We’re seeing the number, type, performance, and sophistication of advanced and low code artificial intelligence and machine learning solutions proliferating at lightning speed. Despite the wide availability and options for these technologies, government agencies may not be leveraging them. Some are reliant on legacy systems. Others simply are not ready for an advanced automation solution.

Nonetheless, options exist for a careful transitional plan. For example, rather than going full-in on automation, taking a human-in-the-loop approach that could include desktop-level automation or simple bots, monitored automation is a manageable approach.

Implemented through a series of methodical steps, these small additions and changes yield significant results. This article provides options for an incremental approach to successful adoption of AI and machine learning. In this article, we provide an operational roadmap to help organizations navigate a digital transformation capable of supporting their mission goals.

Starting the digital transformation process

In theory, most government leaders agree that automation offers benefits that range from improving production efficiencies to quality of service delivery to customers. When it comes time to select and implement a solution, the reality is that digital transformation can seem daunting and out of reach. It requires transitioning legacy systems, shifting workload, and even finding, training, and retaining qualified staff. For those reasons (and others), agencies frequently find themselves mired in less-than-ideal situations to rapidly implement advanced technologies.

With an iterative framework, small, temporary solutions take hold and phase out as the environment matures.

Regardless of the real or perceived hurdles, digital transformation is a necessity for future resilience and organizations simply need to start somewhere to reach their automation goals. An incremental approach makes it possible (or at least palatable) for agencies to begin with basic automation solutions and gradually evolve processes to support more advanced ones. With an iterative framework, small, temporary solutions take hold and phase out as the environment matures.

Implementing and maintaining processes designed for obsolescence might seem counter-intuitive, but applying a solution that your organization is not prepared to support can lead to disastrous outcomes. According to a report from EY, as many as 30-50% of initial RPA projects fail. Most often, this failure isn’t because of technology. It demands a cultural shift from reliance on outdated, manual practices to an environment enabled by machines to strengthen processes, improve scalability, and accommodate future innovations. This is because even simplified, automation integration is part of a digital transformation. It’s more complex than a software implementation. The move to modernize and streamline processes may leverage certain technologies, but it’s critical to keep in mind that those technologies are tools — not solutions. Taking the time to develop a plan that best leverages automation is key for deployment and digital transformation success.

Small Changes for Big Results

Particularly for government agencies, small, temporary changes are the building blocks necessary for more advanced solutions. This conservative approach gives agencies several advantages, including:

  • Build on successes
  • Create a proof of concept
  • Foster agency confidence with automation
  • Learn on a small scale and experiment
  • Make iterative changes without interrupting productivity
  • Work with a legacy system

Together, these incremental wins add up to the most important benefit of a conservative approach: getting buy-in from departmental leadership or executive decision-makers for a full transformation.

Here’s how you make that happen

1 Identify Automation Opportunities

To start, examine your current processes. Identify existing opportunities to apply automation that don’t require significant support. Chances are that resistance to updated processes might be heavy. In addition to roadblocks caused by stringent requirements for technology, advancing an RPA solution is often met with deep institutional resistance. People don’t trust what they don’t understand, and not all of us are IT professionals. Most often, the concerns center on:
  • Job loss due to replacement by bots
  • Data security without a paper trail
  • Removing the perceived comfort of physical paper for notes, signatures, and archived information
The result is a general technophobia that prohibits or slows change. These blockers are not insurmountable, but care must be taken to explain why the changes are necessary and nonthreatening (at least). Highlighting process improvements that support higher-level, shared project goals supports a more comfortable transition for users who are fighting to let go of codified institutional ‘norms.’

Users are able to leverage small but impactful changes and experience the value of automation without requiring a drastic and uncomfortable shift in their day-to-day environment.

Herein is one of the most important areas where we see the positive difference of an agile, iterative approach. Users are able to leverage small but impactful changes and experience the value of automation without requiring a drastic and uncomfortable shift in their day-to-day environment. Low effort, high impact tasks are the target here. Find a process where the upgrade will provide “more bang for your buck.” Try starting with simple tasks like converting data from paper or non-readable portable document files (PDF) to formats that allow users to filter and use the data is incredibly impactful. It not only streamlines everyday tasks, it creates new opportunities to use the data in other applications. Improving internal processes limited by legacy systems, such as electronically generating contact lists or developing form emails to improve communications, also provide strong opportunities to show the value of automation.

2 Outline Process and Set Benchmarks

Begin mapping out the process and establishing benchmark performance data. Ask pointed questions about the nature of the process:
  • What output is required?
  • What are the quality requirements?
  • How will this affect the steps before and after this process?
During this in-depth analysis, critically examine the process for improvements using an adjusted flow or by eliminating unnecessary steps. Consider the benefits of automation in the context of the process. Where will it have the greatest impact and add the most value? These questions are key. Regardless of the process, keep the focus on showing value. Audit current benchmarks and show performance gains throughout implementation. Sketch the updated process flow for your results-oriented automation solution, remembering to show how it addresses policy or contractual needs.

