As organizations develop automation strategies, implementing a 90-day playbook on how to approach it makes the process more manageable.
Starting small and scaling smart is a key part of an automation playbook. It involves first putting all the instrumentation in place and “baselining” to get a sense of how a system is performing, says Corey Ercanbrack, chief product technology officer for Vasion, a serverless printing and orchestrated automation company.
Here Ercanbrack takes us through a 90-day plan companies can follow to create an automation playbook.
Month 1: Discovery & Prioritization
During the discovery phase of the automation process, organizations build momentum and trust for what follows, according to Ercanbrack.
In the first month, he recommends building a cross-functional discovery team that includes representatives from IT, operations and business.
“This ensures you capture diverse perspectives on pain points,” Ercanbrack said. “At Vasion, we include team members who perform manual tasks daily, as they often have the most insightful suggestions for improvement.”
The discovery and prioritization stage involves mapping high-cost manual tasks. This mapping should comprise a two-pronged approach, according to Ercanbrack.
“First, gather quantitative data through time-tracking and measuring effort,” he advised. “Second, conduct facilitated workshops where teams can share their automation wish lists.”
Although the human element is still essential, modern data analysis tools, which incorporate AI, allow organizations to process information more effectively as part of this data gathering approach.
Prioritizing low-risk, high-impact opportunities is a key part of an automation playbook. Creating a matrix with axes for business impact and implementation complexity can help organizations plan for automation, Ercanbrack suggested.
In month one, organizations should also document key findings on automation and build a roadmap that continues past the first automation target.
“This helps stakeholders see the bigger picture while focusing execution on a well-defined starting point,” Ercanbrack said.
“Setting clear success metrics at this stage is crucial,” Ercanbrack added. “Define how you'll measure improvement before you start building, whether it is time saved, error reduction, or faster delivery cycles.”
When proceeding with testing, do not automate too much at once. Start in small, testable increments.
“This allows you to validate components individually before integration, providing natural checkpoints for stakeholder feedback,” Ercanbrack said.
Testing should incorporate expected scenarios as well as edge cases. Involve people who perform manual testing because they can spot issues that developers may not see, Ercanbrack added.
“Leveraging AI can help generate more comprehensive test scenarios, but human testers remain essential for validation,” Ercanbrack said. “At Vasion, we implemented a ‘human in the loop’ capability in all our early automation, allowing staff to intervene if the automation behaved unexpectedly.”
Month 2: Pilot Automation
In the pilot phase, organizations should choose a single workflow, whether that means onboarding or deployment. Workflow automation could entail a DevOps methodology.
Ercanbrack recommends a “focused approach” to pilot automation with clear boundaries.
“At Vasion, we found that constraining our pilot scope was essential for delivering meaningful results quickly,” he shared.
As part of this focused approach, organizations should consider employee onboarding because it can span multiple departments, presents clear metrics for success and has a positive impact on an entire organization, according to Ercanbrack.
Building and testing automation tools during the pilot phase should begin with a process map that outlines the current state of the automation project in extensive detail.
“Have the people who perform the process manually validate this map before writing any automation,” Ercanbrack suggested. “At all companies, we discovered that many standard processes had undocumented variations that would have broken our automation had we not identified them early. Today's process mapping tools incorporate AI to help identify patterns, but human validation remains crucial for ensuring accuracy and discovering gaps.”
During the pilot stage, organizations must decide whether to go with custom solutions or use existing tools. These tools should reduce implementation risk even if they do not offer every single function you desire from the start, according to Ercanbrack.
“Perfect is the enemy of good, especially in early automation efforts,” he said.
Month 3: Expand & Monitor
In the third month, organizations can expand on the progress in the first two months and scale as well as add monitoring, logging and rollback protocols. Ercanbrack explains that monitoring consists of implementing a comprehensive observability framework.
“Every automated process should log key events, performance metrics, and completion status,” Ercanbrack said. “At Vasion, we ensure that all automation has clear success and failure signals, as well as detailed execution logs, which help pinpoint issues when they arise. Modern monitoring tools now incorporate AI to help identify patterns and anomalies, providing earlier warnings of potential problems.”
Rollback capabilities should incorporate repeatability. Also, Ercanbrack recommends partial rollbacks rather than an “all-or-nothing” approach.
During the third month, track progress, according to metrics that include time savings, error reduction rates, resource usage improvements and user satisfaction via simple surveys. A basic dashboard helps organizations visualize these metrics.
This month will also involve training teams and documenting new workflows. This documentation is essential for organizations to troubleshoot and scale.
Ercanbrack recommends three tiers when training employees on automation workflows: They include using central repository for documentation with process diagrams, hands-on training and a buddy system where uses work with a colleague during the transition to provide support.
As organizations plan out their automation playbook, they should prioritize continuous learning.
“The rapid pace of AI advancement means that automation professionals must continually update their skills,” Ercanbrack said. “Organizations need to budget for ongoing education and provide structured learning paths that help team members stay current with the rapidly evolving capabilities and applications of AI in automation.”