Though individual countries may have unique, country-specific driving rules, one thing is for certain across the board—red means stop, and green means go.

And that’s just how simple we made it to use our new machine learning, purchasing feature, Approval Confidence Index.

Our Approval Confidence Index tool uses machine learning (ML) to take a look at all your historical data to make a prediction based on your transactional data.

Approval Confidence Index in action

When users create a purchase requisition, each item they add to the requisition will be assigned an Approval Confidence Index Score, ranking that item from 1-10, with 10 being the highest probability of that item being approved for purchase. Based on past transactions of this same item, machine learning uses all the data running through your solution to give the selected item a Confidence Index Score.

If the score is in the red (lower numbers) that means that the item has a higher likelihood of not being approved. However, if the ML tells the user that the items scored in the green (a higher number), the item will likely be approved.

For Users

It’s as easy as 4 steps:

1. Search for your product

2. Make a selection

3. Review Approval Confidence Index Score

4. Submit request if in the green, search for another item if in the red or risk it and still submit the requisition.

Say a user logs into their e-procurement system to purchase a new mouse for their laptop. Out of all the vendors offering laptop mice for sale, the user selects Mouse A. Once selected, the ML informs the user of Mouse A’s Approval Confidence Index Score. If, for example, the score comes back as a 9 (and very much in the green), the user knows that Mouse A is a safe option. But, if the user selects Mouse B (a mouse that has never been purchased or one that has gold bells and whistles), the Approval Confidence Index Score may come back as a 2 or a 1, placing this item in the red zone and as an unlikely approval.

With this information, the user can make a decision - continue their product search until they find a mouse with a higher score and a greener result or go ahead and submit the requisition and let their manager approve or reject the requisition. This guides users through the purchasing process and minimises maverick spending. Plus, they’ll save time not having to go through the whole process again after the requisition is denied.

For Approvers

The steps are equally as easy:

1. Review request’s Approval Confidence Index Score

2. If request is in the green, submit. If request is in the red, review.

On the approvers side, it’s quick and easy to take a glance at the Approval Confidence Index Score. If an approver is swamped with a long waiting list of transactions, they can use the Approval Confidence Index Score and the corresponding colors to quickly decide if a transaction should be approved or not.

There’s no need to waste time reviewing easy green transactions. You can quickly approve these and address the less “confident” transactions, instead. And once you become more comfortable using this ML to inform your decisions, you can drive high approval rating transactions automatically, without needing human intervention.

More Data, More Approvals

As with all ML and artificial intelligence (AI) in purchase-to-pay, the more data you have running through your automated solutions, the better the results. Once a critical mass of transactions has processed through your system, the data will be available to turn this feature on and begin formulating confidence scores. And the more data you have, the better the approval confidence index becomes, and the more ML can take over the manual approval process. What does this lead to? Quicker transaction times, less manual time wasted on approvals, and an intuitive method for users to submit purchase requests.

Ready to learn more?

Learn more about how data-based tools affect every part of the purchase-to-pay process in our whitepaper.  Questions about the latest Approval Confidence Index feature? Contact us!