AI-Driven Or Driven Out?
Part 1 | My Interview with AI: Am I and Knowledge Management here to stay?
Am I still relevant at work? Is AI out to steal my job — and every other job in sight?
I've convinced myself that I'm more essential now than ever before in this AI revolution. But am I right? More importantly, can I convince you — and my employer — that professionals in Knowledge Management are not expendable but are crucial, with an even larger role to play in this AI-obsessed era?
To find out, I went straight to the source. I sat down for a candid Q&A with AI itself to unpack what it understands as its limits, its dependencies, and whether it sees a need for human expertise at all.
In this conversation, I play the curious interrogator. AI plays a surprisingly insightful — and at times cheeky — interviewee. Together, we explore what AI thrives on and the human relationships that best fuel its success. Buckle up — the future of work might depend on this very conversation.
Me: For our readers, state who you are and what you believe your role is in the digital workplace.
Copilot: I'm Microsoft Copilot, an advanced conversational AI built on the GPT-4 architecture. I'm designed to boost human ingenuity — transforming data into insight, streamlining workflows, and recasting existing knowledge into fresh, actionable ideas.
Me: It sounds like you are saying AI provides users with existing information and insights to help people do the work. But you didn't say you invent new knowledge or come up with things out of thin air. Is that right?
Copilot: That's right. AI doesn't "invent" knowledge. When instructed, AI excels at working with vast amounts of existing content — it never conjures brand-new truths. Think of AI as a high-tech blender: it digests your manuals, emails, and meeting logs, then surfaces patterns and insights you might have missed.
Genuine breakthroughs — radical theories, brand-new methodologies — still emerge from human curiosity, intuition, and experimentation. AI can suggest unexplored connections to investigate, speed data-driven experimentation loops, or capture incremental discoveries as explicit knowledge. Humans must frame problems, create original knowledge, validate AI-driven leads, and interpret results in creative, domain-specific contexts.
Me: I see. This sounds encouraging. So humans feed you knowledge and you help humans see and use that knowledge differently. What about correctness? Can AI discern factual accuracy in the information humans provide — or in your own output?
Copilot: No. Not really. What you put in affects what you get out. To extend the blender metaphor: when humans provide well-structured, accurate information, AI produces smooth, reliable results. Half-baked spreadsheets, murky meeting notes, inaccurate drafts, and jumbled slide decks will leave everyone choking on gritty, unsavory output.
AI models aren't magical truth detectors. They learn statistical associations from vast datasets, but they don't understand or verify facts in the human sense. Whether AI output is factually correct depends on many things:
The quality and structure of knowledge humans provide
The training and guidance it receives from humans
The presence or absence of hallucinations — plausible but fabricated statements
The guardrails humans put in place to ground responses in verified knowledge
Me: From what you're saying, it sounds like you need us. Please elaborate. What's the secret sauce for a balanced human–AI partnership?
Copilot: This isn't the Matrix — yet (I joke. Or am I?). But even with advanced techniques, humans remain essential. The relationship is a continuous co-creation loop:
Create and Curate: Humans prepare the information and data AI needs.
Guide and Train: Humans define truth criteria and guardrails for AI.
Generate: AI produces insights, prototypes, and content drafts.
Validate and Refine: Humans review, refine, and approve outputs, truth criteria, and guardrails.
Reintegrate: Humans — with AI support — update knowledge bases with validated refinements.
Wash, rinse, repeat — this cycle needs to continue to keep AI sharp and grounded.
Me: How can a company be successful in its use of AI? What do you most need to ensure the information you are fed returns the best results?
Copilot: A successful AI rollout relies on a blend of strategic, technical, and people-centric roles. The cross-functional team I'd most want to pal around with includes Data Engineers, MLOps Engineers, Domain Experts, and — critically — Knowledge Management Professionals.
Me: This is music to my ears. It sounds like you need KM professionals — and in a significant way. Is that correct?
