By Kellye-Rae Campbell, Alumna of the M.S. in Information & Knowledge Strategy (IKNS) Program, School of Professional Studies
As part of my work at the intersection of privacy, AI governance, and digital transformation in the Caribbean and emerging markets, I recently attended a cross-functional strategy meeting of an established and heavily regulated organization with a footprint across multiple markets. The discussion was framed as the start of an “AI transformation” project, and the organizer opened with a sentence that alarmed me: “How can we use AI to boost productivity, drive revenue, and reduce costs?" It's an understandable impulse to want to keep up with industry advancements, but the question reflects a common misconception. It treats AI as the starting point rather than as a tool for addressing defined organizational challenges.
As the meeting progressed, we tried to map processes that AI could improve. But the team could not articulate how some of the organization's core operations worked. We were attempting to design an intelligent system for an organization that lacked a clear understanding of its own operations.
The moment was emblematic of a pattern I’d been seeing: Companies are using AI to solve poorly defined problems within broken organizations. According to a RAND 2024 study, more than 80 percent of AI projects fail, and the most common cause isn't technical: It’s leadership and problem framing.
Diagnosing Before Designing
Later in the meeting, someone suggested engaging consultants to define the organization’s AI strategy and manage its deployment, based on approaches used elsewhere.
Consultants can bring valuable perspectives and surface blind spots that internal teams are too close to see. However, there is a failing playbook I've seen, especially in emerging markets like the Caribbean: Organizations engage external expertise to define rather than diagnose. Frameworks built for mature institutions are often transplanted into organizations still building their foundations.
Consultants do not live inside organizations. They often don’t know which workarounds teams depend on, which informal practices quietly hold operations together, or which data sources could be corrupted. That knowledge resides with the people doing the work, and that is where meaningful transformation begins.
The Right First Questions
In my experience, before discussing AI tools, organizations should ask:
- What unsolved problems keep recurring?
- Which decisions take too long, and why?
- What knowledge about how work actually gets done has never been documented?
These feel mundane, but surface structural realities and expose what we in IKNS call the “information architecture problem”: Most organizations treat data, information, and knowledge as separate domains. Data belongs to IT; insights/information belong to business intelligence; and organizational knowledge resides in the employees. But AI should operate across all three layers. As organizational theorist Ikujiro Nonaka explains, critical knowledge is tacit, living in people and practice rather than in pipelines, and it cannot be automated until it is understood. If those layers remain disconnected, AI tools produce output that may sound sophisticated but is actually built on incomplete foundations.
A Different Starting Point
This led me to develop the POISE Framework, an approach to AI transformation that moves not from data to insight, but from people to intelligence.
Building on the Knowledge Discovery in Databases (KDD) model, POISE expands the approach to include organizational context.
People: Surface Knowledge
Organizations run on two sets of processes: the ones that are documented, and the ones that actually work. The first stage of transformation is capturing tacit knowledge and making it visible; not just to identify expertise, but also to establish accountability. Where does expertise reside in the organization? Which processes depend heavily on human judgment? Where do employees rely on informal practices rather than formal systems? Where will human oversight be if automation is introduced?
Operations: Map Reality
Documented processes and operational reality are rarely identical. Workflows evolve, shortcuts emerge, and informal practices fill gaps that formal systems never anticipated. Mapping operations means going beyond the process documentation to understand what is actually happening: where delays occur, where decisions rely on incomplete information, which tasks are genuinely ripe for automation, and which decisions carry enough consequence that automation would introduce unacceptable risk.
Persistent inefficiencies exist simply because no one has ever examined how different systems interact with each other. Mapping makes the invisible visible and shapes what governance needs to look like.
Information: Structure What You Have
The biggest barrier to effective AI is rarely a lack of data, but a lack of alignment between data, processes, and the organization’s operational reality.
This stage examines whether the information environment can support transformation. We must ask whether data is reliable, accessible, and properly governed, and whether it reflects what the operational mapping revealed.
Systems: Designing Conditions
Only after these foundations have been established can organizations begin designing the systems that enable transformation: i.e., the structures that connect people, operations, information, and governance into a coherent operating model. They include technology platforms, information architecture, workflows, controls, oversight mechanisms, and decision-support capabilities.
The objective is to build the conditions under which intelligent technologies can be introduced responsibly and effectively. The NIST AI Risk Management Framework provides governance structure for ensuring that what is introduced is explainable, accountable, and free from unexamined bias.
Execution: Scaling Responsibly
Successful transformation rarely begins with large-scale deployments, but with targeted pilots: controlled environments where technology is tested, governance refined, and learning happens.
Execution is never finished. It includes monitoring outcomes continuously, understanding what happens when systems fail, reassessing risk as capabilities expand, and ensuring accountability remains intact as scale increases. NASA's approach to knowledge management demonstrates that sustainable success depends not on deploying the most advanced systems, but on building the organizational intelligence to operate and adapt them responsibly over time.
What Effective AI Actually Requires
Digital transformation is often framed as a technological challenge. In practice, it is an organizational one. In my experience, transformation happens when people, processes, information, and systems evolve together within structures that maintain accountability and transparency.
I have seen this over and over again not only in my own career, but also when following the IKNS Capstone projects for the fall 2025 semester: AI has extraordinary potential. But realizing that potential requires organizations to understand themselves first: how work is performed, how decisions are made, how knowledge flows and where information is lost.
Without that understanding, AI remains fragile. And without governance, it becomes a source of risk.
Views and opinions expressed here are those of the authors, and do not necessarily reflect the official position of Columbia University School of Professional Studies or Columbia University.
About Columbia’s IKNS Degree
Columbia University’s M.S. in Information & Knowledge Strategy (IKNS) degree integrates data, people, and strategy skills for the AI age. The flexible and interdisciplinary curriculum trains leaders across the entire value chain of data-driven management: Getting the data and analytics right (e.g., AI adoption, business analytics), creating a high performing, people-centric culture (collaboration, team/project management, organizational psychology), and finally the right change management to turn your strategy into reality.
IKNS is available full-time or part-time, online or in-person on Columbia’s landmarked campus right here in New York City. To maximize opportunities for networking and community building, our online students join our New York-based students on Columbia’s campus for three in-person Residencies during their studies. The STEM-designated Master of Science degree offers International Students (F-1/J-1 visa) an opportunity for Curricular Practical Training during their studies (CPT) and three years of work authorization in the US upon completing their studies (OPT).
Students train under world-class faculty, including former and current executives from Google, IBM, NASA, and Oliver Wyman, and join a powerful global alumni network in coveted positions, including at Alphabet, Goldman Sachs, Nike, Pfizer, and the World Bank.
For more IKNS insights, news, and events, please go to our website, connect with us on LinkedIn, or attend one of our online info sessions. Visit the School of Professional Studies website to learn more about the SPS Student Experience.
