Intelligent automation, or the combination of artificial intelligence and automation, is the next big disruptor. While conventional automation automates routine and repetitive processes, intelligent automation systems synthesize huge swathes of information to automate entire processes and workflows.
Intelligent automation systems perform complex industrial tasks, make decisions based on textual information, guide autonomous vehicles, and do much more. Early adopters have already reaped rich rewards through efficiency and quality improvements, customer delight, and path-breaking product offerings.
2020 will be a breakout year for intelligent automation with entrenched companies seeking Artificial Intelligence (AI) based tools to break out or maintain a dominant position in their niche or industry. Upstarts will be using AI to innovate and leap ahead of entrenched competitors. Enterprises who are currently applying task-based automation for low-value opportunities will be seeking to incorporate more advanced analytical and AI technologies for bigger opportunities.
A May 2019 Deloitte survey covering 523 executives across a cross-section of industries, and spread over 26 countries, reveal that companies have moved beyond robotic process automation (RPA) to embrace intelligent automation. 58% of the companies surveyed have started their automation journey. 38% of them are in the piloting stage, 12% are in the implementing stage. 8% are automating at scale.
Why such widespread interest and adoption?
More Productivity with Improved Accuracy
Standard automation automates high-volume, complex, repetitive tasks. It speeds up the process and ensures very high accuracy. For humans, it offers deliverance from dull, repetitive tasks, and allows them to focus on cerebral work. For instance, applying RPA in retail spares workers from the tedious task of checking orders and calibrating data, and allows them to work on creative tasks such as forecasting inventory orders or tending to customers.
But traditional automation has its limitations. It suits only a high volume of repetitive tasks. It falters as soon as it encounters natural language text, images that require categorization, or anything that require judgement.
Intelligent automation applies Artificial Intelligence (AI) to mimic the human brain. It takes automation to a new level, enabling the automation of complex repetitive tasks. With AI, it becomes possible to automate processes based on both structured and unstructured inputs.
A major application of intelligent automation is in quality control. When the human brain spends hours repeating tasks, fatigue creeps in and induces errors. In huge volume processes, even a minuscule mistake creates a nightmare. Intelligent automation processes and analyses orders in real-time, streamlining operations and ensuring 100% accuracy.
Kiva Systems “mobile-robotic fulfilment system” automates retail distribution. The robots travel around a distribution centre to transport shelving units to workers, who prepare customer orders for shipment. The system improves worker productivity by 200%. Amazon has acquired the company.
A European land registry uses cloud-powered deep learning and computer vision to compare property records and aerial images. The insight allows employees to investigate discrepancies and non-compliance with 80% accuracy.
For more: Intelligent Process Automation: The next wave of RPA
Automation saves costs through reduced workforce requirements, lesser errors, optimal use of resources, and faster time-to-market. Intelligent automation expands the scope of cost savings to new areas.
Consider the case of an insurance company that receives four million customer emails per year in three different languages. The company deployed a workforce of 80 full-time employees just to route these emails. Applying intelligent automation allows categorising these emails into 400 different categories, with 98% accuracy. A sophisticated multi-layer categorization model route these emails. The reduction in staff is proportional to the level of accuracy required. For 98% accuracy, 25% of the routing is automated. Humans double-check the rest. Relaxing the accuracy to 95% allows automated processing of 70% of the emails.
Likewise, applying natural language processing (NLP) allows organizations to analyze text documents in double-quick time, with a high level of accuracy. This frees up an army of workforce otherwise required to perform manual transcription.
The possibilities are endless. Artificial Intelligence-powered automation offers enterprises access to virtually infinite amounts of human-like, cheap, and easily deployable workforce. These bots do not go through a learning curve, do not suffer from fatigue, and do not require emotional support. Enterprises can scale up to infinite levels on the fly.
Marketing success depends on knowing when, where, and how to push a product or service. Automation spares marketers the manual tasks associated with dispatching emails or generating notifications. But the marketer still has to decide who to send the message, what to include in the message, and so on.
