A new era for AI-powered AP automation

Getting information from a supplier’s invoice into a buyer’s accounting or ERP system might sound straightforward. But for many companies, achieving this is a major struggle.

In the past, invoice processing meant collecting paper invoices and manually keying the information into the buyer’s system. While this may have got the job done, it is a time-consuming and inefficient approach, and one that is prone to human error.

Technology may have moved on, but for most companies, invoice automation remains a distant goal. The PDF – a universal format favored by many suppliers – is easy to produce. But it is also a relatively unstructured format, and one that accounting systems typically struggle to ingest. As a result, buyers may still need to re-key data from electronic invoices to get it into the right place.

Existing solutions

The need for a more efficient approach is clear. But while electronic channels have brought new ways of getting invoicing data into buyers’ systems, these can have their own shortcomings for some businesses:

• EDI integration – One of the more sophisticated options is to create full EDI integration between the buyer and supplier’s accounting systems. But this can be expensive and time-consuming, and only really works for large businesses that have a significant number of invoices to process. It is also dependent on suppliers adhering to strict parameters when submitting invoices, which for multiple buyer relationships can be a difficult and time-consuming process for suppliers to maintain.
• Supplier web portal – Another option is to provide a web portal that suppliers can use to key in their own information. However, this simply shifts the burden for inputting data from the buyer to the supplier and only suits those who have a small number of invoices to process.

More compelling for businesses who don’t process enough invoices for EDI integration – and for those who process too many invoices for a supplier web portal – is the prospect of capturing the data contained in supplier invoices automatically and transferring it seamlessly into the buyer’s system.

Optical Character Recognition (OCR) goes some way towards achieving this – but is only usually successful if the relevant information is positioned in narrowly defined locations within the invoice file. At scale, compliance with the parameters for OCR to work is troublesome, meaning the resulting data often needs to be ‘fixed’ using manual intervention – another time-consuming process. It’s clear that a better solution is required to streamline the invoicing process.

Harnessing AI for invoice automation

The challenges may be considerable – but help is at hand. In 2019, collaboration between engineering teams at Taulia and Google developed a ground-breaking AI-powered AP automation solution, Cognitive Invoicing, which can autonomously extract data from any type of invoice.

The solution combines two powerful tools to solve the invoice automation challenge. On the one side is Taulia’s intelligent platform, which has a network of over two million suppliers and handles over a million supplier interactions each day. On the other is Google’s Document Understanding AI – a platform which uses machine learning to help companies process and analyze documents.

Cognitive Invoicing takes full advantage of the integrated OCR capabilities that come with Document Understanding AI. The new solution will be able to take data from 100% of suppliers’ invoices – simplifying the submission process for suppliers and eliminating the need for manual handling.

The AI engine will be able to parse invoices and extract data accurately from any format, including machine-generated PDFs and scanned image PDFs, and send the data to Taulia for end-stage processing. What’s more, as the solution processes an ever-wider variety of invoice layouts, machine learning will enable it to get smarter every time.

Invoice automation and collaborative exceptions

The solution will be able to capture 100% of unstructured invoices. But there will still be times where fixes are needed – for example, if information is missing from the file itself. For any exceptions that do occur, Cognitive Invoicing offers further opportunities to streamline processes.

The answer is a cloud-based collaborative resolution process that draws on Taulia’s expertise in facilitating collaboration between buyers and suppliers. Instead of relying on time-consuming exception handling processes, companies will be able to fix any issues in collaboration with their suppliers – or, if relevant, with a third party BPO center.

Of course, not all exceptions are alike. Buyers will also be able to create rules determining how exceptions should be handled in a variety of different situations:

• Some buyers may prefer to pass all invoices back to suppliers so they can add missing data, thereby freeing up the buyer’s AP team from manual handling.
• For certain suppliers, the buyer’s AP team may step in and complete the validation to provide the smoothest possible supplier experience.
• Buyers may even opt for more complex rules. For example, they might state that if 95% of an invoice’s fields are in place, the remainder should be corrected by the AP team.

What are the benefits of invoice automation?

The benefits of invoice automation, and more specifically implementation of Taulia’s Cognitive Invoicing, are unprecedented opportunities for buyers to streamline the processes involved in capturing invoice data – saving time, reducing manual effort, and reducing the risk of error. Meanwhile, suppliers may no longer need to key information into a dedicated portal, or field avoidable queries.

The collaborative resolution process included in the solution means that buyers will be able to reduce the amount of time their AP teams spend correcting errors and tracking down missing data. What’s more, existing Taulia customers will be able to benefit from the new solution with minimal additional work.

Last but not least, the new solution will give companies the opportunity to take fuller advantage of working capital solutions. Companies can use solutions like supply chain finance and dynamic discounting to pay their suppliers early, thereby boosting their working capital or deploying excess cash more effectively. But if it takes weeks for a company to process suppliers’ invoices, this can eat into the possible gains. The faster invoices can be processed, the sooner suppliers can be paid – and the greater the opportunity to free up cash across the supply chain.

Given these benefits, we believe our AI-powered Cognitive Invoicing is a game-changing solution that will make it easy for companies to get data out of supplier invoices and into their own systems – while also supporting Taulia’s wider adoption of AI to support decision-making and strengthen buyer-supplier relationships.