Cognitive tools, relying on artificial intelligence (AI), are considered one of the most disruptive technologies in today’s business climate. These new automation and cognitive technologies represent an opportunity for organizations to increase productivity and effectiveness. But doing so will require rethinking and redesigning employees’ relationships with machines and ultimately, how work gets done.
In “The Rise of Cognitive Work (Re)design,” Deloitte shows how businesses would be best served adopting cognitive technologies in conjunction with an effort to rethink workflows and processes, as well as the main principles to account for when taking this approach.
Cognitive technologies represent an opportunity to kick start a renaissance in today’s business environment by reinventing how work gets done and paving the way for far-reaching impact. In a partial sense, we’ve experienced a similar inflection point before – as Mark Twain said, history doesn’t repeat itself, but it does rhyme. In the early 90s, a process change approach known as “business process reengineering” (BPR)[i] centered around information technologies as a driver of improvement in broad business processes. Deloitte believes that the rapid rise of data science, AI and cognitive technologies calls for a bold new approach to reinventing work. It is likely that some jobs will go away, others will emerge. And in many other cases, jobs will be transformed by tapping into the complementary abilities of humans and machines.
Beyond simply improving business processes, cognitive technologies enable fundamental shifts in human/computer collaboration. Today’s cognitive technologies enable both automation and human augmentation. They are based on a cluster of algorithmic and data-driven methods including robotic process automation (RPA), statistics and machine learning (ML) used to train (narrowly) “smart” algorithms, natural language processing (NLP) and generation (NLG), information retrieval technologies and recommendation engines. In the cases of tasks that are routine and/or associated with massive data trails (such as driving cars or answering common customer queries) algorithms can increasingly perform autonomously or near-autonomously. In other applications (such as making complex medical diagnoses, underwriting decisions, or judicial decisions), algorithms are well-suited to de-biasing and augmenting human judgment. In other words, data and algorithms enable forms of both AI as well as human-computer extended intelligence.
Cognitive technologies create and apply knowledge learned from data, making them proficient in data-intensive situations, like when there is too much data for the human brain to master or regulatory requirements need to be followed. They fit best where a substantial amount of knowledge is required to be effective because they function, and thrive, on information. The most capable systems, like IBM’s Watson, can “learn” to make better decisions as they get fed more data and can then recall previously ingested information.
One characteristic of cognitive technologies to note is that while they support a mix of skills and tasks, they do not replace entire processes. To maximize the impact of these technologies, an organization should systematically rethink the interaction between AI and humans, then redesign the underlying processes and tasks. Whether making a medical diagnosis or approving a loan, human experts can be taught to make better decisions as they consider the output of an algorithm along with their domain and institutional knowledge.
Deloitte believes that a lack of process changes to complement cognitive technologies has hindered the opportunities. They note that “cognitive technologies need a set of management structures and best implementation practices to yield the benefits of which they are capable.” They go on to argue that human-centered work redesign must complement cognitive technologies. They state that “smart technologies are unlikely to engender smart outcomes unless they are designed to promote smart adoption on the part of human end users.”[ii]
Thus far, with respect to using cognitive technologies, companies are just starting to scratch the surface—many are simply automating existing workflows. Deloitte argues that while this may be the quickest way to implement these tools, substantial gains can be made from augmenting the entire process and not just existing tasks.
This work redesign is an example of “design thinking,” which cognitive technology expert Manoj Saxena, chairman of Cognitive Scale and former GM of IBM Watson, believes is a useful implementation of cognitive technology.[iii] Design thinking can involve the design of products, strategies, facilities, or work processes.
Deloitte suggests applying the following principles in creating a synthetic approach to cognitive process redesign:
Understanding customer (end user) needs. Cognitive process designers should identify and consider the needs—both met and unmet—of the customer (or user who receives the output of the process). By understanding the process and the output, an improvement may become evident.
Work collaboratively. In addition to the technology experts, process experts, and customers, including the people who perform the actual processes is essential for making sure a redesign is successful. They can shed light on the procedures and workarounds necessary to meet customer needs, and will be more receptive to change if they have input.
Design iteratively and experimentally. Take an agile approach to cognitive work redesign by creating prototypes and pilots to assess different aspects of the design and break the design effort into stages; accomplishing a small part each week and scaling up later.
Keep the cognitive enablers in mind. Those in charge of the redesign should have a high level of familiarity with cognitive technology capabilities and use cases in order to connect customer needs with the technological possibilities.
Consider multiple alternatives. Since cognitive technology has many variations, it is valuable to think of multiple technologies and processes that can be tested against customer and process needs, as opposed to focusing solely on one particular design or tool.
Start with easy and relatively inexpensive problems. Going after the low-hanging fruit is a more successful strategy in cognitive redesigns because the cost of being wrong is lower. Larger, more important problems and decisions may still need human decision making while reengineering processes.
Author: Tom Davenport
[i] Thomas H. Davenport, Process Innovation: Reengineering Work through Information Technology (Harvard Business Review Press, 1993); Michael Hammer and James Champy, Reengineering the Corporation: A Manifesto for Business Revolution (HarperCollins, 1993).
[ii] James Guszcza. “Smarter together: Why artificial intelligence needs human-centered design” Deloitte Review, January 2018.
[iii] Personal communication from Vanguard spokesperson with Deloitte, May 5, 2017.