Creating a Chatbot Business Case

Getting started with chatbots is easy, but you need to have a business case to make it a long-term success. In this article, I’ll show how to decide on a business case and determine the ROI.

Do I need that bot?

The hype has passed, but it’s still an important question to ask: do I really need a bot?

Like websites and apps, millions have been spent on fixing problems that don’t exist.

You can think of three proven use cases:

  • Customer care
  • Marketing and sales
  • Conversational UIs (for lack of a better name)

Customer care

Expectations for bots in customer care were extremely high, while in practice there is a lot of disappointment.

For customer care, you should consider a safe 10–15% automation success…

You cannot expect a bot to automate 90% of customer interactions.

For customer care, you should consider a safe 10–15% automation success using a bot.

Let me remind you this is for a well-designed, ask-me-anything bot, that doesn’t frustrate customers.

Top Articles on How Businesses are using Bots:

1. Chatbots: What Happened?

2. How Businesses are Winning with Chatbots & Ai

3. The road to a conversational banking future

4. Chatbot Conference in San Francisco

Now image you have 10 agents working on customer care; you’d save 1 FTE (full-time employee) based on full automation.

Is it that simple? I’m afraid not. You still need to factor in the costs of running and maintaining the bot.

This can cost you about 0.5 FTE to maintain and train the AI — at least with a cloud-based solution. If you decide to build your own technical solution this will increase drastically — but more on that later.

…you should see an increase in customer satisfaction of around 5 to 15%.

Another advantage of using bot automation is further decreasing handle time and reducing customer effort. It’s a bit harder to measure but think of the time being saved when a bot does the intake of customers. For example, asking customer details before connecting to a human agent.

I’d call this semi-automation instead of completely resolving conversations automatically.

On the customer effort part, you should see an increase in customer satisfaction of around 5 to 15%. This is particularly higher on social channels, like Facebook, in comparison to live (web) chat.

… messaging can come with serious costs when connecting to SMS or WhatsApp

Looking at costs, there is another detail you should take into consideration: Bots will increase the number of outbound messages.

Does that matter?

Yes, because messaging can come with serious costs when connecting to SMS or WhatsApp channels.

From a business case perspective, if you have 10+ agents working on customer care, it’s practically a no brainer to start working with bots. If you have fewer FTEs you still save a lot of time, money, repetitive work and improve your customer satisfaction, but you need to consider what you want to invest to create and maintain the bot.

Larger enterprises, like banks and insurance companies, tend to build their own tech first until they figure out they suck at it

Now, how about the ROI if you want to build your own solution instead of using an existing cloud-based solution?

Ok. Let’s say we create a bot that reduces about 100k costs a month. That’s over a million yearly.

We build a bot from scratch that costs us a million (let’s keep it simple). We’d still break even within 12 months?

This sounds too easy, and it is. I’ve seen this mistake being made a couple of times, the reality is:

  • You won’t start to reduce costs immediately after the bot is live
  • Building a bot is one thing, but for maintenance, you need to factor in additional costs.
  • You get stuck in an IT delivery process. In other words; can marketeers or customer service agents manage the bot themselves or do they need tech support? If the latter is true, at least triple the running costs and reduction in time to market for changes.

Larger enterprises, like banks and insurance companies, tend to build their own tech first until they figure out they suck at it.

Think of building your own CMS or payment system, you will always follow, not lead the market and end up spending millions on external consultants.

Marketing and sales

Next use case, marketing, and sales.

This is a bit harder to estimate. There is a big difference between social media campaigns and for example, capturing leads on your website.

With messaging you can achieve far higher engagement and conversion rates…

For social media campaigns, you can use your current campaigns as a performance baseline.

With messaging you can achieve far higher engagement (> 600%) and conversion rates (> 300%). Of course, if your target audience is actually on social media…

Messaging reduces the number of steps in a funnel, it’s faster, asynchronous and, if necessary, can have human agents step in to close the deal.

From a costs perspective, your first campaign would be a bit more expensive but after that, it’s the same effort to launch, run and maintain these campaigns as you do right now.

So, is it interesting to my business?

  • Do you spend on social media advertising?
  • Do you run email campaigns?
  • Do you want to capture leads on your website?
  • Any or all of the above?

The answer is simple: yes. Take some budget, run a chatbot powered campaign, experiment and review the results.

Conversational UI

Track a bag, onboard new users using a chatbot or create a complete conversational website or app.

..does this fix a problem for my customer or improve my product?

Conversational UIs — or CUIs — run on the same basis of any user experience or CX investment:

  • Improve the user experience
  • Innovate customer interactions
  • Decrease customer effort

For dedicated apps, it’s good to first ask the question “does this fix a problem for my customer or improve my product?”. The answer would be the core argumentation of your business case.

There are a number of reasons why CUIs can be a big success:

  • Customers don’t need to install separate apps, nor visit a mobile website
  • They can choose their preferred channel
  • Customers already use CUIs every day (messaging apps) so there is little to no learning curve
  • CUIs can simplify or even remove complex navigation
  • They are easily personalized and can be updated real-time
  • It’s easy to gain fine-grained insights on how people interact with your app

It is a pretty long list and the business case is almost entirely focussed on the customer/user experience. Does your app need to be conversational? CUIs are no silver bullet, but a good UX designer can choose it as a solution.


You should have already gotten a sense of numbers.

To start with, I usually recommend doing a short pilot. Approximately 4–10 weeks for customer service or marketing.

During that time you can experience how people interact. Planning could look like:

  • 3 weeks setting up a bot
  • 2 weeks running a pilot
  • 2 weeks of refinement
  • Go/ no-go and move to production

The exception is for CUIs. You might want to consider starting with user tests before creating a bot.

There are plenty of cloud-based, low-code Chatbot platforms out there and most of them are cheap. Designing and developing a chatbot isn’t, that's why aiming for a 2-4 week creation period is what you should aim for.

What platform to choose depends completely on your organization, channel, and functionality. After creating the bot the biggest costs are in training the AI, coding functionality and keeping it online.

About The Author

Gijs van de Nieuwegiessen is CEO and Co-founder of, a Chatbot design, and creation tool. He has a background in UX design, a creative technologist, and started his first company in 2005. You can connect with him on Twitter and LinkedIn.

Don’t forget to give us your 👏 !

Creating a Chatbot Business Case was originally published in Chatbots Life on Medium, where people are continuing the conversation by highlighting and responding to this story.

Source link

Related posts

Wearable Fall Detector Using Recurrent Neural Networks.


Binary Weighted Memristive Analog Deep Neural Network for Near-Sensor Edge Processing. (arXiv:1808.00737v1 [cs.ET])


More than half of Maine prescribers are now EPCS-enabled


This website uses cookies to improve your experience. We'll assume you're ok with this, but you can opt-out if you wish. Accept Read More

Privacy & Cookies Policy