Recently, in this place, I reported in the first part of my blog about my experience with artificial intelligence in everyday life. Today, I look at the topic more closely in terms of accounting processing.
Intelligent incoming mail processing
If an input channel is not used exclusively for incoming invoices, but also serves as an input gate for other information, we are dealing with the so-called mailbox or digital mail-room solutions. Here, too, AI-based solutions have been used for a long time. Classification of documents is no longer based purely on the recognition of key terms or isolated, formal criteria. A semantic analysis unit analyses and “understands” the content of a letter or an e-mail to the extent that not only the type of information is reliably detected, but also all relevant information is available for further processing. It is obvious that by reducing manual steps, process speed increases and costs can be saved.
The most developed systems not only capture the information received, but also recognise the sender’s mood (mood analysis). This information particularly supports applications such as complaints management and claims processing. Related techniques, for example, are already analysing certain channels in social media continually, and are predicting the success or failure of products with high accuracy and at a very early stage.
Speech control helps
When we look at solutions for accounting in practice, we find that despite all digital support auditing and approving invoices is often a time-consuming part of the overall process. Frequently, the reason is that the responsible persons are often and long in transit. The above-mentioned specialised apps for processing on mobile end devices already provide noticeable relief.
Further solutions from the field of AI are used here. From the private sector, we have known assistance systems such as Alexa, Google Assistant or Siri for some time. If we then use the ability of such systems to provide information in natural language meaningfully and respond to spoken instructions, it closes a further gap in companies’ financial processes. On the road and, for example, on car trips, between starting a speech-controlled play list and retrieving e-mails, what is there against also releasing some invoices and thus securing the cash discount savings?
Systems learn independently
In the near future, we will work with intelligent IT systems as intuitively as today with human assistants. Intelligent software systems will relieve us of today’s many, still manual or semi-manual steps, and communication will take place wherever possible through natural language. The systems are self-taught and support users actively.
By analysing large amounts of data (big data), transparency of available information increases and trends are detected at an early stage. Intelligent assistants use this information, network it in real time with numerous other internal and external sources of information, know their users’ needs, and actively assist users in their activities and daily decisions.