Intelligent automation gives automation a mind of its own. Bots analyze large swathes of data to optimize marketing campaigns. The technology looks at the success rates of a certain type of product or a certain target and acts accordingly.
Smart companies have already leveraged intelligent automation to develop marketing systems that presents offers to customers based on their profile. The system scans the data associated with a target and decides which content to despatch. It changes website content depending on the user who has logged in, based on their digital footprints.
Advanced Artificial Intelligence even automates content creation, using highly engaging prose! Credit Suisse, one of the early adopters of intelligent automation, analyzes millions of data points, to generate research reports in English. These automated reports assess expectations, upside, and risk factors related to companies targeted for investment. Applying intelligent automation has helped Credit Suisse triple the volume of reports while improving quality and consistency, when compared to analyst-written reports.
Speed up Complex Tasks
Businesses use intelligent automation to streamline processes and speed up complex decision-making.
Automated systems powered by machine learning ingest and analyze massive amounts of textual information to respond quickly to complex inquiries. Some common applications already in vogue include:
- credit card processing systems to identify and block fraudulent transactions
- e-discovery systems to classify documents according to meaning and relevance, in litigation.
- Insurance approvals systems, to approve medical treatment plans after considering circumstances
A large hedge fund applied intelligent automation tools to extract content from research notes for compliance and quality. The tool applied computational linguistics technology to read and interpret the text and unearthed inconsistencies in the expressed sentiments and ratings. The sentiment diverged from the analyst ratings in almost 50% of all research notes! The insight allowed the company to redesign its processes and incentive structures. Manually reviewing all the 100,000 published research notes to unearth such inconsistencies was an impossible task.
Applying artificial intelligence may not even be a luxury going forward. The relentless pace of data and information creation makes it impossible for human brains to play catch-up. For instance, the amount of medical information doubles every five years. AI-based power tools may become indispensable to understand and use humongous volumes of information.
Intelligent automation unlocks new possibilities
Normal automation eases processes and tasks, making work faster, easier, and smoother. Intelligent automation allows enterprises to fix or improve the outputs rather than tasks.
Intelligent automation allows enterprises to leverage the power of collective experience by allowing them to train algorithms with the decisions of the experts and develop prediction models. Such models allow applying the intelligent decision making capability of the top experts anytime, anywhere. If offers junior employees a guide to towards the best decision.
The “intelligent” system learns and adapts on the go, meaning each application becomes better than the previous application.
Intelligent automation has already given a fillip to several business opportunities. Intelligent automation help home seekers locate a perfect apartment without the real estate agent having to painstakingly search the list of available property, applying filters based on customer’s preferences.
Singapore’s Safe City test bed deploys bots to monitor CCTV cameras. The human employees focus on the alerts these bots emit, rather than question their purpose of life by staring at a monitor screen for hours on end. The network of cameras, sensors, and data feeds links to an artificial-intelligence-powered system automatically flag images to a human analyst watching for threats. Similarly, London’s network of video cameras employs intelligent technologies to identify crime suspects automatically.
In an economy marked by slowdown and cutthroat competition, enterprises apply intelligent automation to not just improve efficiency and cut costs, but also to differentiate products and innovate.
Many businesses have started to apply intelligent automation to drive innovation. Enterprises seek to disrupt before being disrupted. Competitors have access to the same technology and the same possibilities. Innovation is the key to differentiate and get the first mover advantage.
Intelligent automation technologies power robots that analyze and respond to situational data emitting from sensors and other real-time data sources. The Nest Thermostat, for instance, adapts as the user’s life and the seasons change. The thermostat programs itself after a week of use, sparing users the hassles of making changes every time.
WellPoint, the leading health insurance coverage provider leverages IBM’s Watson cognitive automation technology to provide fast and precise decisions on patient care. The robust algorithm, developed with 15,000 hours of training by nurse clinicians, recognizes the unstructured English spoken by physicians, and recommend treatments based on WellPoint’s policies and clinical guidelines.
