Step 1.  On the Helpshift toolbar, click Settings (Settings).

Step 2.  In the navigation pane, scroll to the WORKFLOWS area. Then, click Bots.

The Bots page opens. Multiple tabs are visible near the top of the page.

Step 3.  To see and work with your feedback bots, click the Feedback Bots tab.

UI elements on the Feedback Bots tab:
  • list all of your configured feedback bots
  • distinguish among feedback bots that are published, unpublished, or disabled
  • help you to create, configure, publish, unpublish, disable, and reenable feedback bots
Step 4.  On the Feedback Bots tab, click + New Feedback Bot.

The New Feedback Bot dialog box opens.

Step 5.  In the New Feedback Bot dialog box, enter a name and nickname for the bot.
  • Bot Name — Unique “bot names” help your team to distinguish among feedback bots and use them deliberatively. For example, bot names could help your team to represent the results of A/B testing.
  • Bot Nickname — When your brand allows the display of feedback bot “nicknames,” support messaging sessions show these nicknames to your end-users. There is no requirement to have or use more than one bot nickname. For example, you could call each bot "Product Support," and this is the only name that your end-users would see, irrespective of which bots chat with them.
Step 6.  Click CREATE AND CONFIGURE BOT.

Visual Bot Builder opens in its canvas view. This is where you choose which steps should run when anything triggers the bot.

Step 7.  In Visual Bot Builder, click Add First Step.

A contextual list opens, citing supported bot steps. Some of the bot steps may be familiar to you because custom bots also support their use. Other bot steps are new to, and exclusive to, feedback bots.

◼️◼️◼️

Get Information from User — Prompts the user to answer a question that you control. The prompt itself can incorporate any number of placeholders (variables) that you choose.

TIP

Pay close attention to grammar and diction in sentences that incorporate placeholders. Their actual output can differ from what you expect.

Permitted end-user response types are as follows.
  • text (valid Unicode characters.)
  • email address
  • number
  • date
  • list of options
  • yes/no
You can:
  • consume the end-user response as input that auto-populates a custom issue field (CIF) of your choosing.
  • allow users to skip your request for information and proceed directly to the next bot step.
  • consume the end-user response as input that auto-populates a customer satisfaction (CSAT) survey comment in the end-user's name.
◼️◼️◼️

Send a Message — Shows a message to the end-user.

A common use case is to announce what the end-user should expect to observe next in their support journey, such as issue handoff from the bot to a customer service agent.

◼️◼️◼️

Branch Based on Custom Issue Fields — Choose among your custom issue fields (CIFs) and set which values should branch the conversational flow.

Helpshift executes branching logic rules in the same order as they are saved, per step. To adjust their order, edit a branching step, drag its rules to new positions in the list, and then save your changes.

◼️◼️◼️

Make an API Request — Trigger an API request (GET, POST, or PUT) of your choosing, with different outcomes depending on whether the request succeeds or fails.

An administrator must set up external API requests before they become available in the API request to be triggered list that this bot step shows. Administrators set up your brand's external API requests at Settings > External API Requests > [+ NEW API REQUEST]
  • Success — The API response code is 2XX and the response returns all keys that your administrator specified for the current request.
  • Failure — The API response code is 3XX, 4XX, or 5XX, or the returned response is missing one or more keys that your administrator specified for the current request.
◼️◼️◼️

Confirm Issue Resolution — Prompts the end-user to confirm that their issue — now that it's marked as resolved — is really resolved.
  • The text for this prompt can incorporate any number of placeholders (variables) that you choose. However, you are not obligated to use even one placeholder.
  • The end-user can choose only Yes or No as their reply. You must configure what happens next in either case.
  • You can set when this step expires — a duration that's measured in hours or days, for example. When you enable and configure an expiration, the countdown begins as soon as this step runs. Then, when this step expires for a particular conversation, your Helpshift instance stops waiting for the end-user to respond, and proceeds without having received any response. You must configure what happens next, after a bot step expires.
  • Map User Reply to a Custom Issue Field (CIF) — Automatically set the value of a CIF to match the user reply.
TIP

Pay close attention to grammar and diction in sentences that incorporate placeholders. Their actual output can differ from what you expect.

