Alice

We've just published a detailed tutorial on how to combine ChatGPT with your own customer data and display the results in Zendesk on the FactBranch blog and I thought you all here might be interested. Combining your own data with a GPT prompt is such a powerful way to enhance what ChatGPT can do for you.

How it works

• Load data from your data source (PostgreSQL, MySQL, Microsoft SQL Server, Google Sheets or any RESTful API).
• Merge that data into a custom GPT prompt.
• Display GPT's output in the FactBranch Zendesk app.

In essence you can use your JSON-formatted data in your prompt and let GPT do the rest.

Here are the steps how to do it:

You need three accounts: An OpenAI account with API access, a Zendesk account (of course), and a FactBranch account.

Connect FactBranch to your data source and create a data pipeline in which you first load data from your data source and then send an augmented prompt to OpenAI's API like this:

Writing a prompt

We've tried many prompts and output formats. What we found is that short prompts work best because they don't confuse GPT what's important to you. And a powerful keyword to use is "bullet points". This tells GPT you want concise sentences, each on a new line.

Then we include the data and the ticket text and let GPT summarize the relevant data into 1-3 bullet points.

For example the final prompt (with the placeholders for "question" and "data" replaced) that FactBranch sends to GPT could be:

----------------------------------------------------------------------

Summarize data into 1 to 3 short bullet points that are relevant to the question.

Question: "Where is my last order?"

Data: [{"order_id": 123, "status": "shipped", "items": "3 Hoodies, 2 T-shirts"}, ... ]

----------------------------------------------------------------------

And GPT may return something like:

- There are 2 orders that aren't delivered yet
- Order 123 includes 3 hoodies and 2 t-shirts and is shipped.

Final thoughts

It's really exciting to see GPT work through your own raw data and I hope this is useful to a lot of people here.

I'd love to hear how you're using GPT to make your operations more efficient! Simply comment here or send a short email to support@factbranch.com to get in touch!

Thanks, Raphaël Péguet - Officers.fr! :)

OK, because I tried this out and couldn't get it to work. You can filter the sheet for while you're editing it, but not from the published web page.

The use case in the video is: You have dozens to thousands of rows in your sheet, then a customer creates a ticket and when you open the ticket, the FactBranch app automatically only shows data from the matching row(s).

E.g. say you have a ticket sent by charlie@example.com, when you open the ticket, only data from the row with charlie@example.com in the "Email" column gets displayed automatically without any interaction needed:

Interesting idea! Does that simply display the whole spreadsheet or can you automatically filter the rows to only show the ones that relate to the ticket requester?

Every company stores data in spreadsheets. For those using Google Sheets I have good news! You can show relevant rows in your Zendesk tickets, contacts, deals and chats.

We just published a step-by-step video guide on how to display Google Sheets data in the Zendesk sidebar and I thought some of you will find this useful!

Here is the video:

It only takes a few minutes to set up and all you need is:

• A Google Sheets spreadsheet that contains your data (e.g. customer data)
• A Zendesk Support, Sell or Chat account
• A FactBranch account

FactBranch is a service that connects to different data sources, queries those sources and in the end displays your data in Zendesk.

It only takes a few steps to set everything up:

1. Connect FactBranch to your Google Sheets account
2. Create a data flow in FactBranch
3. Add a Google Sheets query to that flow
4. Auto-generate the output that you want to see in the Zendesk app
5. Install the FactBranch app in Zendesk Support, Sell or Chat

Once you're done, the app will look like this for example in the Zendesk ticket:

Those data are loaded live from a Google spreadsheet and setting this up only took a few minutes.

We'd love for you to try this out! Let me know what you think.

This is a use case we come across often: Show a dropdown that lists data from a live database. Either loading something specific to that ticket (like the customer's most recent orders) or a general list of items (like a list of locations or products for example). And when the agent selects an entry from the dropdown, save that value to a ticket field.

We've just added that feature to our app FactBranch and here is how it looks:

Step 1 - Setup FactBranch

First you need to create a FactBranch account and connect your data source. This could be a PostgreSQL, MySQL or SQL Server database.

Step 2 - Adapt your Query

Make your query return an array of JSON objects. This is a bit different in every SQL dialect and so we wrote specific documentation on how to do this here:

• PostgreSQL
• MySQL
• Microsoft SQL Server

Step 3 - Add the dropdown

In the FactBranch Designer set the type of the item containing the JSON-response to "HTML" and add the dropdown like this:

{{ value|form_select('name', 'order_id', 'ticket.customField:custom_field_123456789') }}

Here is what's going on: name is the key in the JSON object that should be displayed in the dropdown, order_id is the value that will be assigned to the ticket's custom field 123456789.

If you need help …

… do reach out to support@factbranch.com. This is an advanced feature and we’re happy to help you set it up or hear what else you need.

If you want to display a collapsible (an item that expands when clicked showing more info) in Zendesk, there now is a way! Here is how this will look:

As you can see this is part of the FactBranch app, that you can install in Zendesk Support, Sell and Chat. FactBranch loads customer data from your database, Google Sheet or REST API and displays it in the sidebar in Zendesk.

Step 1 - Setup FactBranch

First you need to create a FactBranch account and connect your data source. This could be a PostgreSQL, MySQL, SQL Server database, a Google Sheet or any REST API.

