Assistants (AI powered)

XLReporting offers several Assistants which are powered by artificial intelligence to help you in working with your data. The assistants enable you to type in your requests in natural English, and use artificial intelligence to help you transform and improve your data.

Terms of Use

We use OpenAI and its most recent "gpt-4" version. The requests and data that you submit may be kept for no more than 30 days, but only to identify abuse and support issues.

• Your data will never be used to train or improve OpenAI's models or XLReporting models.
• There is no "memory" of any previous requests, each request will be processed as a single isolated interaction, without any awareness of you, your tenant, or any other requests.
• Use of Assistants is conditional to a special permission "Use assistants" in User roles, giving you granular control to decide which of your users are allowed to use this function.
• Basic requests (clear/trim/strip/copy/move/swap/merge/mask, change to lower/upper/proper case, lookup reporting codes) are performed by XLReporting without interaction with OpenAI. All other requests are sent to OpenAI for processing.
• If you want to prevent any processing by OpenAI for your tenant, you can set the security option Disable OpenAI requests in Manage your Profile. Users will still be able to use the Assistant for basic requests, but any requests that require OpenAI processing will be denied.
• Use of OpenAI requests is included in your license, but is subject to our Fair Use Policy with regards to the type of requests you can make. Please note that requests may be monitored.
• Standard subscriptions include a limit of 50 OpenAI requests per tenant per day. Extra charges will apply if you are sending more. Contact us for details if necessary.
• You should take care to always check the results yourself. Interpretation by artificial intelligence of your request and your data may not always be accurate, correct, or consistent with your expectations.

Data assistant

The data assistant is available in Define Data sets and Define Imports via the Actions menu, provided you have the relevant permission in your user role. You can use it to transform column data in your data set. You can type your request in natural English (or select from the list of common requests) and select the source column and the target column in the data set. The assistant will process the source column in all rows in your data set, and will populate the result into the target column of every row. Source and target may be the same column, in which case the contents of the source column will be overwritten by the results. You can visually inspect the data before you decide to save it.

The requests clear, trim, strip, copy, move, swap, merge, mask, and change are standard requests for simple transformations of column data. These requests are pre-programmed and are not processed by artificial intelligence.

The request lookup reporting codes performs a specific purpose: based on the account descriptions of your Chart of Accounts (the source column) it will lookup the standardized reporting codes that are used in our Essentials templates. We have trained our model on a large variety of different account descriptions.

Another approach to this, as an example of a simple and generic request, could be "match to: asset, liability, equity, revenue, salaries, expenses". Based on the texts and descriptions in your source column, each row will be matched (as near as possible) to any of the items you specify.

Another example request could be "find synonyms", which will instruct OpenAI as follows: "find synonyms for every item in this list". Only the data in the selected source column will be processed by artificial intelligence, and no data in any other column.

Example requests:

clear
trim
strip
copy
move
swap
merge
mask
change to lowercase
change to uppercase
change to proper case
lookup reporting codes

translate to english (or any language)
match to: asset, liability, equity, revenue, salaries, expenses

find synonyms
find definition
find currency code
list random number
list random person names
increment with 1, starting at 10
multiply with 100
round to 1 decimal
split on comma and list first element
add "XX"
prefix "YY"
remove "ZZ"