Automating a complicated process doesn't necessarily make it better. Sometimes it simply allows the confusion to happen faster.
Artificial intelligence and automation are creating new opportunities for public agencies to improve service delivery, reduce administrative work, and make information more accessible. For organizations with limited budgets and growing responsibilities, the appeal is understandable. Technology promises to help employees accomplish more without continually expanding staff.
But there is an important question that deserves attention before an agency deploys an AI agent or automates a workflow: Do we understand the process well enough to automate it?
Across the Monterey Bay region, public organizations are working in an environment shaped by changing technology, workforce needs, and expectations for responsive services. The Santa Cruz County Regional Transportation Commission's November 2026 IT consulting solicitation, for example, describes an operating environment involving public communications, records management, hybrid meetings, and coordination with technology vendors.
That solicitation is not evidence of an AI project. It does, however, illustrate the variety of interconnected responsibilities that public-sector technology must support.
The Real Problem May Be the Process
Consider a public agency receiving requests for services from residents, businesses, and community organizations. Requests might arrive through email, telephone calls, web forms, or in-person conversations. Employees then determine which department is responsible, gather missing information, and coordinate a response.
If the process is inconsistent, introducing automation can create new problems. An AI agent might categorize requests incorrectly because departments use different terminology. A workflow could route incomplete information to the wrong team. An automated notification might tell a resident that a request is complete when only one administrative step has finished.
The technology may be performing exactly as configured while the organization produces an undesirable outcome.
This is why workflow redesign should precede automation. Agencies need to understand how requests enter the organization, how decisions are made, where information changes hands, and which exceptions require human judgment.
Bring the People Doing the Work Together
One of the most effective ways to uncover these details is to bring employees from different functions into the same conversation.
A department manager may describe a straightforward approval process. Frontline employees may explain that unusual requests routinely require additional calls, missing documentation, or informal coordination. Finance may identify compliance requirements that aren't reflected in the documented workflow. IT may discover that two departments maintain conflicting versions of the same information.
Each perspective reveals part of the problem. The objective is not simply to collect requirements but to develop a shared understanding of how the work actually happens.
At DKomplex, we refer to this as building coherence. Our collaborative approach recognizes that people support what they help create and know they benefit from.
Through our Problem Solving Workshops, we help organizations uncover assumptions, examine competing perspectives, and develop practical solutions before committing to technology changes.
Identify What Should and Shouldn't Be Automated
Not every activity requires artificial intelligence.
Many routine tasks can be improved through straightforward workflow automation. A submitted form can create a record, notify the appropriate team, request missing information, or establish a follow-up deadline.
Microsoft Power Apps and Power Automate can support these kinds of processes, depending on an organization's licensing, systems, and security requirements.
AI may be useful when a task involves interpreting unstructured information, summarizing lengthy documents, or helping employees locate relevant guidance. However, these capabilities require additional controls, particularly when personal information, eligibility decisions, or public records are involved.
A useful starting point is to separate repetitive administrative work from decisions that require discretion, professional judgment, or formal authorization.
The goal should be to make employees more effective, not to remove accountability from the process.
Establish Clear Data and Governance Rules
An automated process depends on reliable information. Before implementation, organizations should determine which system holds the authoritative record, who can access sensitive data, and how information is corrected when errors occur.
They should also establish what happens when automation fails. Can an employee override an incorrect classification? Is there a clear record of who approved a decision? Can residents reach a person when a digital process doesn't resolve their concern?
These questions become especially important when AI agents interact with multiple business systems.
The NIST AI Risk Management Framework offers a useful reference for considering reliability, transparency, accountability, and risk throughout the lifecycle of an AI-enabled system.
Governance should be designed into the workflow rather than added after deployment.
Start Small and Measure Real Outcomes
Public agencies don't need to redesign every service simultaneously. A more practical approach is to select one well-defined process with measurable problems.
For example, an agency might examine internal service requests that require employees to copy information between email and spreadsheets. Before making changes, measure the average time to assign a request, the number of incomplete submissions, and the frequency of reassignment.
Then redesign the process with employees, implement a limited solution, and compare results.
If assignment times improve but employees spend more time correcting errors, the automation may not be delivering the intended benefit.
Measures should include both efficiency and service quality. Faster processing matters, but so do accurate decisions, employee confidence, accessibility, and the experience of the people being served.
Employee Adoption Is Part of the Solution
Technology change and organizational change do not happen at the same speed. Employees may understand a new application long before they trust it or incorporate it into their daily work.
Training alone cannot resolve every concern. People need opportunities to question assumptions, explain exceptions, test proposed solutions, and influence how the system operates.
Leadership also has a responsibility to communicate why the change matters, what success looks like, and how employee feedback will be used.
When the people doing the work participate in designing the solution, the organization is better positioned to identify practical risks and build lasting adoption.
Better Public Services Begin With Better Understanding
AI agents and workflow automation can help Monterey Bay public agencies improve operations, but their value depends on the processes and people behind them.
Before asking what technology can automate, organizations should ask what problem they are trying to solve, whether the existing process makes sense, and whether everyone involved understands the intended outcome.
DKomplex helps organizations address these questions through collaborative problem-solving, business process redesign, Microsoft Power Platform, CRM implementation, and people-first technology adoption.
Explore our collaborative workshops to learn how shared understanding can lead to better technology decisions.
DKomplex | Complex problems made simple.
Sources and Regional Context
The local context includes the Santa Cruz County Regional Transportation Commission's IT consulting solicitation, the Monterey Bay State of the Region event, and the Technology in the Workforce Summit. The governance discussion references the NIST AI Risk Management Framework.