← FrontierBase
Guide

Channel Settings

Give a chatbot answers from Excel or Google Sheets

Upload Excel or CSV, or connect Google Sheets, so a chatbot can answer from that material, then check status and sync. Based on Learning Data in FrontierBase Console.

Give a chatbot answers from Excel or Google Sheets — FrontierBase
Quick answer

What this guide covers

Upload Excel or CSV, or connect Google Sheets, so a chatbot can answer from that material, then check status and sync. Based on Learning Data in FrontierBase Console.

  • Learning Data screen layout
  • Dataset list
  • Add Dataset panel
  • Add from Excel or CSV
  • Connect Google Sheets
  • Pause, delete, recover

The chatbot's answers are based on datasets under Channel Settings → Learning Data. Upload a file or connect Google Sheets. New content is used in replies after the status changes to Ready.

If you have not set up the channel yet, start with Add a chatbot button to your website and set it up. When the channel status is Pending, you usually prepare its learning data on this screen.

Sign in at the FrontierBase Console and open Channel Settings → Learning Data.

Use Guide on the page and inside the dataset settings panel to walk through the same flow on screen.

1. Learning Data screen layout

The page title is Dataset Settings. Inside the card you get toolbar actions and the dataset list.

On the left: Generate with Prompt and View Template. On the right: Add Dataset. An empty list shows the no-data message.

Dataset Settings

Generate with Prompt View Template
Add Dataset

Click a row to open the settings panel and edit that dataset.

Generate with Prompt

Opens a modal with a prompt you can Copy into ChatGPT, Claude, or similar. Edit the [My service] block (industry, topics, tone, language, row count) before you paste it.

The expected Excel shape uses exactly five columns: Category, Item, Detail, Condition, Notes, with a header row. Download the file from the AI tool, then upload it through Add Dataset.

View Template

Opens sample templates by industry (customer support, online store, retail store, education, community, and more). Download an Excel file and adapt the rows instead of starting from a blank sheet.

2. Dataset list

Typical columns:

Column Meaning
Name Dataset name
Type Excel, CSV, Google Sheets, and so on
Rows Row count used for learning
Columns Column count
Status Processing / ready state
Created / Updated Timestamps

Type shows an icon and label. Status uses colored tags. While processing, a remaining-time progress bar may appear. On Failed, hover the help icon to read the error when one is stored.

Click a row to open the edit panel. Use that to check the source and preview, not only the status tag.

Name Type Rows Columns Status Updated
Shipping FAQ X Excel 90 5 Ready 2026-09-17
Product sheet G Google Sheets 120 5 Processing 2026-09-17
Legacy CSV C CSV 40 5 Failed 2026-09-16

New data is used in chatbot answers only after status becomes Ready.

Status meanings

  • Ready: Processing is complete, and the dataset can be used for answers.
  • Processing / Indexing: Work is still running after upload or sync. Save and upload may be blocked until it finishes.
  • Failed: Processing failed because of the file format, encoding, row limit, or a similar issue. Read the error, fix the file, and try again. Pause and resume are not available for failed datasets.
  • Paused: Not used for answers until you resume in the settings panel.
  • Disabled / Deleted: The dataset is not in use or has been removed. Deleted datasets may be recoverable for a limited time from the recovery screen.

If replies look unchanged after an upload, check whether status is still Processing. Until Ready, the bot may keep older data or look empty.

3. Add Dataset panel

Add Dataset opens a panel on the right titled New Dataset. Clicking a list row opens Edit Dataset, with the status tag and pause/resume actions in the header.

Name and description

Name is required. Description is optional team notes.

Source type

Pick one of two cards:

  • Excel / CSV file upload: one-time .xlsx or .csv file
  • Google Sheets: connected sheet with optional auto-sync

After a dataset has a source, you cannot change the source type. Create a new dataset if you need the other source.

Free plans may block Google Sheets. Available sync intervals depend on the plan. Starter can connect without automatic sync, Pro supports intervals of one hour or longer, and Enterprise supports shorter intervals. Follow the tooltips in the panel.

New Dataset Processing
Shipping FAQ
Shipping and returns FAQ
Excel / CSV file uploadUpload file
Google SheetsConnect Google Sheets
Upload file
File nameshipping-faq.xlsx
Number of rows91
Columns5
Cancel Save

4. Add from Excel or CSV

  1. Choose Excel / CSV file upload.
  2. Click Upload file and pick .xlsx or .csv.
  3. Check file name, rows, and columns in the preview.
  4. Click Save to start processing.
  5. Wait until the list shows Ready.

Use the first sheet and include a header row. Save CSV files as UTF-8 (in Excel, choose CSV UTF-8). Processing will fail if rows have mismatched column counts, the dataset exceeds the plan's row limit, or cells exceed the per-language character limit.

In edit mode, View file content reopens the uploaded file contents. If another dataset is already processing, new uploads may be blocked. If tokens are under 10%, you may see a continue/cancel warning.

5. Connect Google Sheets

  1. Choose Google Sheets.
  2. Click Connect Google Sheets, sign in, and allow access. Unblock pop-ups if needed.
  3. Pick the sheet and set Auto Sync Interval. Depending on the plan, None may be the only available option.
  4. Use Sync Now for an immediate refresh.
  5. Click Save.

Only the first sheet is used. If the connection expires, reconnect when prompted. Sync uses tokens, so very short intervals can raise usage. The panel notice reminds you to watch token use when setting a schedule.

Edit Dataset Ready
Excel / CSV file uploadSource type cannot be changed after first upload.
Google Sheets
Auto Sync Interval
1 hour Sync Now

Pulls Google Sheets data on the schedule you set. Sync uses tokens.

6. Pause, delete, recover

Pause at the top of the panel stops the dataset from being used in answers. Resume turns it back on. These actions may be unavailable while a dataset is processing or after processing has failed.

Use the delete action in the settings panel. Deletion may be blocked while the dataset is processing. A deleted dataset can be recovered for 30 days; after that, it cannot be restored.

7. When something fails

  • Widget shows, answers empty: Check Ready status and whether the channel is still Pending.
  • Upload blocked: Another dataset may be processing; check extension, UTF-8, columns, and row limits.
  • Sheets will not connect: Plan limits, blocked pop-ups, expired auth, or not using the first sheet.
  • Synced but answers unchanged: Wait until status is Ready after Sync Now.

Use this order:

  1. Prepare a sheet with a template or prompt.
  2. Add Dataset with Excel/CSV or Google Sheets, then save.
  3. Wait for Ready in the list.
  4. Adjust sync or pause as needed.
  5. Confirm widget and profile in Channel Management.

After the learning data is ready, review the response settings in the FrontierBase Console.

View another guideFrontierBase →