For enterprises the path to artificial intelligence is through robotic process automation

Today most AI applications are focused on narrowly defined tasks, such as predicting machine failure rates, text analytics for sentiment detection, or facial image recognition. However, leading companies are leveraging AI to make the most of all of the data that is available to them by adding the key element of prediction. This means that the typical process sequence is one of Data -> Judgement -> Action. Understanding the complexities and challenges of these steps will be critical when it comes to enterprise automation.

Data

Data is the most valued currency in business, and it can be classified into structured data, unstructured data, and conversational data.

Achieving predictive capabilities in automation involves securing the right data and converting it into meaningful information or insights for further processing. A typical process requires data-extraction, data-transformation, and data-cleansing.

Unstructured data

Similar to spoken language, unstructured data is difficult, even impossible, to interpret by algorithms, and the potential to achieve zero-touch operations lies in a company’s ability to handle it. This class of unstructured data further consists of subgroups such as images in documents, texts, images in picture form, audio, and video. Each might pose different challenges or technical solutions when it comes to extraction.

With such a wide range of unstructured data sources, it is hardly surprising that some predict that by the year 2020, more than 90 per cent of all data in the enterprise will be unstructured.

Unstructured documents require optical character or intelligent character recognition capabilities (OCR/ICR) to extract the data. If an image has a consistent format, such as payable invoices, payment remittance and so forth, the images can be converted using OCR/ICR, and the output will be readily consumable by the downstream process. If the format is inconsistent, then OCR/ICR technologies will deliver unstructured text data, which needs further processing.

Unstructured text requires natural language processing technologies to interpret the different attributes that are relevant to understanding the data, such as context, entities, person, place, and so on. Unstructured documents require interpretation with vision technologies to extract information. One example would be an engineering diagram of a building that needs to be converted into a bill of material rapidly due to the competitive nature of the bid process.

Unstructured audio has value because it helps companies in particular scenarios, such as analysing customer calls to understand satisfaction-levels.

Judgment

The next step in cognitive automation is judgment, which involves combining information with past trends and rules to decide on a course of action. It can be easily split into two types: rules-based judgment and trend-based judgment.

Rules-based judgment involves decision-making based on configurable rules. For example, a payable invoice is compliant if it has a set of key information present. These rules can easily be configured to deliver touch-free automation. Trend-based judgment involves decision-making based on past patterns, such as the decision to write off short payments from customers.

Humans are still important however, playing a vital role when much of the fuzzy decision-making in an enterprise process can be special events such as marketing campaigns, period-ends, or cash positions. These can call for intuition-based decision-making that can be learned through experience, but cannot be documented as rules. Human intervention is required to make such decisions fast and accountable.

While many of the trend-based judgment decisions will need human input, AI will reduce the need for some processing exceptions by predicting the best decision. These predictions can be automated or may need human-in-the-loop when the confidence level in an organisation does not meet the threshold for automation.

Action

The last step in the process involves taking an action based on the outcome of the first two steps. Action can be system-based, such as automatic data-transfer, data-processing, or email communication, or it can be a presentation at a meeting.

Digital transformation achieved through RPA

Today, however, most initiatives tied to RPA are tactical and are focused on cost-cutting.

They should not be. Business processes can be transformed through the introduction of cognitive automation within RPA technology.

Investment in image-recognition and the incorporation of deep learning enable the robots to understand any screen, similar to the way humans do. An image-recognition engine uses powerful algorithms that are optimised to find images on screen in under 100 milliseconds.

Machine learning increases the confidence in processing of a variety of semi-structured and unstructured content, where reasoning is prevalent. Emails, annual statements, contracts, or other types of documents hold data that can be extracted through keywords and logically organised for robots to drive decisions, with predictive decisioning technology which has self-learning capabilities.

RPA’s integration with the technologies behind natural language understanding and natural language generation also has significant consequences. With 20 per cent of the searches performed with mobile being voice-based, conversational interactions are set to become increasingly pervasive, even in an enterprise context.

For any organisation to achieve true digital transformation, a balance must be struck between the best technologies available. As we have seen, they are many and varied. But RPA can be the platform to introduce them one- by -one where they can be managed easily in one place. This is the simplest and most effective way of implementing them with the additional benefit that transformational efficiencies can be delivered in weeks. In summary, RPA is the perfect enabler of cognitive capabilities at the enterprise level.

-End-