Another high-tech manufacturer applies intelligent automation to monitor suppliers and customers for risks and opportunities, and make decisions on customer credit. The Artificial Intelligence powered software extracts key metrics from the financial statements, ingest news sources, data feeds and all other available info, to generate credit models. Human analysts who monitor the credit ratings validate it 95% of the time, without change.
A big future of intelligent automation is in autonomous or self-driven vehicles. Most leading automakers, such as Audi, Nissan, BMW, Volvo, and Mercedes-Benz are developing self-driving cars. Google’s autonomous driving technology powers driverless cars in California and Nevada. International mining company Rio Tinto uses autonomous hauling trucks to navigate with limited human intervention.
The potential of Intelligent Automation is still mostly uncharted territory. The future lies in human-machine collaboration.
Motivating Human Resources
Traditional perception is that automation eliminates jobs. Automation does make certain low-level jobs obsolete. But in today’s age of talent crunch, the threat of job losses is not as relevant as before. In many sectors, including labour-intensive sectors such as manufacturing and agriculture, there is a shortage of workers. An ageing workforce and the unwillingness of the millennials to work in traditional occupations create a scarcity of labour and increase HR costs. By 2028, Europe will have 8 million fewer workers from current levels.
Forward-looking enterprises regard talented employees as a source of competitive advantage and try to expand their activities with the pool of available workforce, rather than shed talented workforce just because automation saves time.
Intelligent automation, rather than cut jobs, enhances the quality of jobs. It shifts tedious and hard-to-do jobs to computers to robots, leaving the existing workforce to focus on creative things.
Volkswagen’s “collaborative robots” work side-by-side with humans, and perform physically demanding steps in the engine assembly process.
The Roomba robotic vacuum cleaner spares humans the painstaking task of moving around with a vacuum cleaner.
Intelligent automation works. Enterprises combining Robotic Process Automation with Artificial Intelligence increase revenue by 9% on average. In contrast, businesses that do not combine these technologies have only an average revenue increase of 3%. But for intelligent automation to work, the pre-requisites are mature process definitions, robust standards and processes, and a clear-cut understanding of how to capture value.
Intelligent automation succeeds when enterprises identify their weak points and time-consuming activities upfront and develop a strategy to fix it. Successful application of intelligent automation requires a gap analysis. In some cases, it requires altering or shedding existing systems and processes to incorporate Artificial Intelligence. Integrating old databases with new Artificial Intelligence systems often pose a big challenge.
There are other significant challenges as well.
Resistance to Change
Resistance to change is the biggest obstacle towards implementing intelligent automation in enterprises. Intelligent automation comes with changes in role definitions and requires change management interventions.
The onus is on the champions of Intelligent automation to:
- Demystify: Provide clear and in-depth information on how intelligent automation will change jobs and make the workplace better. Counter popular misconceptions head-on.
- Demonstrate. Always apply a pilot of the proposed change. Fine-tune the model for any changes. Co-opt stakeholder feedback to the extent possible.
- Advertise. Communicate the benefits of intelligent automation. Sell the changes. Give ownership and the onus of implementation to the affected stakeholders.
- Demonstrate again: Implement intelligent automation across different business. Target low hanging fruits first to convince sceptics of the benefits.
Another big stumbling block towards enabling intelligent automation is the talent crunch. 59% of enterprise automation leaders believe they lack the workforce capacity and skills required to drive automation. The shrinking demographic trends and the millennial’s changing approach towards work add to the talent crunch. Enterprises will have to explore “alternative workforce” such as part-time or freelance experts to drive automation.
There is also a need to reskill existing employees based on how human workforce will interact with machines.
36% of enterprises perceive process fragmentation or the way the enterprise manages process as the main barrier to adopt intelligent automation. 17% of enterprises believe their IT department is not ready to unleash intelligent automation. In both instances, it requires the active intervention of the top management to invest in new systems, procedures, and capabilities that eradicate these pain points.
Growth comes through changing practices, technologies and business models. Letting go of obsolete tools and traditions is as important as acquiring new tools and competencies. Unlearning is an underrated concept and is the bedrock on which intelligent automation rests.