What is AI-Enabled Automation?
Artificial intelligence (AI) has created a new era of automation that is revolutionizing how work is done. This technology is already making our lives easier by automating tasks that were once difficult or time-consuming, such as scheduling appointments and data entry. But the potential for AI automation extends far beyond the mundane. With machine learning, natural language processing, computer vision, and predictive algorithms becoming more sophisticated every day, AI could soon be helping us tackle more complex business and operational workflows.
There are many repeated tasks, processes, and workflows in every job. Think processing invoices in finance and accounting, updating CRM records in sales, scheduling employees in operations, tracking inventory in merchandising, and so forth. Freeing up time from those tasks allows employees to focus on more strategic work. This benefits businesses with increased output and decreased costs, and increases morale amongst workers.
Over the past few years we have seen three distinct waves of AI enabled automation, each progressively demonstrating higher business impact. The evolution of these technologies has taken us from very simple transactional automation just a few years ago to self-learning business solutions as the state of the art today. Importantly, the evolution of AI-enabled automation solutions is congruent with Omega’s longstanding prediction that the most promising use cases of AI will involve people and AI working together.
The Evolution of AI-Enabled Automation
Wave 1: Basic Automation
The first wave of AI-enabled knowledge work software was characterized by a focus on discrete transactions based on mimicking the actions of a human. Simplistic chatbots, often using templatized options and keyword matching, were among the earliest solutions. Generally these approaches provided more novelty than utility, deploying rules-based logic rather than real AI. Given the mostly negative reception, the phrase chatbot has come to be perceived as slightly pejorative and companies developing advanced solutions now prefer the term ‘digital assistant.’
Wave 2: Robotic Process Automation (RPA)
Next came Robotic Process Automation (RPA), which focuses on automating the operational flows of knowledge work. RPA solutions leverage process mining software, which observes what a knowledge worker does on a computer including specific mouse clicks, cursor movements, text input, and data transfer, and then aims to mimic and automate this sequence of actions. Even though companies such as UI Path demonstrated the ability to build big businesses based on this approach, in practice RPA can be brittle and requires a heavy lift in implementation. Clients often need to work with professional services firms to map out use cases, build the solutions, and integrate the technology with other vendors’ software.
With RPA, companies have central control over the technology which facilitates governance over its actions and access. Additionally, most RPA solutions integrate with commonly used enterprise systems of record, such as CRM (Customer Relationship Management), ERP (Enterprise Resource Planning), HCM (Human Capital Management), and PLM (Product Lifecycle Management) systems. In practice, RPA software is limited to high-volume, predicable, and repeatable use cases where discrete actions can be defined and templatized in a sequence.
Wave 3: Digital Workers for Advanced Automation
Today, we are seeing a third generation of ‘digital worker’ automation solutions come to market, which focus on lower friction adoption and offer greater versatility in use cases. A defining characteristic of these next-gen solutions is that they seek to automate discrete tasks that are typically associated with specific roles and functions within a business, but embrace a human in the loop for more complex and nuanced challenges.
Another characteristic of digital worker solutions is the native ability to process unstructured data to learn and improve over time. These solutions not only integrate with business systems of record but also into business systems of engagement, such as email, telephony / voice, messaging, documents, notes, and spreadsheets. These solutions often feature pretrained models giving them out-of-the-box value, combined with the ability to learn and improve as they ingest additional unstructured information.
This third wave of intelligent automation shifts the automation paradigm from replicating human actions for specific sequence flows to the higher-order concept of automating a business task or objective, with contextual understanding as well as failover to human handlers for exception handling. These solutions do not try to promise 100% automation for anything other than the most simple use cases. As a result, they more accurately reflect the complexity of the real world, recognizing that even as the AI improves over time, there will be new scenarios that require human intervention. The real world regularly presents novel situations that require improvisation and adaptiveness; simply put, AI-enabled automation systems cannot anticipate, nor be pre-trained for, every possible real world scenario. (A fascinating emerging area that is narrowing the gap involves AI training itself on imaginary scenarios through the use of adversarial neural networks, where an AI system constantly practices against an adversarial AI. Over time, this creates headroom for AI systems to get progressively better).
Augmentation Is the Future
There is little doubt that AI-enabled automation will play an ever-expanding role in the future of work. The choice of which solution to use comes down to business use cases, priorities, and existing technology stack. We believe that automation will be an integral part of companies’ strategy moving forward, but we also believe that automation will be complementary to human judgment for the foreseeable future.
Companies and their workers will welcome solutions that lighten the tedium of repetitive work but will do so in a context where such solutions serve to empower people to focus more on strategic endeavors that require judgment, creativity, intuition, common sense, and insight.
The future of work involves people and AI working together to achieve more than what each can achieve on a standalone basis. For more articles like this one, visit Omega’s website.
##########
