Configurations and Transaction Manager Settings

Contents

Configurations tab

Within the Configurations tab of Transaction Manager, there are two sections: Automation Settings and Additional Automation Settings

Automation Settings

In this section, you will map a Smart Connect account to a GL account from the QuickBooks Online (QBO) or Xero Chart of Accounts.

Once the account is mapped, you will choose from AutoPush, ManualPush, or No Automation.

AutoPush

AutoPush automatically posts transactions from the Smart Connect bank feeds into QuickBooks Online or Xero. Using our Machine Learning models, we predict the payee and category for a transaction. Any transactions that are low or medium confidence will automatically appear on the Needs Review tab for your review.

• Last Updated By column
  This column provides audit visibility into the most recent changes to automation settings, showing who made the update and when. The column displays updates made directly in the Configurations tab, as well as through the Add Client Flow and Smart Connect.

Low-confidence transactions will have a red AutoPush icon next to the Category and/or Payee fields; medium-confidence transactions will have an orange AutoPush icon next to the Category and/orPayee fields.

High-confidence transactions will appear on the Processed tab and will have a green AutoPush icon next to the Category and/or Payee fields. If you hover over the AutoPush icon, you will see a tooltip with additional details.

You can also see the confidence level in the transaction details window by clicking the Action button, and then the Details button.

Knowing when AutoPush is Enabled
Accounts connected to AutoPush display an AutoPush icon inline with the financial institution name. Upon expanding the financial institution line, any account in that financial institution connected to Smart Connect will be identified with the same icon.

Additionally, you can see if AutoPush or ManualPush is enabled by opening the configurations tab. The radio buttons will indicate which is on:

When adding new entities

Once the first AutoPush/ManualPush transactions are ready for review, the Partner Admin who connected the GL, the Client POC, and the TxM Default Assignee will all receive an email notification.

The start date will be when transactions should start “AutoPushing” to the GL. The AutoPush start date can be set to cover transactions up to 24 months in the past.

ManualPush

Once configured, manual push holds all transactions in the Needs Review tab of Transaction Manager as bank feed transactions. These transactions will be automatically categorized, but will not be automatically posted to the GL, regardless of their confidence rating. They will need to be reviewed by a user and manually pushed to the GL. 

After the user has reviewed the transaction they will need to take the action to Add to GL & Mark Reviewed, or to Add Transactions to GL (which adds the transaction to the GL without removing it from the Needs Review tab).  This can be done on individual transactions, or in bulk by using the checkboxes to the left of each transaction and clicking on the Other Actions bulk button at the top of the transactions list.

1. Go to the Configurations tab in Transaction Manager.
2. Click the Edit button in the upper-right corner of Automation Settings:
   
   
3. Select the ManualPush radio button under Push to GL.
   
   
4. Click Save.

When adding new entities

Once the first AutoPush/ManualPush transactions are ready for review, the Partner Admin who connected the GL, the Client POC, and the TxM Default Assignee will all receive an email notification.

No Automation

We won't automatically categorize or post transactions to the GL. But you can still use Transaction Manager to collaborate on categorizing uncategorized transactions with your team and clients. To enable, simply check the box under No Automation.

Additional Automation Settings

The Additional Automation Settings section is where you can configure additional settings, such as: ML Training Dates, GL Automation, Statement Automation, Materiality Threshold, Default Assignee and Default Settings for Low/Medium Confidence Transactions. 

• The “ML Training Dates" Start Date allows you to set a specific date when the ML models should start training based on the historical GL data. If no date is defined, the models will look at all GL activity. This is useful in case there’s historical data we do not want the model to train from. 

  • Users may choose both a start and end date to create a specific range of data to train machine learning.

  • When the start or end dates for the machine learning models are changed, the models will retrain immediately. Predictions based on the retrained model will begin within 24 hours.

• The GL Automation setting is the recommended approach for accounts not connected in Smart Connect. All transactions posted to these accounts will be processed daily. Any transactions requiring review will appear on the Needs Review tab. To ensure processing, transactions must be added to the GL. To learn more, jump to GL Automation.

• The Statement Automation setting is the perfect option when a PDF statement is the only way to obtain transactions. It automatically extracts transactions from the uploaded statement, so they can be added to the GL and categorized by our AI. If a match is found, no new transaction will be posted.

• • A GL holding account must be assigned. The extracted statement transactions will be posted there temporarily before the AI processes them. You can select one in the Statement Automation Holding Account field.
  • PDF statements must be uploaded to the appropriate “Bank Statements” folder in the client’s account—either directly from the folder or via document upload task.
  • Transactions will appear for review within 24–48 hours in Transaction Manager.
  • Please note:
    • As transactions are processed, the account is reconciled in Auto Bank Rec using our standard matching logic, automating the full transaction lifecycle.
    • When extraction confidence is low, the Botkeeper team will manually review transactions before they are posted.

• The Uncategorized Accounts setting allows you to exclude accounts from machine learning to help improve the accuracy of your future automated transaction categorizations.

• The Materiality Threshold allows you to set an amount where all transactions greater than or equal to a certain amount will automatically post to the Needs Review tab for further review. 

• The Default Assignee Settings allows you to select a user that should be assigned to all transactions on the Needs Review tab.

• The Settings for Low/Medium Confidence Transactions allows you to select a default category and payee to which all low and medium confidence transactions should be posted. For example, you can post all of these transactions to “Uncategorized Expense” and “Uncategorized Vendor.” If these settings are not configured, the transactions will post to the category and payee that the ML model predicted.