Step 2 - Adapt your Query

Make your query return a JSON object or array. This is a bit different in every SQL variant and so we wrote specific documentation on how to do this here:

• PostgreSQL
• MySQL
• Microsoft SQL Server

Step 3 - Add the Collapsible

In the FactBranch Designer set the type of the item containing the JSON-response to "HTML" and add the collapsible like this:

{% set preview = [
  'total',
  'summary',
  'status'
] -%}

{% set labels = [
  'order_id:Order ID',
  'total:Total',
  'status:Status',
  'item1:Item 1',
  'item2:Item 2',
  'item3:Item 3'
] -%}

Recent orders

{{ value|collapsible('order_id', preview, labels) }}

This is just an example so let’s walk through that code: First we define the list of fields to use in the preview. This is simply an array of strings. Use the field names from your SQL query.

Then we define the fields to show when an item is expanded. Also use the field names from your SQL query here. To define a custom label (because the raw field names aren’t that nice) use this syntax with a colon:

'field_name:Custom label to display'

If you need help …

… do reach out to support@factbranch.com. This is an advanced feature and we’re happy to help you set it up or hear what else you need.

You can now use FactBranch to automatically look up data in an SQL database and save that data into Zendesk ticket fields.

FactBranch is an app in the Zendesk Marketplace and it's been around for years now. Most people use it to display customer data in the ticket sidebar when an agent opens the ticket. So far the data was just displayed and couldn't be pushed into Zendesk fields. That changed this week!

FactBranch can now pull customer data from your database and save it into ticket fields - including your custom fields! This will happen automatically whenever a new ticket is created.

Here is a quick guide for how to connect FactBranch to your Zendesk account and activate "Push to Zendesk".

1) Getting started

First you need a FactBranch account. If you haven’t signed up yet, you should create an account now. Every account comes with a 14-day free trial and you don't need a credit card to sign up.

Once you've signed up, you have to connect FactBranch to your database. You can follow one of our guides for PostgreSQL, MySQL or Microsoft SQL Server.

Be sure to write a SELECT statement that looks up a single record and returns all the fields you'd like to push back into Zendesk fields. You'll assign the fields in the next step. You can ignore the Output Designer for now. Just click "Finish" to move on.

2) Activating "Push to Zendesk"

Now that you've successfully created an SQL query, return to the Integrations page in FactBranch and in the "Push results into ticket fields" box at the top click on "Connect to Zendesk".

Enter the Zendesk subdomain, user-token and API-key. Here is a guide how to generate an API-key in Zendesk.

Once you've connected FactBranch to Zendesk, head back to your Integrations page, which should now look like this:

Click "Push to Zendesk" inside the box of the integration you've just created. You'll be forwarded to this screen, where you can select which fields from your query should be pushed into which ticket field.

Click "Save" and you'll see a list of fields you've just assigned. It will look like this:

Now there's only one step left: Click "Activate push" and from now on new tickets will be automatically looked up with the query you've written and the fields you've assigned will be saved back into the ticket.

This is especially great if you want to create reports in Zendesk Explore that are based on values of these fields.

Questions or Feedback?

If you have questions about the app or are stuck at any point, please email support@factbranch.com and we’ll help you within a few business hours.

And if this article made your life easier, please also share it on your favorite social network! We love to read those posts by enthusiastic users.

Recently we’ve come across a very common use case: People often have customer data in a Google Docs Spreadsheet. This makes a lot of sense: the online spreadsheet is easy to keep up-to-date across the whole organization and you can use Google Forms to collect the data. But when agents are replying to a ticket it’s difficult for them to look up customer details in the spreadsheet.

This is where our app comes in. We’ve added Google Sheets support to our Zendesk app and I thought I’d share a short guide on how to set everything up. I hope this is useful to all people who use Zendesk and Google Sheets. If it is, let us know at support@factbranch.com :)

Our Solution

FactBranch is an app that loads customer info from a variety of external sources and displays them in Zendesk when you open a ticket. Now that data source can also be a Google Sheet. The app looks like this:

Let’s get started

1) Prepare the column labels

Open the spreadsheet from your Google Docs account and make sure every column you want to display has a title in the first row.

If you want to look up your customers by email, name that column “Email”. If you want to look them up by phone number (when you get calls via Zendesk Talk), call the column “Phone”.

If you want to display a column as a link, add “[link]” to that column’s name. Your spreadsheet should look something like in the screenshot below. First row is for data labels. The rest is for the data itself.

2) Share the spreadsheet with the app

Hit the blue SHARE button.

In the box that pops up click on Advanced.

Then on Change... to change the sharing settings.

Now select On - Anyone with the link. This allows the app to access the spreadsheet.

Click Save in this window and Done in the next ones.

You should now be back to your spreadsheet. Copy the spreadsheet's link from the browser’s address field. It should start with https://docs.google.com/spreadsheets/d/ followed by a lot of random characters.

3) Connect the app to the spreadsheet

Now head to factbranch.com and create an account if you haven’t signed up already. Don’t worry, you can try everything out for 2 weeks for free.

If you have to be GDPR compliant, you can also accept the data processing agreement and add your GDPR contacts.

Go to My Account > Configure Integration and select Google Sheet as your data source.

Paste the link to your spreadsheet that you’ve copied previously:

Click Next, then select how many agents use Zendesk, click Finish setup.

Now the hard part is over. The only thing left to do is installing the app in Zendesk.

4) Install the FactBranch app from the Zendesk Marketplace

Copy the FactBranch API-key and then click on FactBranch App from the Zendesk Marketplace.

On the app’s page hit the Install button and when prompted paste the API-key you’ve copied one step earlier.

Click that Install button one last time.

5) Done!

Now when you open a ticket, FactBranch looks up the customer in your spreadsheet and displays their data in the ticket sidebar.

Contact

If you have questions, feedback or love how the app works, please leave a comment below or send an email to support@factbranch.com