Common sense dictates that besides paying for the development of the chatbot it’s worth considering the cost of running the bot itself. Let’s take a deeper look into the so-called total cost of chatbot ownership.
It consists of 5 components that have to evaluate for every project:
Cost of licenses for the tools. There are many available tools that could speed up the development process of a chatbot and implementation of the corresponding AI. I could just name a few: Smooch, DialogFlow, IBM Watson, Amazon Lex, Chatfuel and many others. Also, you will have at the end of the day to pay money their usage, it could be a fixed subscription fee or cost per message. Please check the details on the website of the vendor.
To tell the truth, there is an option to consider the free open source projects such as Botkit, Botpress, Node-red or others, where there is no need to pay for licenses.
You will also need to pay for access to live agent software if you’re going to use it in your project.
Running costs of sending messages. Implementation of a chatbot for WhatsApp or for SMS will also require making payment for sending messages. We extensively using Twilio and highly recommend it.
Servers and environments costs. Your chatbot should be running somewhere. Even if you’re using a SaaS solution mentioned in the item, you will probably need to have a back-end part that should be hosted somewhere, otherwise, the whole bot should have hosting either in the cloud or on-premise.
For cloud installations, we generally recommend to our clients that they start with the lowest configuration of environment (about 25 USD per month for AWS ) and increase its usage while the bot will increase in popularity. In terms of the total cost of ownership, all the expenses are considered to be operational.
While on-premise installations are more complicated for TCO calculations, it’s generally considered the cost of hardware, cost of its maintenance, cost of server-side software (operational system, application servers, etc.) I advise talking with your local IT team to get that figure.
Support from the development team and bug fixes. The bot is the application that requires support from the developers. This includes normal, run of the mill, bug fixes as well as the allotted amount of time for the improvement of Chabot. Generally speaking, the customer and team of chatbot developers have a Service Level Agreement (SLA) wherein the response time and limit of monthly reserved hours are agreed between all parties.
Support and tuning of the conversation. The conversation of a chatbot is rarely perfect from the get-go. It requires tuning and training on a daily basis from the side of user experience. We estimate that on average, about 30% of Chabot development costs should be allocated to that particular item during the first 3-4 months after publishing the chatbot.