When each automation type runs

Choosing the right Automation

Option 1: GL Automation (If the accounts are already connected in the QBO/Xero GL Bank Feed, then this is the preferred method)

1. Account Connected to GL (QBO/Xero) Bank Feed
   1. Credentials are required to establish a connection in the GL bank feed. This can either be third-party credentials the client sets up for the firm, or the client’s primary credentials.
      1. Note: Some institutions (Bank of America for example) do not allow the third-party credentials to connect. Connections must use the client’s primary credentials.
   2. Note: If the firm does not have the credentials and you do not want the client to set up the connection in the GL, you will need to choose another option.
   3. Note: If you still wish to use GL Automation, but the account isn’t connected in the GL (or can’t be connected), you may upload a .csv file of transactions to the GL:
      1. Firm or their client will need access to the transactions in .csv format so they can be imported into GL. 
      2. This requires credentials to access the bank. This can either be third-party credentials the client sets up for the firm, or the client’s primary credentials. If access is unavailable, the client will need to provide the .csv file.
      3. Note: If the firm does not have the credentials and they do not want the client to provide the .csv file, they will need to choose another option.

Once GL Automation is enabled, Botkeeper will ingest new transactions from the GL every day at 7am ET and categorize them. Any transactions needing review will be on the Needs Review tab in Transaction Manager, where they can be updated. Users can then select Mark Reviewed to sync them back to the GL.

Option 2: Smart Connect (If the accounts are NOT connected or can’t be connected in the GL Bank Feed, then this is the preferred method)

• The institution needs to be supported by Mastercard Data Connect.
• Credentials are required to establish the connection. These can be primary or third-party credentials set up by the client for the firm’s use. 
  • Note: Some institutions (Bank of America for example) do not allow the third-party credentials to connect. Connections must use the client’s primary credentials. Major Institutions with Specific Setup Suggestions
• If the client doesn't wish to share their credentials with the firm, there are two options to establish a Smart Connect connection:

1. Invite client into the Botkeeper platform as a user so they can establish the connection themselves; OR
2. Send the client a Smart Link (No Botkeeper login required) so they can establish the connection themselves.

Once Smart Connect is connected and automation is enabled, Botkeeper will download new transactions from Mastercard Data Connect every day at 7 am ET and categorize them. Any transactions needing review will be on the Needs Review tab in Transaction Manager, where they can be updated. Users can then select Mark Reviewed to sync them back to the GL.

Option 3: Statement Automation (Preferred method when accounts aren't, or can't be connected in the GL or Smart Connect)

• The firm or their client will need access to the PDF bank statement
  • This requires credentials to access the bank (third-party or read-only). If access is unavailable, the client will need to provide the PDF statement.
  • Note: If the firm does not have credentials and they do not want the client to provide the PDF statement, they will need to choose another option.

Once a statement is uploaded to the account’s Bank Statements folder (either via direct upload or a document upload task) Botkeeper’s OCR extracts the transactions and posts them to the designated GL holding account. Transactions remain there briefly while the AI processes them, beginning immediately after the statement is processed. The entire process is completed within 24 hours of the statement being uploaded. Once complete, transactions will appear in the Needs Review tab of Transaction Manager, where you can review and edit them as needed.

Some scenarios for choosing your connection:

1. Account can be connected in Smart Connect, but is already connected in the GL via bank connection
   The options listed below are in order of preference:
   1. Option 1: Use GL Automation for transaction categorization and Smart Connect for statement downloads ONLY. Take care when using this approach NOT to enable transaction categorization in the Transaction Manager Configurations tab, or duplicates might result.
   2. Option 2: Use GL Automation for transaction categorization and don’t connect Smart Connect*
   3. Option 3: Connect Smart Connect for transaction categorization and statement downloads. This configuration requires the GL Bank Feed(s) to be disconnected.
2. Account can be connected in Smart Connect for transactions, but not statements (and is not currently connected in the GL). Connect Smart Connect for transaction categorization (Option 2 from the section above).*
   *To use Auto Bank Rec, you will need to manually upload the statement to the corresponding bank statement folder within the Documents module.
3. Account can’t be connected in Smart Connect and is currently connected in the GL via bank connection.
   Use GL Automation for transaction categorization (Option 1:1 from the section above).*
   *To use Auto Bank Rec, you will need to manually upload the statement to the corresponding bank statement folder within the Documents module.
4. Account can’t be connected in Smart Connect, and is currently being imported into the GL via CSV uploads
   Use GL Automation for transaction categorization (Option 1:2 from the section above)*
   *To use Auto Bank Rec, you will need to manually upload the statement to the corresponding bank statement folder within the Documents module.
5. Account can be connected in Smart Connect and is currently being imported into the GL via .CSV imports
   The options listed below are in order of preference:
   1. Option 1 - Connect Smart Connect for transaction categorization and statement downloads.
   2. Option 2 - Use GL Automation for transaction categorization and Smart Connect for statement downloads.*
   3. Option 3 - Use GL Automation for transaction categorization and don’t connect Smart Connect. * **
      *.CSV imports are highly manual; we always recommend prioritizing a connection over an import.
      **To use Auto Bank Rec, you will need to manually upload the statement to the corresponding bank statement folder within the Documents module.

Connection Quick Reference

Train Your Model Tab

Editing, Deactivating, or Deleting Logic

The Action button (three dots) on the right side of the logic's line will open the ability to edit, deactivate, or delete that logic:

Edit logic opens the full settings for your logic so you can make changes:

Deactivate logic retains your logic for later use, but turns it off. Deactivated logic will not execute once this is selected, until/unless it is reactivated.

Delete logic removes the logic from your list of saved options. It cannot be recovered.

Finding transactions affected by Train Your Model

On the Needs Review tab, you'll find a filter allowing you to select just the transactions where logic was applied, as well as one excluding transactions in which logic was applied:

Transactions benefitting from Train Your Model logic have a green lightbulb indicator next to the attribute defined by the logic: