From pilot wins to enterprise impact, scale your intelligent automation program with purpose.
A major appeal of intelligent automation (IA) is its ability to deliver quick results. Compared to other technology platforms that also promise increased workflow efficiency and productivity, IA tools like robotic process automation (RPA) and agentic AI can be implemented relatively quickly, with minimal personnel and infrastructure requirements.
While an intelligent automation program could always be planned and started at a large scale, the nature of the technology makes it ideal to start small. A business can get its feet wet with a pilot IA implementation and see results within a matter of a few weeks or months. Starting this way allows the team to learn while building, blazing a trail and producing assets that can accelerate future development. Regardless of business size, this is a practical, logical entry point.
Then, after that first automation goes to production and everyone celebrates the achievement, the easy thing to do is grab the next candidate off the top of the pile and repeat the process with the same people. Doing this means you don’t have to transfer knowledge to new team members, expand the program’s information technology (IT) footprint, or add management overhead. This light and lean team could theoretically operate indefinitely, ticking off one process after another, delivering the benefits of the highly flexible IA technology. In fact, it might seem like a mistake to interrupt that chemistry by introducing new costs and complexity.
However, at some point, this “if it’s not broken, don’t fix it” philosophy will hold them back from realizing their full potential. We will examine how scaling and maturing your IA program can compound your successes at a relatively low cost, and help you identify when it’s time to take these steps.
Step 1: Formalize Pipeline Generation
One of the first signs that your IA program is being hindered by its small size is an inconsistent pipeline of project ideas. A small pilot team will likely consist of a small number of developers, analysts, and subject matter experts (SMEs). The role of the SME is to use their in-depth knowledge of business processes to provide ideas to analysts and developers, who turn them into automations. Initially, this relationship is efficient, as the SMEs will have multiple high-priority ideas. Over time, though, the ideas will come more slowly as the SME finds themselves reaching farther outside their field of knowledge to source projects.
One solution might be to move the development pod to new SMEs in new departments; however, each time you do that, you start the learning curve over again.
Instead, this is the ideal time to construct a system for managing the pipeline.
- Create a template for automation proposals
- Set up a mailbox for submissions
- Dedicate a small block of time weekly to evaluate submissions
Your stalling pipeline is now accelerated, as your SMEs no longer need to put in the legwork of sourcing ideas; now the ideas come to them. Furthermore, individuals across the organization with ideas have a formalized pathway to participate in the program. The pipeline becomes a public document that anyone can interact with.
Step 2: Initiate Federated Development
In addition to ideas starting to trickle in more slowly, another trouble point you may experience is long, sluggish discovery phases. You have an idea that you know is a good one, but the process of transferring all necessary knowledge from the SMEs to the development team becomes the longest phase of implementation.
This is a great time to implement a federated development program. While the SME can be invaluable in teaching business processes to the development team, an even shorter path is for the expert to do their own development.
Employees who know a process they want automated can be encouraged to build their own automations and add them to the program’s work pool.
What they initially lack in expert automation skills will be overcome by their knowledge of the business and the processes they choose to automate. They are the people who know best which tasks consume the most time, and how to ease that burden.
With these individuals at the controls, the discovery process becomes simple. Granted, it might seem impractical or even risky to open the doors to anyone without those basic automation skills to start building automations, but you wouldn’t do so without guidance and oversight.
Step 3: Generate Training Playbook
The first step in guidance and oversight is training. One option for training citizen or federated developers is a classroom program where a teacher is selected or brought in to instruct a specific group of individuals. However, the appeal of intelligent automation tools in the first place is their ability to be deployed without major investments, including personnel.
Rather than scheduling classroom time with dedicated trainers, employees interested in federated development can be enabled and empowered to train themselves. After a few development cycles with the pilot team, you should have accumulated a substantial collection of assets to assist in development.
These can be compiled into a custom playbook:
- List of online training courses
- Sample practice projects
- Infrastructure enablement checklist
By making this accessible to any potential federated developer, you democratize development, giving anyone with initiative the ability to train themselves to create their own automations.
Step 4: Establish Enablement Repository
Once those federated employees are ready to begin developing in earnest, they will need a resource to provide guidance. Following the training playbook, additional assets can be created to guide them along the way.
A central repository can be hosted online, serving as a one-stop shop for everything they need to know:
- A technology selection worksheet to guide the developer through options for tools, ensuring they are using the right one for the job
- Documentation templates make it easy to document their build
- A best practices handbook lays out all the guidelines they need to follow to stay in compliance
- A test plan tracker allows them to thoroughly and comprehensively test their work
- Last, a build and deploy checklist can be a step-by-step guide through the entire process
When complete, this repository becomes a tremendous tool and a home base for all federated development. As a final safeguard, once the developer has trained and guided themselves through the development process, they can submit their project to a review board for approval. You now have a main hub and roadmap for the entire federated IA program.
Step 5: Build a Champions Network
Up to this point, every step described for maturing your intelligent program has prioritized minimizing staffing overhead. Through the power of self-actualization, a company can watch its employees trade repetitive job responsibilities for the highly engaging and rewarding practice of automation engineering. Eventually, though, most people can only make it so far independently. They will need support to get past roadblocks.
With the tiered approach of a champions network, support can still be provided while continuing to prioritize self-sufficiency:
- Communications channel for peer-to-peer exchange of ideas
- Nascent developers can support one another by answering questions, sharing reusable code, and offering ideas and inspiration.
- Recurring open-door sessions
- More accomplished developers can deliver lectures, demo projects, or conduct live brainstorming sessions.
- Contact list for support escalation
- When developers encounter problems that their peers can’t solve, this list shows them who to contact directly for help, without the runaround or hassle of opening a general ticket with their IT department.
A network like this, connecting developers of all skill levels, doesn’t just answer the FAQs of first-timers. It transforms your democratized group of individuals into a collaborative and innovative workforce far greater than the sum of its parts.
Step 6: Codify Change Management
The last major bottleneck that crops up in immature intelligent automation programs is usually user awareness and acceptance. As the small, dedicated development team goes through the phases of discovery, development, testing, and release, they perform their due diligence by identifying every process and individual their automation will impact and communicating the impact to all relevant parties.
Without this communication, users who could benefit won’t buy in, conflicting processes will remain in place, and confusion about what is happening and why will lead to wasted time and resources. It’s reasonable to expect the team to handle change management on the first few automations they build, but as they accumulate a library of automations, even their most carefully laid plans will have blind spots.
The solution to this is a structured change management system.
Introduce a standardized template for announcing the launch of new automations that clearly indicates:
- Function
- Schedule
- Impacted systems and parties
- Support contact information is the first step
Next, the same repository created for the federated model can also host an information sheet detailing each automation in production, with links to comprehensive documentation. Now everyone in the company understands what to expect from the automation program and where to find more information.
Finding the Right Balance
We just walked through the steps of maturing an intelligent automation program. What began as a small, dedicated team picking off the lowest-hanging fruit has evolved into an enterprise-wide perpetual-motion machine, transforming the company’s efficiency. It is important to emphasize that these are not competing visions.
There is a temptation to view this growth as a departure from the principles that originally made IA effective. By introducing new people, processes, and regulations, it appears that there is a risk of compromising the rapid gains delivered by a lean program. As we saw, however, this method eventually hits bottlenecks and slows down, failing to maintain the initial rate of delivery. Facilitating growth becomes the only way to keep moving forward.
Likewise, attempting to operate at this scale before you are ready will lead to inefficiencies and create overhead that might prevent the program from ever reaching its potential. The IA tools we implement are well understood, and this foundation of knowledge allows us to effectively deploy from one organization to the next, incorporating all the lessons learned along the way. Yet, each organization is different. Starting small allows us to deliver great results for your organization while establishing relationships, defining processes, and generating custom assets that will facilitate an effective maturing process.
Main Digital Can Help
Main Digital’s IA process is designed to work with your business, identifying inefficiencies in both your business processes and your technology programs themselves. The result is a collaborative engagement that grows and evolves with you, ensuring your intelligent automation program delivers maximum results from the outset and continues to yield benefits into the indefinite future. Contact us to learn more.
Contributed By: Tom Weaver