At what point should you invest in a bot?

Chatbots are an investment, but they usually pay for themselves in < 1 year. Here's what to consider before you invest in your next bot.

‍

A majority of our prospective customers come to us with the question: When’s the right time to start investing in a chatbot? It’s a great question because building a chatbot is an investment. You need to dedicate resources to building a bot before you experience the value.

‍

You probably figured out already that our answer to this question is related to return on investment (ROI). And how do we measure ROI for a chatbot? It’s simple, we use the time saved by the chatbot. There are other KPIs such as recognition rate and CSAT score, but the main KPI is time saved vs. bot cost. Basically, whatever you invest in the chatbot, you need to make back in time saved.

‍

There are a few reasons why we usually recommend starting with a pilot project, but one of the main reasons is that you should always validate that you can make back what the initial project costs in time saved by the bot. Below are the questions that we ask our prospective customers, as well as a Chatbot ROI Calculator to help you build your next chatbot business case.
‍

Questions to ask yourself before investing in a bot​

There are two sets of questions that matter when deciding to invest in a chatbot project. The first set of questions helps you figure out whether or not you have a business case. The second set of questions helps you figure out the scope of your business case.
‍

Part 1 - Do you have a business case?​

These are the basic questions that help you decide whether or not you have a solid need for conversational automation. At the end of the day, we’re very practical people. If you won’t experience a return on your investment, then it’s not worth it. Period.

‍

Early on a potential customer asked us to build them a bot. They had 20 conversations a day. We said no because we’d spend a lot of time building a fancy solution that would cost far more than the client would ever experience in saved time. It wasn’t going to be a win-win. We’re in business to enable people to focus on what truly matters by using AI to eliminate boring tasks. If we can’t do that, then there’s no point in launching a project.
‍

How many questions/interactions do you receive per month via chat?​

Automation works when there are high volumes. Because time saved is the main KPI, you obviously need enough conversations to make automating them worthwhile. There isn’t a one number fits all answer, but generally, if you receive +2,000 conversations per month from one channel in one language, then you’re already in a good position to be considering a chatbot project.
‍

What percentage of those questions/interactions do you want to automate?​

Not every conversation can and should be automated by a chatbot. You want to select the percentage of incoming questions/interactions that can be automated. A modest estimate for a pilot project is 25% of the total. For more tips on which types of questions/interactions are best for automation, check out this article.
‍

What’s the average time spent on answering a question or resolving an interaction​

How long does it take for a person to open the conversation, read it, find the answer, respond, confirm that the user doesn’t have any remaining questions and close the conversation? Usually, this takes anywhere from 3-5 min. for easy-to-answer questions. For more complex flows that require data capture, the time can quickly increase to +10 min.
‍

Part 2 - How much should I invest in my chatbot pilot project?​

To realize a return on investment, you need to figure out how much saved time it will take to cover the initial chatbot project costs. You should never invest more than you will save because it’d be faster and easier to flush your money down the toilet.
‍

What’s the cost per day for someone to follow up/resolve those conversations?​

This is pretty straightforward. Just figure out the hourly rate of the person(s) handling those conversations. If it differs per profile, then take the average cost.
‍

What’s your recognition rate?​

The recognition rate is the percentage of incoming questions that your chatbot recognizes correctly within your automation scope. At Campfire, we have a standard of 90%, but this can vary per implementation partner, so make sure to ask before agreeing to a project.
‍

What are your costs?​

This includes your annual SaaS license fees, the full cost of your implementation project and the annual costs for general maintenance.
‍

Calculating your chatbot’s ROI​

Now that you know which questions to ask yourself, it’s time to plug in all the numbers and figure out the details of your business case!

We'll walk you through an example, but download this free calculator to follow along and calculate your own business case 👇
‍

‘Smores Stadium is an arena in Brussels that hosts +100 events per year and its customer support team receives ~10,000 English messages per month via chat. The Customer Support Director, Marshall Mallow thinks that a chatbot should be his team’s next hire, but he needs to convince the CEO with a solid business case.

‍

After talking with Campfire, Marshall thinks that he can automate 25% of the incoming messages. He also estimates that it takes on average, 5 min. for an agent to resolve a chat. It costs ‘Smores Stadium 60,000 EUR per year to employ a customer support agent, which comes to €285 per day.

Marshall plugs all of this info into the ROI Calculator and sees that he should save €80,000 per year! The initial estimate for the pilot project and the first year of the license is €45,000. This means that he has a 78% ROI and will recover his investment in less than a year! Now that’s a deal that’ll spark his CEO’s interest.