The goal of an ROI calculation is twofold: to ensure that the project makes sense economically speaking, and to draw the backbone of the chatbot features .
The first goal is an estimation based on volume data, financial data and the experience we have gained from more than 50 chatbots built in less than 3 years.
The second objective is to calculate the interest of inserting such or such functionality in the chatbot, based on certain volume data and subject typologies.
Let’s focus on the number one interest of ROI calculation: knowing how much money my company will save thanks to a chatbot.
This spreadsheet is mainly used to calculate the automation potential of a chatbot. It is only valid for projects whose objective is to reduce reccuring tasks, guide users and customers in a new path or a new application, etc...
The first step is to know how the support is managed on the task you want to automate.
The second step is to know how you are going to set up your chatbot. The objective is to calculate the cost of this implementation.
Then we focus on the features to be integrated into the chatbot. We will start quite simply either from data that you can extract from your systems (ITSM tickets,…) or even from the feelings of your business teams who can identify in less than 30 minutes the questions they answer recurrently, as well as the way they answer them.
The first step is to list all the topics in the column “Top 20 most recurrent topics”. (if you have more than 20, just add rows and remember to take into account the new rows in the sum calculation)
Then you will have to enter the percentage that this subject represents in the total of the requests.
Finally, you will have to enter the type of action that your team is carrying out today to respond to this request. Behind each type of action, we apply an appropriate automation percentage, based on our experience.
Step-by-step procedure performed on the phone: 30% automation
Action in an application (Salesforce, ServiceNow, Azure,…): 80% automation
Request for clarification by email or phone: 50% automation
Qualification of the request and referral to the right team: 30% automation
Referral to an FAQ: 10% automation
This gives you at the end :
The total percentage of subjects addressed by the chatbot
The percentage of automation of the chatbot according to the way they are answered
Here the objective is to identify the topics that have the best ratio of simplicity of implementation / automation.
If a topic represents only 1% of the requests but requires an API integration with an on premise system, we will drop this functionality.
If a topic represents 30% of the requests and is currently only handled by an FAQ, we will try to increase the chatbot’s competence to make it manage this task autonomously and thus reach an automation level of 30%*80% instead of 30%*10%.
All the elements are in place. All that remains is to calculate the final ROI. Here are the steps:
Calculate the cost of internal or external support on the estimated perimeter of the bot: we take the cost of support and apply the percentage of ticket volume of the top 20
Calculate the gain of the chatbot: we apply to this cost the percentage of automation according to the positioning of the bot
Calculate the ROI: we remove the Vizir license from the previous calculation
Here are some examples of nice ROI :
if you manage 2000 tickets per month, the chatbot is profitable from an automation potential of 50%.
if you manage 5000 tickets per month, the chatbot is profitable from an automation potential of 20%.
if you manage 10 000 requests per month, the chatbot is profitable from an automation potential of 10%.
Sur ces belles paroles, il ne vous reste plus qu’à remonter cette page et télécharger le fichier Excel qui vous permettra de calculer le ROI de votre chatbot 🙌