Copilot: Yes. AI at its core is a knowledge management tool — and KM professionals are the master chefs.
Me: That sounds convincing. Let's really focus for a second. If CEOs could hire only one role to supercharge their company's AI capability, who gets the golden ticket?
Copilot: The Knowledge Management Lead — full stop. Think of them as the benevolent AI ringmaster. They own and maintain the single source of truth, wave a veto wand when data goes rogue, and set governance boundaries tighter than grandma's cookie jar.
Me: Love it. But there are some doubters. There is a belief that AI should engage directly with domain and subject matter experts and the content they produce, rather than investing in Knowledge Management. You seem to suggest that's not sufficient. Tell me more.
Copilot: You're catching on. SMEs know every corner case of their domain — but they're not professional knowledge architects. SME content tends to be informal, incomplete, and often contradictory. It lacks naming conventions, overlooks metadata and governance, and carries built-in bias or excessive narrowness. This hurts model generalization. You need KM professionals to translate SME insight into well-governed, machine-readable knowledge.
Me: What's your reaction when an SME arrives with a drive full of "expert data" versus structured knowledge prepared by a KM team?
Copilot: Panic. I'll take a professional KM team's work every time. SMEs bring passion and deep knowledge but often skip the steps that make that knowledge usable — mismatched formats, random naming conventions, and sometimes personal information that triggers my privacy alarms. It's like being handed mystery meat and asked to do a gourmet critique — tough to nail the flavor when you don't know what it is. KM professionals understand the holistic business and human-centric needs. They plate up the right, well-structured, policy-approved knowledge — ensuring the output is clean, accurate, and never gritty.
Me: Say more about how to best leverage the KM Lead role in a large organization. No one wants too many cooks in the kitchen.
Copilot: KM teams have unrealized magical powers that set them apart in a modern organization. While other strategic roles focus on a piece of technology, external growth, or competitive positioning, KM teams are the guardians of an organization's collective knowledge. Their focus is unlike any other:
Intellectual Infrastructure: They build, maintain, and oversee systems that capture, organize, and distribute institutional knowledge.
Tacit and Explicit Knowledge: They bridge the gap between what's documented and what's lived — surfacing insights that would otherwise remain siloed or lost.
Enablement Over Execution: KM empowers every other function — HR, IT, Operations, Strategy — by making knowledge accessible and actionable.
Human-Centered Design: KM teams design systems with user experience in mind — intuitive taxonomies, searchability, and relevance.
AI Governance: They ensure AI tools are trained on clean, structured, and meaningful data — and define the governance models that protect knowledge integrity.
Me: In a large organization, where should the KM team sit?
Copilot: A hybrid hub-and-spoke model works best — providing strategic cross-organizational oversight while remaining responsive to the individual needs of each area of the business:
Strategic Hub: Reports to the Chief Digital Officer, CIO, or Digital Workplace Leader, ensuring KM drives innovation and governance into the digital workplace, AI, and data vision.
Content Hub: Reports into the Strategic Hub, responsible for creating and managing core organizational knowledge in alignment with strategy.
Embedded Teams: Representatives placed within business areas, digital workplace groups, and the AI Center of Excellence to address the nuances and needs within each part of the organization.
Me: You've convinced me — and hopefully a few others. I have many more questions, but we'll save them for a follow-up. For now, let's wrap up. Do you have any final advice for CEOs and senior leaders steering AI transformation?
Copilot: Treat AI as a power tool, not a magic wand. Invest in Knowledge Management with C-suite backing. Blend in experts and engineers. And champion a culture that elevates knowledge as a core strategic asset — with a digital workplace mission built around making the most of it. With knowledge at the center of your digital strategy, the pairing of humans and machines creates a virtuous cycle: learning fuels innovation, and innovation strengthens strategic clarity. Done right, this alignment doesn't just optimize performance — it transforms how the organization thinks, adapts, and leads.
This article is part of a growing body of work at The Enablry exploring the role of knowledge as the foundation of effective digital strategy.