◼️◼️◼️

Get Feedback from User — Prompt users to rate the quality of customer service they received for their now-resolved issue.

◼️◼️◼️

Close Conversation —
    • Stops messaging about the end-user's now-resolved issue.
    • Shows a Start New Conversation button to the end-user.
    • (Optional) Triggers one or more actions, based on your selections from a list of supported actions.
        • Workflows

          Add private note — Enter free text for Helpshift to save as a private note in Agent Dashboard. To tag particular agents, supervisors, or administrators, type the '@' sign and choose their names from the list of registered users.
        update default issue fields

        • add issue tags — Choose from a list of administrator-managed tags that are not currently applied to the now-resolved issue. Choosing tags from the list adds them to the issue.
        • remove issue tags — Choose from a list of administrator-managed tags that are currently applied to the now-resolved issue. Choosing tags from the list removes them from the issue.
      • update custom issue fields — Choose among whichever CIFs are configured for use on your Helpshift instance.
TIP

Feedback bots run for resolved issues exclusively. Nonetheless, updating CIF values and/or CSAT comments for a resolved issue after the fact can still help to surface important trends in analytics and reporting.

◼️◼️◼️

NOTE

Although the Reopen Issue bot step is available for use in feedback bots, we do not support its use as the first step in any feedback bot.

Reopen Issue — Reopens the issue, assigning it:
  • to a queue that you choose
  • back to the issue's original queue
  • to a bot that you choose
  • to an agent whom you choose
  • back to the issue's first-assigned agent
(Optional) Triggers one or more actions, based on your selections from a list of supported actions.
  • Workflows
    • Add private note — Enter free text for Helpshift to save as a private note in Agent Dashboard. To tag particular agents, supervisors, or administrators, type the '@' sign and choose their names from the list of registered users.
  • update default issue fields
    • add issue tags — Choose from a list of administrator-managed tags that are not currently applied to the now-resolved issue. Choosing tags from the list adds them to the issue.
    • remove issue tags — Choose from a list of administrator-managed tags that are currently applied to the now-resolved issue. Choosing tags from the list removes them from the issue.
  • update custom issue fields — Choose among whichever CIFs are configured for use on your Helpshift instance.
TIP

Feedback bots run for resolved issues exclusively. Nonetheless, updating CIF values and/or CSAT comments for a resolved issue after the fact can still help to surface important trends in analytics and reporting.

NOTE
  • You can delete bot steps from an unpublished feedback bot. [Link to procedure]
  • You cannot delete bot steps from a published feedback bot.
Just as adding a bot step changes fundamentally how a feedback bot behaves while chatting with your end-users, so does deleting a bot step. Such deletions may also change what information is collected and how your Helpshift instance applies or uses the information in subsequent steps.
We recommend that you consider all possible effects before, during, and after your addition or deletion of any bot step.

Step 8.  In the contextual list, click the name of a type of bot step.

The Edit Step dialog box opens. This is where you configure how the step should function in your feedback bot.

Step 9.  When you finish configuring the first step, click Done.

The Edit Step dialog box closes, restoring your view of the Visual Bot Builder canvas.

A visual representation of the newly configured bot step uses a plus symbol to indicate each outcome that you configured the bot step to support.

Next, you must:
  • choose which bot step each supported outcome should trigger.
  • configure behaviors for each such bot step.
Step 10.  To begin configuring what should happen after a particular outcome, click the associated plus symbol.

A contextual menu opens.

Step 11.  To choose a type of bot step, click its name in the contextual menu.

The Edit Step dialog box opens. Its options are particular to the type of bot step that you chose.

Step 12.  Configure the new bot step, and then continue as before.

a.  Choose and configure other bot steps, where needed.

TIP

You can use as many as 25 steps per bot. Even so, you may find that it's better to use only three or four bot steps, possibly lowering any risk of end-user drop-off.

b.  To return to the Visual Bot Builder canvas when you finish setting up a step, click Done.

c.  Repeat as needed.

NOTE

Visual Bot Builder saves your work automatically.

Nonetheless, a feedback bot is not available for use in any automation until you publish it.

Step 13.  (Conditional) To publish your feedback bot, click Publish on the Visual Bot Builder canvas.