AI

Changing Information for Your Chatbot Flows in Teneo

Image title

Changing information in Teneo

Every now and then something changes in your business, it might be that you expand the business or launch new products. These kinds of changes require changes to your chatbots’ answers in order to keep your chatbot updated with the latest information. Let’s go through how you can do this in Teneo using an example from Longberry Baristas.

You might also like:  7 Powerful Chatbot Building Platforms

Longberry Baristas has decided to open two new stores, one in Berlin and one in London. Your flow about stores in different cities that you created earlier needs further updating:

User: Do you have a store in London?
Bot: We have recently opened a brand new store in London! We’re looking forward to seeing you there!

User: Do you have a store in New York?
Bot: Actually, we opened our first café in New York in 2016. You’re very welcome to visit us there!

User: Do you have a store in East Leake?
Bot: We have stores in New York, London, and Berlin. We might consider opening additional stores in East Leake in the future.

Here you’ll learn how to add a branch at the beginning of your flow and to specify different paths depending on what the user said:

  • if the user asks about London or Berlin, we want to say that we have just opened a store there,
  • if the user asks about New York, we want to say that our first store was opened there,
  • if the user asks about any other city, we want to tell the user where we do have stores, and that we may consider opening one in the requested location in the future.

These are the steps you’ll need to take:

  • Extend the flow structure:
    • Add a junction, and set it to be start-node
    • Add two new output nodes
    • Link the junction to the output nodes
  • Populate the output and transition for the Berlin/London response
  • Populate the output and transition for the New York response
  • Update the output for other cities

This is what your flow will look like when you’re done:

Final flow structure

Extend the Flow Structure

First, you should open or create the ‘User wants to know if we have a store in city’ flow in edit mode:

  1. Select the flow ‘User wants to know if we have a store in city’.
  2. Click ‘Edit’ in the ribbon. The flow will open in a new window.

Follow these steps to re-draw the structure of the flow:

  1. Add a junction by clicking ‘Junction’ in the top ribbon’s ‘Add Node’ section. Make sure nothing in your flow is selected before doing so, or the new junction will be connected to the selected item.
  2. Keep the new junction selected and click ‘Set Start Node’ in the ribbon’s ‘Flow Entry’ section. The graph will be re-drawn so that the junction comes straight after the trigger.
  3. Connect the junction and the output node ‘Our only store is….’ as follows: Click on the junction, and then drag the cursor to the center of the output node. If you want to cancel the transition while drawing, click Esc.
  4. Select the junction and click on ‘Output’ in the ribbon’s ‘Add Node’ section. This will add a new output connected to the junction.
  5. Repeat step 4 to add yet another output. In total, you should now have three outputs after the junction.

[embedded content]

Populate the Output and Transition for Berlin/London

We’ll start with the branch for questions about the recently opened stores in Berlin or London:

  1. Select one of the new output nodes, and add the following answer text: We have recently opened a brand new store in ${city}! We're looking forward to seeing you there!. Name it: We recently opened a brand new store there!.
  2. Now we need to make sure this answer is only given when the user is asking for Berlin or London. Select the transition and open its ‘Examples’ panel. Set the toggle to from ‘unconditional’ to ‘conditional’, and paste the following into the condition field: %LONDON.NN.LEX / %BERLIN.NN.LEX. Name the transition: Recently opened stores. Make sure it is ‘Continuing without input’ using the toggle.
  3. This transition should be tested first. You can change its order at the bottom of the ‘Examples’ panel: change it to 1.

[embedded content]

Populate the Output and Transition for New York

Now, we populate the New York part of the flow. The steps are the same as before:

  1. Select the other new output node, and add the following answer text: Actually, we opened our first café in New York in 2016. You're very welcome to visit us there!. Name it We have a cafe in New York since 2016.
  2. Select the transition. Make it conditional and add the following condition: %NEW_YORK.NN.LEX. Name the transition New York.
  3. Set the order of the transition to 2.

[embedded content]

Update the Final Output Node

As a last step we will update the answer text of the final output node:

  1. Select the output node ‘Our only store is…’
  2. Change its answer text to the following: We have stores in New York, London and Berlin.
  3. Replace its name by We have stores in....
  4. Leave the order of the transition at 3
  5. Hit ‘Save’

[embedded content]

We don’t have to add a condition to this transition, we thus leave the toggle set to ‘unconditional’ as it is by default. This transition will be tested after the other two transitions, thus any question about London, Berlin or New York would already have gone down either of the other transitions. And questions about any other city, will go down this path!

Try It Out!

Now go ahead, open the ‘Try out’ panel and give it a go!

Add More Variation to Your Outputs

Right now, we have a working flow. However, you may have noticed that we always get exactly the same answers for exactly the same questions. That may make your bot seem a little repetitive and not very humanlike. What we can do about it is to add more variation to the bot’s answers. We illustrate the process for one output node, but it can then, of course, be repeated for the other output nodes as well:

  1. Go back to the flow and select the output node called ‘We have stores in…’.
  2. In the ‘Answers’ panel on the right-hand side, simply add more possible answers. Ideally, these should vary but still retain the same content. For example, you can add the following answers using the little green ‘add’ icon at the top. Click the add button and then add one answer at a time:At the moment our only stores are in New York, London and Berlin.
    You can find us in New York, London and Berlin.
    Come and visit us in one of our stores in New York, London or Berlin!
    As soon as you have added a second answer text, the first one will appear as the ‘Fallback’. That means that if the user asks the question triggering this output more often than we have a different answer for (here: more than 4 times), the answer given will always be the ‘Fallback’ one from then on.
  3. Now go ahead and select the subsequent output node ‘We might consider opening a store in city’ and add a few more answers there as well, for example the following ones:However, we may open a new store in ${city} soon.
    You may find us in ${city} in the future as well!
    But who knows, maybe we will soon open a store in ${city} as well?!
  4. Hit ‘Save’

[embedded content]

Try It Out Again!

That’s it! You may now give it another try in tryout. Ask your bot similar questions along the lines of, “Do you have a store in the city?” and you will get a variety of different answers. What happens internally is that Teneo randomly picks an answer from the first node and then combines it with a randomly picked answer from the second node. This results in a great variety of different answers and will make your bot appear less repetitive and more humanlike.

Further Reading

A Beginner’s Guide to Creating an Interactive Chatbot Flow in Teneo

Most Common Mistakes to Avoid When Building a Chatbot

Comment here

This site uses Akismet to reduce spam. Learn how your comment data is processed.