Many executives and managers don’t understand what governance is and what goes into implementing it. However, they do recognize the risks and inefficiencies of opening access to data, technologies, procurement, and workflow to all employees in whatever way they want.
Executives also recognize the transformative opportunities of AI, but leaders generally don’t support open experimentation without defined goals, approaches to validate quality, and a plan to get successful experiments into production. Many organizations, especially businesses in complex industries like manufacturing and construction, seek pragmatic AI, where AI is applied and tested against defined opportunities and problems, not moonshots or endless experimentation.
While many governance implementations focus on what employees can’t do, executives are more open to discussing what they can and should be doing. I call this empowering governance, which calls out the behaviors, establishes the guardrails, and evolves standard best practices that enable employees to leverage technology, data, and pragmatic AI in their workflows.
Empowering governance: A pathway to productivity and growth
Empowering governance provides employees guidelines on using technology, data, and AI while minimizing risks. Leaders start by setting clear objectives, explaining non-negotiable compliance requirements, and establishing measurable metrics. I recommend leaders take an agile and iterative approach to defining and rolling out empowering governance to employees so they aren’t overwhelmed and have opportunities to provide feedback.
Empowering governance should have several objectives and benefits, such as delivering efficiencies, improving customer experiences, or reducing risks. In 2025, empowering governance should also establish pragmatic AI objectives and AI governance so businesses can establish organizational and technical practices that enable today’s and future AI benefits.
Here are five ways to get started, and I recommend -gradually building on each of these areas:
1. Eliminate gray work to boost productivity
Trying to use machine learning’s prediction capabilities, build automations to integrate workflows, or develop AI agents to improve decision-making doesn’t work if every department, team, or project has disparate practices with kludgey tools. Using long-threaded emails for communications, creating spreadsheets to support operations, and manually developing Gantt charts to represent complex schedules are all gray work practices that inhibit using analytics and AI.
Empowering governance should include a clear statement to managers and employees on the challenges with gray work and a clear explanation of why the organization targets work management. Leaders must manage this evolution with change management practices, especially in organizations with managers with significant subject matter experience who prefer sticking with tools they know.
2. Partner with citizen developers on best practices
Developing applications and workflow automation is no longer just an IT responsibility. It’s also not advisable to deploy low-code tools to citizen developers without a collaborative operating model, or in other words, empowering governance.
A partnership between IT and operating departments fosters standards and best practices, including:
- Creating reusable components, especially for managing data used across applications
- Defining naming conventions and other documentation standards
- Establishing deployment practices such as testing and incident management
- Ensuring applications follow security and compliance requirements
3. Centralize, normalize, and cleanse data
Using spreadsheets puts data quality in the hands of end users and complicates developing trust and transparency around decision-making. The practice is highly problematic when driving pragmatic AI, where models are only as intelligent as the quality of data feeding them. Models are more costly to develop when data is all over the place.
Centralizing data starts by illustrating that end-users can access data without exporting it. For example, I will develop dashboards and reports in Quickbase, but I will also use Qunect to connect Tableau when I want to perform more advanced analysis. I’ll create Pipelines and Zapier Tasks when automating workflows and connecting data from other systems.
The challenge is many employees don’t know about these tools and how to use them, so they fall back on the tech they know, like pasting and importing data into spreadsheets. Empowering governance should include learning, training, developing POCs, and leading hackathons to better educate employees on using tools that support proactive data governance and smart integrations.
4. Create transparent and manageable entitlements
Large language models and AI agents can drive significant productivity improvements by helping employees answer questions, review decision recommendations, and trigger workflow automations. However, organizations must ensure that the appropriate people, including employees and contractors, have sufficient data access, application entitlements, and data security training.
There are several important considerations.
- Are compliance requirements understood, and are related policies well-defined and communicated to employees?
- Are the policies easy to implement and manage? Do they support granular row and column-level permissions?
- Are data owners assigned, and do they understand their responsibilities?
- Can employee requests for access and entitlements be reviewed and addressed quickly?
- Is confidential and protected data secured and masked when required?
Addressing these starting principles is key for progressing from data-driven organizations to delivering pragmatic AI capabilities.
5. Establish regular communications to engage employees
An important governance practice that’s often overlooked and underinvested in is communications. Empowering governance requires behavioral changes, education, and time to experiment as AI and other technologies evolve. Organizations that engage employees with clear communications, guidelines, and incentives are more likely to experiment with AI in the most pragmatic areas, address risks during the implementations, and deliver business outcomes.
If we look at how new technologies gain adoption, they become less expensive to procure, easier to deploy, and more scalable over time. Addressing human factors separates faster and smarter organizations from ones lagging behind competitors or taking on too much risk in their implementations. Evolving empowering governance, setting clear objectives, and communicating with employees regularly are all key steps to engaging employees on pragmatic AI applications.
Isaac Sacolick is President of StarCIO, a technology learning, leadership, and advisory company that guides organizations in building digital transformation core competencies. He is the author of Digital Trailblazer and the Amazon bestseller Driving Digital and speaks about agile planning, devops, data science, product management, and other digital transformation best practices. Sacolick is a recognized top social CIO and a digital transformation influencer, with over 900 articles published on his blog, Drive, and other sites. You can find him sharing new insights on the Driving Digital Standup or during his weekly Coffee with Digital Trailblazers.
