Marc Potter is the CEO of Actian.
AI has become increasingly available to companies and is transforming businesses. When companies all have access to the same models and AI technology, the real differentiator is how your organization can uniquely apply AI. It’s not about the data. It’s all about the human layer: the contextual knowledge that your employees have about the business and your customers that the data alone cannot capture.
In my conversations with customers, there is a debate over the context layer and where it should “live” in an organization. It can exist in systems, data, agents or within your employees.
People understand the who, what, when, where and why behind the data. They know the history and relationships between people, systems and data that are not interconnected by digital systems. Their undocumented knowledge enables them to identify anomalies. With this insight, they can help train the AI solution so it can understand the context around the data, provide proper guardrails and governance and determine the appropriate next steps rather than guessing.
Giving AI Guidance
The human layer of context is frequently overlooked when scaling AI across an enterprise, and its influence goes well beyond what someone types into a prompt. AI requires guidance such as understanding its purpose and the company’s objectives, the company’s philosophy on customer experience or the company’s rules about data usage.
AI should be treated like a non-human employee (NHE) that works alongside people and technology to accomplish a common goal or outcome. Training AI should be approached like you would onboard a new employee to the company, where a human sets the direction and purpose. Not a “human in the loop,” but “human as the lead.”
Your employees provide the human context required for AI to have understanding, purpose and governed actions. Your people know the business and can validate whether the AI solution’s data is relevant or accurate, so your best ideas come from them. That’s what makes your people your real intellectual property.
The Risks Of Using AI To Justify Layoffs
That intellectual property is at risk when leaders use AI to justify layoffs. The people leaders remove may possess precisely the context required to make their AI investments successful.
Challenger, Gray & Christmas, an outplacement and executive coaching firm, reports that AI is the leading reason for layoffs year to date. Through September 2026, “AI has been cited in 120,136 job reduction announcements, approximately 21% of all cuts.”
It’s a pattern playing out across industries: leaders implement AI, treat it as a replacement for people and only later realize what they actually lost. Gartner “predicts that by 2029, 30% of employees laid off due to replacement by AI will need to be rehired.”
Layoffs can also make companies less diverse in perspectives and experience; years of company know-how, expertise and ideas can be lost. I’ve found that diversity amplifies the value of AI outputs.
To Replace Or To Augment?
Whether AI replaces someone or augments their work often comes down to how it’s used. Used well, AI takes lower-value tasks off someone’s plate so they have more time to do higher-value work. That’s augmentation: using AI to make people more effective rather than replacing them.
Executives who assume an AI tool can absorb the work of the employees it’s replacing aren’t accounting for how much direction it needs from someone with understanding and expertise. Much of that knowledge isn’t written down anywhere. It lives in the person’s head. A sales operations manager with a 20-year career knows what a customer account discrepancy means. An inexperienced employee, stepping in if the manager gets laid off, won’t.
Without that insight, the AI may generate a plausible but incorrect explanation for a data discrepancy, and no one catches the error before it shapes a business decision.
How Leaders Can Get This Right
Above all, leaders should define the job AI is going to do before deploying it. If you don’t define AI’s job, someone else will define it for you. Skip this step, and you risk ending up with governance blind spots, low adoption and AI decisions no one trusts.
Leaders should also clearly communicate that purpose to their teams. That transparency can help ease employees’ fears about AI.
More than that, leaders should mean what they say about AI’s purpose. If they stress that AI will streamline everyone’s work across the organization, they shouldn’t turn around and use it to justify layoffs later. Doing that only undercuts the trust needed for employees to use AI in the first place.
Leaders should also give everyone on their teams access to AI tools, but pair it with intentional training and education. Even if people have valuable expertise, it won’t automatically translate into value if they don’t know how to direct AI with it. Instead, leaders should make training a strategic investment and educate employees on best practices.
Cross-team collaboration matters too. Different teams provide different pieces of context, and AI fails when one team assumes its one slice represents the whole business. Clear processes around sharing ideas keep AI projects from happening in silos. That way teams learn from each other instead of duplicating work.
Context Makes AI Work
The human layer is what makes or breaks the success of AI initiatives. Your people are the foundation of your intellectual property because they know your business and understand the undocumented intangibles about your customers. AI can’t replicate the personal experience your people have learned over years on the job. That is why you need both the human layer to provide insight and direction and governed context to produce accurate outcomes that people can trust.
AI should complement human collaboration, not replace it. The human context provided by your employees is necessary for AI to succeed and be aligned with your desired outcome. Make sure you still have them when you are ready to scale AI in your business.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
