Many companies placing automation at the center of digital transformation aren’t achieving the gains they expected. We believe the problem isn’t with the technology itself, but rather with the digital transformation approach they are using. The result: They tap but a fraction of the potential of today’s digital automation technologies.
Advances in technologies like artificial intelligence, analytics and cloud computing can provide unprecedented improvements in operational efficiency. But most companies have only scratched the surface of such technologies. A Forrester survey of 1,600 business and IT decision-makers found that only 17% of firms undergoing digital transformations will use AI – even though three-quarters see it as vital to transformation.1 Gartner notes that just 2% of customer service and support operations have used virtual customer assistant or chatbot technology across their engagement channels, though it predicts the proportion will rise to 25% by 2020.
When Traditional Approaches Aren’t Enough
Why are companies slow to use automation technologies? We believe it’s because they are using outdated business process improvement approaches. These methodologies (such as business process redesign, pure task automation, and Six Sigma) were invented in an era when very little manual work – both labor-intensive and knowledge work – could be automated.
Since that time, AI and other automation technologies have advanced significantly beyond robots bolting car parts or word processing software eliminating tedious writing work. Today’s technology can electronically connect companies to their customers, suppliers and other stakeholders around the clock.
What else is different about today’s workplaces that old approaches to process improvement fail to recognize? Three are significant:
The explosion in high-powered devices. Smartphones, digital watches, tablets, sensors connected via the Internet of Things and other “edge” devices enable companies today to interact with, monitor and guide customers wherever they are – on the road, in their stores, in a branch office or on a mountaintop. Today’s customer experience designers need to put themselves squarely in the customer’s shoes to create friction-free processes.
The abundance of data and low-cost storage capacity and computing power. In the past, companies had far less data to collect and analyze. As well, they tended to keep the data they did collect in functional silos. Today enterprises gather a lot more data – far beyond bookings or orders – in every business process (including marketing, sales and service) on what customers say (via social media) on what customers do (via geolocation data) and how products perform (via IoT sensors embedded in those products).
On-demand computing power in the cloud. Along with abundance comes ubiquitous access to data sources via cloud computing systems. Cloud-based systems also enable companies to analyze data from many sources through advanced and affordable applications. The machines can personalize data analysis for stakeholders like the CFO or CMO. That requires companies to think bigger about the kinds of immersive digital experiences they can now provide to customers.
Together, these opportunities call on leaders to take a new approach to digital transformation.
The Machine First™ Approach: What It Means
We refer to that approach as “Machine First.” We have named it that because it gives technology systems what we call “the first right of refusal” to perform a task that can be either be automated completely or done manually. Powered by analytics and AI, machine-first automation drives greater efficiencies by performing routine tasks faster and more accurately than before.
Implementing machine-first automation requires an agile development approach to process and product development. That involves rapid trials that apply lessons learned from trials that come before it.
The machine-first approach is not about reducing headcount. Rather, it is a way to free people from routine and repetitive work, and to use their talent for more sophisticated jobs doing the new work their companies now need. When implemented in an agile manner, a machine-first approach pushes the boundaries of human potential, strengthening an organization’s capabilities today so it can solve the business challenges of tomorrow.
Compared to mainstream digital transformation methods, the machine-first approach relies on much more data than was available in the past. That includes:
Structured and unstructured data the company collects;
Data from ecosystem partners;
Data from the public domain.
In addition, a machine-first transformation approach takes an enterprise-wide view rather than a functional or divisional view. The assumption is that data from across an organization will enable a company to share insights on customers, its business processes and other performance indicators with everyone in the company who needs it, and when they need it.
A machine-first automation approach also allows a company to radically rethink key business processes and achieve breakthroughs in performance. It can help a company’s strategists define new business models by reimagining what to sell, whom to sell to, how to generate revenue. For product developers, it helps them reimagine core products and services by driving the integration of digital with analog. For marketing and sales, it can transform segmentation, targeting and selling through of deep analytics. For supply chain leaders, it can improve production and distribution by integrating Business 4.0 technologies (cloud, analytics, automation and agile development processes) across the value chain. For R&D, it brings data-driven insights into product performance and market trends, both of which can boost product and service innovation.
Applying Machine First™ to Customer Service
To illustrate the impact of a machine-first transformation approach, let’s look at how it’s improved customer service:
Designing an end-to-end customer experience. A leading airline used a machine-first approach to its social media channels (including Facebook, WhatsApp, Twitter, WeChat and Google Home), which have more than 20 million followers. AI engines identify the intent of incoming customer conversations (more than 150,000 every week) to automatically respond to common questions ranging from loyalty programs and reservations to flight status and baggage. If a customer’s request is complex, the system passes her on to service agents. After implementing the system, the airline found its Net Promoter Score rose 17%, from 35 to 41.
