it is more a case of experience over time.

Chatbot Positives

  • Single Interface Living in Different Environments. A single chatbot can exist in different ecosystems, hence giving users access via different avenues to the same interface.
  • Rapid Availability. In order to have an interface up-and-running is much faster than a mobile app. Once the interface is released in the wild, it can be improved incrementally.
  • Lower Development Cost. Development cost is lower, hence making it more accessible. Opensource and no-cost tools are in abundance and growing.
  • NLP / NLU Advancements. Natural Language Processing and Understanding as disciplines are rapidly expanding.
  • Footprint. Chatbots can have a very small footprint with no local installation requirements.
Spectrum of Modalities

Chatbot Negatives

  • User Management. With chatbots it is harder to mange the user’s behavior in terms of input. Users have the freedom to input anything and the onus is on the chatbot to disambiguate & detect irrelevance.
  • Technology Maturity. Technology is expanding in terms of what is possible. In many instances chatbots have problems scaling due to initial technology choices.
  • Notifications work better on apps than conversational environments. Notifications which can be set according to time/day, frequency and subject.
  • Shopping still has less friction on apps, but conversational commerce is on the rise.
Messaging App Lead the Way

Chatbot Complimenting Apps

  • Customer Service is well suited to be served within a mobile app via a conversational interface. Many apps offer live agent chat, and the chatbot functionality can serve an Interactive Text Response (ITR), analogous to IVR in a call center.
  • Conversation Rates can increase on mobile apps. Historically it is hard to get user to converse on apps, as oppose to devices with large surfaces. Chatbots can up the user engagement on apps.
  • Payments can be facilitated within the safe confines of the mobile apps.
  • Automation of task sequences on behalf of agents can be leveraged. The chatbot can, by detecting a single user intent, action all relevant tasks automatically. Acting as a Robotic Process Automation (RPA)element.
Most popular global mobile messenger apps as of October 2019, based on number of monthly active users (in millions)


Chatbots must be seen as an additional medium, and a very new one at that, through which a message can be delivered to a user or customer.

The Medium is the Message.

~Marshall McLuhan

In addition, a conversational interface should be seen as a medium which will eventually lead to interfaces or services being ubiquitous. Living in a user’s environment. What is often referred to as Living Services.



Cobus Greyling

I explore and write about all things at the intersection of AI & language; LLMs/NLP/NLU, Chat/Voicebots, CCAI.