3 Design the Solution

Remember that design starts with a dedicated team. Because this will be a lean effort, team members should be fully on board with the transformation, cognizant of the overarching project goals, and committed to working through the implementation process. Discuss and refine the project requirements to ensure that critical quality points are addressed, along with a well-defined scope, objectives, and clearly defined roles and responsibilities. Without these frameworks, the project is particularly vulnerable to scope creep, as participants try to add more (or even unnecessary) requirements in the name of process improvements. At this stage, the solution doesn’t need to be transformative. It should mimic the manual process with automation to help users feel more comfortable. Additional features and functionality can be addressed in later iterations. The initial solution should remain as close to the original manual process as possible for a true comparison that highlights the automation gains.

4 Build and Test the Solution

Rooted in Agile methodology, the project typically takes about two weeks of consistent, on-demand collaboration for the solution team. Keep in mind that the team will be lean. Recruit for flexibility — more than architects and testers, you’ll need team members who have experience developing and integrating code. Build the solution with accessible and open-source languages like Java, javascript, VBA, or Python. These languages are easier to use and don’t require specialized recruiting or a full-stack development team. Follow DevSecOps best practices as closely as possible to maintain a healthy development and deployment environment. Once the automation is ready, run it in a secure, controlled environment to test and fix any bugs until it is ready for an initial release. When ready, schedule a demo for key parties. This is an opportunity to show the solution, discuss anticipated performance gains, and get feedback to inform and strengthen the next iteration. With adjustments in place, expand the testing group to gather more insights and improve product performance. Product implementation should iterate rapidly, keeping detailed records to support the iterative process and accurate troubleshooting. 

5 Plan for Production

With several rounds of testing complete and all of the necessary approvals from leadership, make the automation available to all users. This requires a solid training plan that includes detailed instructions for use, answers to common questions from users, and how the update will impact their daily workflow. In addition to training sessions, share this information and other best practices with leaders and managers to ensure that employees have the support they need. Once launched, continue recording performance data to show the impact of the solution and to troubleshoot, make further improvements, or prepare for more advanced releases.

 

 

The Limitations and Considerations of an Iterative Approach to Automation

An incremental digital transformation process can empower government agencies to experience value using non-intrusive process changes that are less daunting, employ the skills of team members in new ways, and position the organization to accomplish its goals.

That doesn’t mean, however, that it is perfect. The approach has some limitations or concerns worth noting.

Once the organization progresses in its digital transformation and has sunset manual processes and legacy systems, simple automation solutions should give way to more robust, machine-enabled versions.

Incomplete transformation: Initial implementations are designed as temporary solutions designed to serve existing processes rooted in a legacy system. Some agencies will erroneously assume that these solutions are complete fixes. Once the organization progresses in its digital transformation and has sunset manual processes and legacy systems, simple automation solutions should give way to more robust, machine-enabled versions.

The upside here is that even if the organization stops the transformation process here, it still stands to benefit from the efficiencies and capabilities gained through the new systems.

Maintenance needs: The solutions may focus on small processes, but they require a good deal of maintenance. This myopic approach often leads to what’s referred to as shadow IT, a situation in which non-IT resources take on small tech-focused roles beyond their regular responsibilities. Because the added responsibilities are temporary and not part of a core job function, it must be emphasized that the solutions implemented are designed for planned obsolescence. You can reduce the risk of shadow IT with an implementation plan that clarifies roles, project length, and expectations once (if) the solution is handed off to IT.

Security concerns: When handling system information, data security is always a concern. Automation allows information to be moved more quickly and in greater volume, giving users even more power. Partner early on and frequently consult IT security to ensure that the solutions are not creating data risks. Ask your agency’s IT security team to recommend the most up-to-date protections for data handling and storage.

Key Takeaway for Iterative Digital Transformation

The path to digital transformation is rarely as simple as a turn-key solution. Taking an iterative approach is a great way to move forward in the face of blockers. We understand this isn’t the full process, but it’s a good stepping stone. With thoughtful planning and incremental changes, organizations can position themselves to reap the benefits of advanced automation.

NT Concepts helps agencies see big results in process efficiency, security, and growth through small adjustments like improving how existing technologies are used. If you’re ready to start the path towards digital transformation or need help designing and implementing advanced AI or machine-learning solutions, our experience, dedication to process innovation, and flexibility to meet you wherever you are in your transformation can help you achieve your process automation goals.