How to get AI democratization right

One primary IT objective is to reduce the complexity of using technology to empower business teams to continuously improve operations. In doing so, CIOs help departments become more customer-focused, efficient, and data-driven by leveraging their expertise to define their technical operations, often without IT’s direct involvement.

This democratization of technology and data now extends to AI capabilities, including copilots, language models, and AI agents. The question is whether CIOs, CISOs, and CDOs will get the right balance between capability and governance; innovation and security.

“Generative AI is shifting technical skills from IT teams into everyday business roles, allowing customer experience and operations leaders to pull important insights from customer data without relying on engineers,” says Daniel Ziv, global VP of AI and analytics at Verint. “Agentic AI speeds this change even more by automating how insights are found and acted on, reducing the time from decision to action from weeks to hours.”

Learnings from low-code and citizen data science

To help their organizations to get the most of AI and AI agents, CIOs can learn from previous technology evolutions that helped progressive IT departments democratize technology and data capabilities, including:

• The rise of data visualization tools and citizen data science practices, which democratized the creation of reports and dashboards, once primary IT responsibilities
• Low-code and no-code development, which helped CIOs expand application development and enable business teams to enhance their applications
• The introduction of drag-and-drop content management systems (CMSes), which have enabled business users to build websites that once required teams of designers and developers
• The rise of enterprise SaaS tools that offer professionals in marketing, sales, finance, and human resources the ability to achieve through configuration and simple automations what used to require full development teams to write code to customize workflows

In some cases, empowering the workforce with self-service capabilities accelerated innovation and improved operations. But many CIOs will also recount the wounds of shadow IT, major security incidents, technical debt, data duplication, and other issues stemming from poor governance and enforcement.

“The democratization of AI is shifting technical capabilities directly into the hands of business functions, but this expansion of shadow IT requires a new model of IT partnership centered on accountability,” says Andrea Malagodi, CTO at Sonar. “As we move toward a future of agentic AI, where autonomous systems contribute more to our digital infrastructure, IT’s role must evolve from a manual gatekeeper to a provider of automated, intelligent guardrails.”

The impact of getting democratization wrong

There is a rationale for democratizing AI with caution, limiting the number of people, tools, and functional areas where it is piloted. No CIO wants to see rogue AI agents, leaked intellectual property, or lawsuits from customers given incorrect information. Democratizing AI has great rewards, but also significant risks and many unknowns as the technology improves.

CMOs and CISOs are terrified of AI security incidents impacting customers, such as an AI chat app exposing 300 million messages tied to 25 million users. CIOs and CDOs must be concerned about flaws in AI code generators and vibe coding tools that can enable data theft and remote code execution (RCE) attacks. It only took me a handful of prompts to find over 100 AI incidents over the past few months.

In 2025, most CIOs had little chance of fully controlling AI in their organizations. There was too much pressure from Boards to find efficiencies and not to be caught too far behind competitors’ efforts. Additionally, there was too much demand from business leaders and employees to use LLMs in their work, and now many enterprise SaaS platforms have AI agents for employees to try out. So, CIOs can’t become the office of no when it comes to AI.

“AI democratization can be successfully implemented throughout organizations when IT decentralizes decision-making and makes safety standards more accessible and achievable,” says Chris Mierzwa, senior director of global resilience programs at Commvault. “Instead of being the bottleneck, the IT team can balance governance to meet the current needs of employees.”

Mierzwa recommends providing employees with a pre-approved list of AI platforms they can use. AI governance programs must also specify who has access to which tools and provide guidelines on their use. Additionally, organizations must update their data governance policies and tools to address data privacy, protect IP, and monitor for sensitive data when training and prompting LLMs or engaging AI agents.  

Taking this one step further, Malagodi of Sonar adds: “Implementing a ‘vibe, then verify’ culture supported by a rigorous verification layer can empower every business unit to innovate at speed while ensuring quality and security remain intact.”