Ensuring uninterrupted service during peak demand. Predictive maintenance, one of the killer applications of intelligent automation, enables a company to monitor its products and services for signs of potential failure and fix them before it happens. After suffering from service disruptions during its peak holiday shopping season, a leading fashion retailer implemented a machine-first model to run minute-by-minute health checks of its e-commerce and in-store point of sale systems. When the next peak in traffic arrived, the retailer automatically detected and resolved issues that could have led to more than 290 service outages in its point-of-sale systems during the holiday rush while its systems suffered no outages. The retailer saw a 55% spike in new customers served and a 29% rise in its e-commerce sales for the season.
Providing personalized service. The abundance of data and computing power enables firms to offer personalized services at an affordable cost. For example, a large investment company uses a machine-first model to build a service that complements the work of financial advisers. It uses machine-learning algorithms to automate processing and onboarding new clients, assesses their financial status and advises them on investments that meet their needs. The company can now cater to more customers at a lower per-client cost, while freeing up its financial advisers to work with higher-net-worth clients.
Empowering employees to solve problems. Companies can provide advice and tools to customer care employees so they can solve product issues. A German electronics manufacturer uses a machine first approach – including AI and other technologies -to create intelligent agents that receive queries from employees about a product issue. The system automatically provides them with how-tos, FAQs and screenshots from the company’s digitized knowledge base to answer questions and explain how to address problems. This has reduced the time it takes to solve many customer problems from days to hours and even minutes.
Implementing a Machine First™ Approach to Automation: Key Steps
Exploring which service improvements would have the most value for customers – and which they might actually pay for.
Determining which processes to automate to deliver that value, including data collection and analysis to understand changing customer behavior. (Design thinking can be used to simulate, test, and refine new customer service options.)
Selecting the appropriate intelligent automation technologies for high-value use cases.
Making the financial case for the intelligent automation investment, calculating the initial and ongoing automation costs against the customer revenue, loyalty and other improvements.
Identifying the new jobs and skills that the organization will need, and then developing employees’ skills to shift to these new jobs.
After identifying where to invest, managers should launch a pilot project and see how it’s working before deploying the technology more broadly. They should invest in foundational processes, new operating models, and IT systems that use the experimental design and testing approaches in iterative transformation.
But this is just a start. A Machine First approach to digital transformation uses technology advances to build an infrastructure of continuous improvement. It’s up to the leaders to create the capability that not only transforms their business but also encourages their teams to craft new ideas about how work could be improved through automation. Once they do, they can generate new levels of productivity and operational efficiency. What’s more, they can develop new services and deliver phenomenal customer experiences. By doing so, they will take their digital transformation initiative to the next level.
To take a Machine First approach to digital transformation, managers should begin by identifying the physical and knowledge work that machines should do. What tasks (like those mentioned above) can a company automate? The answers will lead to a list of priorities to pursue including:
About the author(s)
PR Krishnan (PRK) is Executive Vice President & Global Head, Enterprise Intelligent Automation & Artificial Intelligence at Tata Consultancy Services (TCS). In this role, PRK helps business and technology leaders drive innovation and integrate digital technologies – AI, smart automation, machine learning and cognitive computing – into their business model for growth and transformation.
PRK has strategized innovations that drive human - machine collaboration that serve as the guiding force for elevated growth and success in the Business 4.0 era. PRK spearheads the MFDM™ (Machine First Delivery Model) initiative in TCS to institutionalize the Machine First™ approach across client engagements and within TCS.
PRK’s team creates new opportunities for, enabling the world’s leading enterprises to harness human ingenuity together with machine precision.
For more than 35 years, PRK has played a pivotal role in coaching large teams, demonstrating thought leadership and innovation towards uncovering insights and creating exponential value for large enterprises.
Prior to leading the EIA & AI portfolio, PRK served as Global Head of the IT Infrastructure Services, overseeing the expansion of the TCS services portfolio. In this role, he successfully led several of TCS’ global delivery centers (GDC), paving the way for the creation of the company’s Global Network Delivery Model (GNDMTM).
Starting in 2004, he was responsible for creating and running the telecom delivery centers for TCS, where he built a 3000+ strong practice servicing over 40 customers, growing the practice in a time when the telecom industry was just beginning to embrace technological advancements.
PRK has contributed significantly towards leading a very successful SEI CMM Level 5 assessment that became a new benchmark and model across TCS.