A checklist for choosing chatbot tools

Doris Chi
Chatbots Life
Published in
5 min readJul 18, 2019

--

Photo by Glenn Carstens-Peters on Unsplash

Chatbots have been around for more than 50 years, but not until recent years did the chatbots start to show great business values. More and more companies start to use chatbots to improve customer satisfaction and employee productivity. The market is so big that many tech companies extend their business line from ticketing system, customer relationship management system, knowledge management system, search engine etc. to chatbot tools. Many start-ups also joined the battle with chatbot tools as their primary products. There is still no clear dominator in this area and no company has claimed that their tool can meet all the conversational needs.

How to choose the chatbot tool for your business from hundreds of candidates? Here is an incomplete list of questions to ask before making your final decision.

Chat experience or Search experience

  1. How are the user intents defined?
  2. How many examples are needed to train a user intent?
  3. How many user intents can the tool handle?
  4. Is there auto-complete or auto-suggestions while users are typing?
  5. Does the tool recognize entities?
  6. Does the tool remember the context?
  7. Does the tool support variations of responses?
  8. Can the tool choose different responses for different users?
  9. Does the tool support response in the rich format like link, image, video, table etc.?
  10. Does the tool support dynamic responses based on real-time data?
  11. Can the tool get the response by calling APIs?
  12. Can the tool save user input and pass them to APIs as parameters?
  13. Does the tool prefer user input as keywords or in natural language?
  14. How well can the tool handle typos, grammar mistake, and informal expressions?
  15. How well does the tool recognize synonyms and different syntactic structures?
  16. Can the chatbot detect positive and negative sentiment?
  17. Can the chatbot transfer to live agent when it detects negative sentiment or cannot find any relevant responses?
  18. Can live agent transfer back to chatbot when there are relevant responses?
  19. Can the chatbot provide escalation channels like an email address or a phone number when it cannot find any relevant responses?
  20. Does the chatbot support voice input?
  21. Does the chatbot support multi-language input?
  22. Does the chatbot auto-translate response to user’s language?

Top 4 Most Popular Bot Design Articles:

1. Ultimate Bot Design Guide

2. How to design a Chatbot

3. Distributing a slack app

4. Chatbot Conference in San Francisco

Front End

  1. Does the tool have out of box user interfaces?
  2. What kind of user interfaces does the tool provide? (e.g. floating widget, embedded widget, sliding widget, support center)
  3. How easy is it to deploy the out of box user interfaces?
  4. How many endpoints can the widget be deployed to? (e.g. Webpage, Facebook messenger, Slack, Google Assistant, Alexa, SMS, browser plugin)
  5. How easy is it to customize the look of the user interface?

Back End

  1. Is there a backend console for the chatbot?
  2. How are user intents, entities, dialog flow, or knowledge articles stored?
  3. How easy is it to edit my content in the console?
  4. Can I import my content all at once?
  5. Can I export my content?
  6. What kind of file can I export my content to? (e.g. JSON, CSV, PDF etc.)
  7. Is there version control on my content?
  8. Can I take my content offline without deleting it?
  9. Can I publish new intents or knowledge articles in a future date?
  10. Can I set the expiration date on intents or knowledge articles and take them offline after the expiration date?
  11. Is it easy to search or filter the user intents?
  12. Does the tool provide a separate production and staging environment?
  13. Is it easy to release the article from the staging environment to the production environment?
  14. Is it easy to test my chatbot before deploying it?
  15. Can I collaborate with other teammates in the console?
  16. Is there a reporting or analytics module?
  17. Is the conversation log stored as a whole or as individual queries?
  18. What kind of report does the tool provide? (e.g. traffic, most popular user intent, user input with no matched intents etc.)
  19. Does the chatbot tool provide APIs or SDKs for automation and integration with other tools?

Security

  1. Where is my content stored?
  2. Are my content encrypted?
  3. Are the user input encrypted?
  4. How does the tool authenticate users?
  5. Does the tool support single sign-on?
  6. How often does the tool backup?
  7. What does the tool backup? (e.g. content, widget configuration, user log)
  8. How to restore or recreate my chatbot if something goes wrong?

Pricing

  1. Is there a free tier or trial period?
  2. What is the pricing model?
  3. Can I upgrade my account if I need more intents and user interactions?

Other

  1. What is the development plan for the tool?

If you are trying to build a chatbot, I hope this list can help you choose the perfect tool. If you are trying to build a chatbot tool or platform, I hope this list can give you some ideas on what features to develop. I will keep adding items to this list and hopefully it will be a complete list one day.

This the first blog I have ever written in my life. Thanks for reading it and please feel free to leave some feedback.

Don’t forget to give us your 👏 !

--

--

I am a full-stack software engineering interested in NLP and ML. Opinions expressed are solely my own and do not represent my employer.