Stop building AI features and start onboarding AI agents
Has your AI experimentation delivered measurable gains?
Most edtech and education companies today are busy building AI features. Accelerating content creation. Automating grading. Recommending content. Or launching chatbots to users. These are useful, but they deliver modest measurable operational gains and only fleeting competitive advantages.
By contrast, AI agents represent a fundamentally different category of opportunity. And organizations that have recognized this are already pulling ahead—in sectors like finance, insurance, and pharma.
So, my advice is: if you want to get ahead, stop building AI features and start onboarding AI agents.
AI features vs. AI agents
The differences between early experimentation with GenAI features and deploying AI agents is profound:
- Features respond. Agents act.
Features answer a query or automate a step. Agents take responsibility for delivering outcomes. - Features add convenience. Agents deliver value.
Features save minutes. Agents change economics—producing 10–100x efficiency or cost improvements. - Features are invisible tools. Agents are teammates.
Features sit in the background. Agents are transparent, auditable, and accountable—with defined jobs, resources, and performance reviews.
Lessons from other sectors
Through my work with Faculty—Europe’s largest applied AI company—I’ve seen first hand how companies in high-stakes sectors are scaling AI agents with extraordinary results. The lessons for successfully deploying AI agents are strikingly human-centered:
- Define their job: document the agent’s role, responsibilities, and level autonomy—just as you would for a new hire.
- Give them context: equip the agent with the resources, data, tools, and policies needed for them to succeed—just as you would for a new hire.
- Oversee their launch: design the interactions you want with your agent, and them with other agents, build in transparency (logging actions and decisiosn), and set them goals.
- Help them to grow: assess their performance, give them regular feedback, and expand their responsibility as your confidence and their performance grow—just as you would for a new hire.
This is how leading companies in finance, insurance, and pharma are embedding AI agents into their daily operations—from compliance monitoring to risk assessment. These companies are not dabbling with AI features pilots. They are deploying AI digital teammates to deliver outcomes in some of the toughest regulatory environments. These companies chose to stop building AI features and start onboarding AI agents.
Why this matters for education companies
Education companies face many of the same challenges as these industries: high stakes, complex processes, and trust as a prerequisite for adoption. These requirements are the sweet spot for (well-designed) agents. Agents present transformative opportunities to edtech and education companies both internally and externally:
- Internal operations—automating complex, labor-intensive, costly processes where improving KPIs like cost, speed, and accuracy remain elusive.
- Next-generation product experiences—creating entirely new and dynamic learner and educator experiences where personalization, flexibility, navigating complexity, and delivering better outcomes really count.
I’ll explore that second opportunity in my next post.
My strategic question to you
For CEOs, CPOs, and COOs in education, the question is simple: have the AI features you’ve developed delivered major gains in your operations or market share? If not, now is the time to stop building AI features and start onboarding AI agents. The difference is not semantic, it’s strategic. Features keep you in the race. Agents change the race.
Need help to refine your business, product, and AI strategy?
If you’d like support assessing your business or product strategy—and the role AI can or should play—please reach out. We’d love to help. We also welcome your feedback or questions on this blog—just use the form below and start “Your challenge” with the word “Feedback.”