Define a vision for AI opportunities

In 2025, AI opportunities more often led to a speed-with-less-rigor response from leaders. Rather than succumbing to the fear of AI risks or prioritizing security-first implementations, many rushed to deploy AI tools and enable experimentation across numerous departments.

The result is that many businesses ended up with too many pilots, few of which were deployed to production, and some that required a churn of rework. McKinsey’s State of AI Report 2025 states that no more than 10% of businesses report scaling AI agents across any individual business function. Workday’s report on driving ROI reports that 85% of employees say AI is already helping them save time, but nearly 40% of that time is being spent reviewing, fixing, and reworking AI outputs.

Democratizing AI shouldn’t mean departments, teams, and employees can do what they want with AI, even if their ideas meet AI and data security standards.

“AI democratization is about people and organizations having broad access to not only AI tools, but also to the know-how and cultural conditions needed to convert AI into actual impact,” says Brandon Sammut, chief people and AI transformation officer at Zapier. “Democratizing AI is about leadership, learning, and culture — these are the ingredients needed to make the benefits of AI truly available to all.”

Leaders must go beyond prescribing what not to do to avoid chaotic experimentation and deliver business value. Top organizations blend AI governance with strategy, communicating a single message to employees on what to focus on and the implementation guardrails they should respect.

Democratize AI by sponsoring on-the-job learning

Between defining a focused strategy and establishing governance lies delivery: the planning, implementation, testing, operationalizing, and change management to deliver business value. Let’s consider several issues facing AI democratization efforts.

First, IT is often a detractor to democratization efforts, even when CIOs sponsor them. In some organizations, complex stage-gate processes, lengthy reviews by architecture boards, and a build-over-buy DevOps culture can hold back the transformation.

“Leaders should focus on technology tools that enable these operators to implement changes quickly, bypassing the long roadmaps and matrixed decisions that eat up ROI,” says Ashley Moser, co-founder and CCO at MelodyArc. “Workflow improvements should be driven by the people closest to the work, a paradigm shift to how process improvement has been done in the past.”

Second, while many departments will have AI early adopters, they are likely to face roadblocks from their later-adopting peers and others who fear AI will eliminate their jobs.

“AI has the potential to add tremendous business value for customer and employee experience teams, but businesses won’t recognize the full value of AI if their employees are reluctant to use it,” says Brad Murdoch, CEO at Deskpro. “The key is demonstrating that AI as an assistant can make their jobs a lot easier, and is not a replacement for the human empathy, compassion, and problem-solving that’s needed for great experiences.”

Adapt change management programs

A key step for CIOs looking to scale AI democratization efforts is to develop an evolving change management strategy and on-the-job learning opportunities.

They should start by partnering with the CHRO on developing tailored programs to improve AI literacy.  These programs should help employees grasp what AI can and cannot do, how to work effectively with AI tools, evaluate AI’s outputs, and review ethical principles.

Change management strategies will need ongoing updates as AI models improve and AI agents are deployed.

• Companies in early stages should guide business teams in reviewing whether their data is AI-ready and provide release-ready guidelines around pilots and deployments.
• Those with AI agents reaching production should consider reviewing the impacts on fundamental IT practices, including evolving agile and devops, upskilling test automation of AI agents, and building AgenticOps.
• As AI democratization efforts reach more departments and impact workflows, CIOs and CHROs must consider career-pathing opportunities for affected roles and jobs.

Ryan Downing, VP and CIO of enterprise business solutions at Principal Financial Group, adds, “When AI is put in the hands of more employees, it quickly shows where processes, data, and decision flows aren’t as strong as we assume. The role of IT is to help the business act on that insight by embedding well-governed, integrated AI into everyday work, so teams can improve how decisions get made, not just work faster.”

AI democratization efforts can be a force multiplier in transforming operations, enabling efficiencies, and driving growth. CIOs should learn from past technology democratization efforts. Getting democratization right requires CIOs to communicate governance requirements, champion strategic experimentation, establish AI literacy programs, and evolve change management.