As artificial intelligence (AI) continues to become common place across workplaces regardless of industries, a common misconception has emerged: AI is often treated as just another tool, akin to legacy software-as-a-service (SaaS) platforms like CRM systems. Or, even worse, people expect AI to be magical in it's ability to immediately know everything or do anything. I'll focus on the former (legacy tools comparison and in the future tackle the do everything angle)
Unlike traditional software tools, managing AI is more akin to managing a junior employee. This distinction is crucial for organizations to understand if they are to harness the full power of AI.
The Junior Employee Reality
Imagine hiring a junior employee to handle a task like answering a large volume of questions about a specific program. A junior employee would start with a base level of knowledge, maybe some training, and the ability to follow instructions. However, they would require guidance, continuous feedback, and a learning curve to adapt to the nuances of the program and the organizational culture. This probably takes 4-6 months if we are being honest and with the way trends are going they may only be there 12 months.
But, for example, let’s say the program is a philanthropic initiative aimed at supporting small nonprofits through grants. The junior employee might start by answering basic questions: What is the eligibility criteria? How do you apply? What documents are needed? However, as more complex inquiries come in—questions that involve specific scenarios or interpretations of guidelines—the employee might struggle.
This is where you, as a manager, step in. You would provide them with the necessary context, refine their responses, correct any misunderstandings, and over time, they would get better at handling these complexities on their own.
AI functions in a remarkably similar way. When deployed to answer questions about a program, AI will start with its initial training data (maybe your website, etc) —similar to the junior employee’s onboarding. But, much like that employee, AI requires supervision, training, and fine-tuning to handle the subtleties and complexities of real-world inquiries effectively. A challenge that many might not say out loud given large ($$$) investments in wiki like solutions such as Sharepoint or Notion (which are amazing in their core use cases, just maybe not leading on externally facing roles or opportunities for organizations).
Differentiating AI from Legacy SaaS Tools
The key difference between managing AI and using legacy SaaS tools like Google Workspace lies in the nature of the relationship. SaaS tools are static; they do what they are programmed to do, and while they might be customizable, they do not learn or adapt in the same way that AI can. Managing these tools is akin to maintaining a piece of equipment: it’s about ensuring they are working correctly, up to date, and integrated with other systems. The interaction is largely transactional.
In contrast, managing AI is dynamic and developmental. AI, like a junior employee, can learn and improve over time, but only if it is given the right kind of guidance. This means that managing AI involves:
Now, lets use another example, helping with a conference.
Imagine you are hired to manage communications for a major trade association. The association is about to host its annual conference, which draws thousands of attendees, vendors, and speakers. You’ve deployed an AI system to handle the large volume of inquiries expected in the lead-up to the event.
At first, the AI handles straightforward questions effortlessly: “What are the conference dates?” “How do I register?” “Where is the event being held?” These are the equivalent of the basic questions a junior employee could answer with minimal training.
However, as the event approaches, the inquiries become more complex: “How do I change my registration from in-person to virtual?” “Will there be vegetarian options at the networking lunch?” “Can you provide more details on the keynote speaker’s background?” This is where managing the AI becomes critical.
Just as you would guide a junior employee, you need to feed the AI more detailed information, adjust its responses based on the feedback you receive, and even intervene when it’s unclear how the AI should respond.
Also, if new issues arise—say, a sudden change in the event schedule due to unforeseen circumstances—the AI needs to be quickly updated with this new information to respond accurately to attendee questions. This real-time adaptability is something you wouldn’t expect from a legacy SaaS tool, but it’s essential for both AI and a junior employee.
One callout in a world of AI wrappers and simple applications, it's VERY hard to course correct models and data you don't fully control. Buyer beware if your boss tells you to buy a $20/mo subscription and create enterprise value.
In this scenario, managing the AI requires the same attentiveness and leadership that managing a junior employee would demand. You are responsible for ensuring the AI is properly trained, continuously improving, and capable of handling both routine and complex inquiries. The AI, much like the junior employee, becomes more effective over time as it learns and adapts to the unique needs of the trade association and its event.
Managing AI is not a hands-off, transactional process like managing traditional SaaS tools. Instead, it’s a dynamic relationship akin to managing a junior employee—one that requires ongoing training, feedback, and guidance. By understanding this distinction, organizations can better integrate AI into their operations, ensuring that it not only meets but exceeds the expectations placed upon it. Just as a well-managed junior employee can grow into a valuable asset, a well-managed AI system can become an indispensable tool, capable of handling complex tasks with a high degree of efficiency and accuracy.
Are you ready to manage AI + humans?
