Artificial Intelligence

The Untapped Potential of Automation in the Public Sector

Automation provides more flexibility, improved productivity, and greater citizen satisfaction for governments facing ever-changing priorities.

By James Lye, Oracle Insight

July 2018

Three of the biggest challenges in the public sector are controlling costs, providing effective and efficient services to citizens, and scaling to meet changing needs.

Cost-wise, the global public-sector budget deficit is close to US$4 trillion a year (approximately 5% of global GDP), with global public debt at US$61 trillion and growing. Despite record spending, citizen satisfaction with public-sector services continues to lag the private sector by a significant margin. Meanwhile, governments are constantly expanding and shrinking as economic shifts occur, new administrations get elected, and the public’s priorities change.

Technology adoption and digital transformation, as they have in the private sector, can decrease costs and create better outcomes. Automation, in particular, can help governments scale, improve productivity, and improve citizen satisfaction.

Yet deployment of automation in the public sector is in its infancy relative to its potential. A McKinsey Global Institute paper estimates half of all workforce activities could eventually be automated. Roughly applied to the Australian public-sector context, as a midsize example, up to US$76 billion in potential efficiencies could be used to deliver much-needed services or to repair the budget.

Achieving the full potential of automation, however, requires significant investment, time, and management focus.

Rise of the Machines

Automation can broadly be categorized into physical (assembly lines, drones, driverless vehicles), digital (software/process automation), and combinations of both. The first wave of automation has been delivering productivity benefits for years in highly structured tasks and predictable environments—for example, robot arms in car factories and data scraping and aggregation on the internet.

However, advances in underlying and complementary technologies such as AI, machine learning, and IoT have now put increasingly complex cognitive tasks within scope, such as tasks requiring judgement or dealing with ambiguity.

In the public sector, robotic process automation (RPA) can deliver the greatest benefit in the short term. RPA uses software robots to perform tasks using the same tools human workers use. The robot can learn by watching humans perform tasks as well as by the trial and error of its own efforts. Combined with natural language processing, robots (for example, chatbots) can also take on customer interactions, improving with every conversation.

In the public sector, robotic process automation can deliver the greatest benefit in the short term.”

As an example, the University of Adelaide in South Australia deployed a chatbot to field questions from potential students during the yearly admissions rush—a critical time to attract prospective applicants. The chatbot, built within four weeks on a PaaS service, reduced average response times from 40 minutes to 90 seconds. Leveraging machine learning and natural language processing, the bot learned from prior conversations to deliver accurate information in natural, conversational responses that 60% of users rated 5 stars.

Driving Forces for Automation

Beyond the headline objective of improving productivity, federal and state governments have identified a number of areas where automation can play a large role.

  1. Next-generation workforce. Government agencies are increasingly called on to do more with less, and as a result, they have been outsourcing jobs offshore. RPA offers the chance to keep workloads and jobs onshore. Many of the tasks offshored are repetitive, high-volume, and highly structured [omit hyphen after highly]—perfect for RPA. Using RPA keeps knowledge onshore and allows employees to develop the next-generation skill sets necessary to support the design and deployment of automation.
  2. Improved services to citizens and businesses. Current eservices and online portals can be further improved through the use of customer-friendly robots (for example, chatbots). The increasing sophistication of the underlying AI and natural language processing allows the majority of queries to be answered quickly. A query on a tax return may have previously required a citizen to navigate through multiple screens to locate the right data. A chatbot could retrieve the same information with one question.
  3. Key back-office processes. Many back-office processes are still manual, or if they are digitized, they are not digitized in an end-to-end fashion. Each handoff adds delays and increases the margin for error. The push to digitize end to end is a perfect opportunity to build automation into the core design of key processes. Record-keeping, registration services, and claims processing should be highly automated with RPA.
  4. Migration to cloud. Historically, organizations have run siloed systems, resulting in a proliferation of subscale data centers and associated workforces. The current drive to consolidate into cloud facilities will benefit from the next generation of highly automated storage and compute—reducing labor costs, improving security and reliability, and enabling flexible self-service for users.
  5. Scaling. By lowering the overall cost of the delivery of services, automation can help the public sector scale efficiently as demand grows. Conversely, during times of austerity, automated processes can maintain a base level of service at little ongoing cost.

Key Considerations for Success

Getting automation right the first time is critical, given the sensitivity of citizens to changes in government services. Here are four key considerations for achieving immediate success with automation projects.

James Lye, Oracle Insight

James Lye, Oracle Insight

  1. Think about the long term. As with most major undertakings, having a clear idea of the end-state in the long term will improve upfront design and investment decisions. Make sure to properly manage competing priorities between individual agencies and centralized functions (for example, IT and procurement).
  2. Move to a flexible and scalable infrastructure. Cloud-based infrastructure is a good first step. However, as new applications enabled by advancing technology come online, workloads will increase significantly. The underlying infrastructure needs to be as automated as possible to avoid support costs ballooning. The infrastructure should essentially run itself, allowing resources and focus to be applied higher up in the stack.
  3. Build on a strong technology platform. Automation is powered by numerous underlying technologies, such as AI and IoT. Major software vendors are making these technologies more easily accessible through PaaS offerings. As with the chatbot example earlier, this allows organizations to rapidly reap the benefits of the vendors’ multibillion-dollar investments.
  4. Engage the workforce early in the automation process. This is critical. UK insurance company Aviva went to its workforce for ideas about what to automate. The frontline employees were able to provide many more-specific and more-accurate details than initial designs had identified. Retraining workers to work alongside robots has resulted in an uplift in employee morale, because employees are freed from the more tedious aspects of their jobs to focus on higher-value tasks.

Automation has to be a foundational piece of the public-sector strategy. The extensive benefits that can be achieved in scaling, productivity, and improved employee and citizen outcomes will only increase as the underlying technology improves.

Action Items

Building Government for the 21st Century

Oracle Insight for Cloud

Photography by Shutterstock