Cognitive computing will become a widely used analog for human data input to provide us with solutions in the same way we learn. How will you use this tech?
Cognitive computing, which mimics how humans learn, think and adapt, will enable a wide range of real-time insights and action.
Eventually, it will replicate human sensory perception, deduction, learning, thinking and decision-making. Enabled by vast amounts of computing power, cognitive computing technology will be able to perceive patterns humans may not be able to perceive and provide solutions.
The field is exploding with potential and promise—as well as with practical applications that can be deployed right now for immediate benefit. Taking a heuristic approach to adapting and adopting cognitive computing technologies will help organizations in two key ways: First, they can identify applications that will deliver immediate return on investment (ROI) and second, by implementing these mature applications and exploring emergent technologies, organizations will gain the knowledge and experience required to transform and thrive in a business landscape that will be completely disrupted by cognitive computing.
Step 1: Categorize the technology landscape across a maturity continuum
When a technology begins to proliferate, deciding where to start with it is like trying to decide what dishes to select at a large buffet. This is especially true in cognitive computing, where a variety of technologies are emerging. These include advanced Optical Character Recognition (OCR) technology; chatbots, robots and intelligent virtual assistants; voice and video analytics; and image analytics.
The first step in making the decision about which technology to focus on is to evaluate its maturity. Has it gone mainstream and is it delivering value? Or is it just on the cusp of that success? Or, is it utterly embryonic, new and promising, but not proven? Technologies that are highly mature are likely to be ready for adoption to address immediate opportunities.
Another key factor is your company’s appetite for cognitive technology. The same conversational AI offering that appears emergent for a bank to use may be considered mature by an entertainment company.
With performance and appetite assessed, it’s on to opportunity.
Step 2: Identify immediate opportunities for adoption
Look for immediate opportunities for deploying cognitive computing solutions that emphasize ROI. The productivity and/or financial returns from these solutions will help free up resources and funding for longer-term projects such as cognitive automation. The goal is to have a high success rate with initial projects and realize value fast, in approximately six months to a year.
Practical use cases for cognitive computing with quick ROI are emerging across a range of industries, including:
Banking, financial services, and insurance: Machine learning is improving automated systems that monitor, track and report suspicious activities to detect, predict and avoid fraud. Robo-advisors help customers make decisions and generate personalized offerings based on past behavior patterns. Cognitive technologies automate document processing, contract management, and other processes.
Healthcare: Deep learning systems and neural networks are learning from diagnostic images and creating systems of intelligence that can analyze medical images, predict risk and recommend whether intervention is necessary.
Travel and hospitality: Applications use machine learning to personalize travel deals and promotions; chatbots collect data and provide suggestions to travelers.
Retail: Cognitive computing apps can act as personal shoppers, making recommendations based on a shopper’s previous purchases, interests, and needs.
Manufacturing: Machine learning can help optimize production capacity, supply chain management, and pricing and sift through data to ensure finance, operations, and supply chain teams have the right data for better collaboration.
The optimal opportunities for quickly realizing returns are those where it will be relatively easy to plug in the technology. It’s analogous to driverless cars: when these are commercially viable, the autonomous vehicles should integrate well in the U.S., with its wide-open roads, traffic control infrastructure, network of departments of motor vehicles. and legislators already at work on legislation. That’s as compared to a country with significant congestion, such as India, where deploying driverless cars today would simply increase congestion.
Vision is important too: what is your end goal? A search company can see a driverless car as the ultimate search tool, one that can automatically take you to the location of a search result, such as a restaurant or medical facility. An initial deployment of cognitive computing capabilities may be to streamline a single process, such as processing a healthcare claim; that process is likely to be part of a broader operation or experience that can be steadily transformed by cognitive computing, such as providing real-time guidance to healthcare consumers about their medical treatment decisions.
Calculating the potential ROI will help get buy-in from management. The ROI can help fund other deployments of cognitive computing opportunities, and that fact should be part of the business case. If you implement automation for a business process that costs you $10 million and get a saving of 30 percent, then you have freed up $3 million for other projects.
Step 3: Pilot and prototype emerging areas
While deploying the quick-win applications, also plan for how mature and evolving technologies could work together to transform the business. Then the organization can layer the goals it plans to achieve via cognitive computing over time. These applications, which may bear fruit in one to three years, should be tested with pilots and prototypes. Developing and deploying virtual assistants with conversational interfaces and using deep learning to create learning systems that can scale are prime opportunities. Building proofs of concept can help organizations understand how these applications may disrupt the organization and develop strategies to manage that transformation.
Candidates for pilots may include areas that have not been automated in the past but that cognitive computing can tackle such as voice-enabled assistants and advisors.
Step 4: Assess moonshot projects
With successes in hand, the organization can begin to envision radically different ways of operating and turn core competencies on their heads. A manufacturing company becomes a technology company; a transactional company becomes a connectivity company. The technologies that will make moonshots possible will be on a three- to five-year trajectory for achieving maturity, with academia their likely source. Remain current on technology developments and assess how they may complement cognitive applications deployed at immediate and intermediate stages.
To come up with moonshot answers to “what if?” queries, organizations need a mix of hardcore technologists and people skilled in divergent thinking. These experts and thinkers should be tasked with developing radically different visions that can be transformed into pilots, prototypes, and proof of concepts.
Recognize that fostering cognitive computing adoption requires looking outside the organization for talent, technology, and innovations. The market moves faster than any one company can. Being part of an innovation ecosystem, such as an academic-business consortium, will enable an organization to stay current on and even influence technology directions.
Cognitive computing can free humans to be more human
Taking a heuristic approach to deploying cognitive computing technology will help organizations understand the technology, prepare for its disruption—and scope out its many possibilities for transforming products, services, experiences and even how we work. Cognitive computing ultimately is how we can flawlessly replicate human sensory capabilities, analyze more data in real time and free us to spend more time using our imaginations to create and innovate. Cognitive computing will not only help businesses contain costs and operate more efficiently, it will equip them to spark new waves of innovation.
[This article was co-authored by Aan Chauhan and Srinivas TK, Director, Global Technology Office (GTO) for Cognizant.]