Throughout history, humans have been using tools to get jobs done. The “Jobs to Be Done” theory of innovation suggests that humans “hire” tools to get things done for them, and AI is no different. It is yet another tool in a long line of tools from cave drawings to the printing press to Google that help humans find and process existing information. The Jobs to Be Done theory is a useful framework to think about how to deploy AI in your job or your organization. (Here’s a great resource on Jobs to Be Done Theory)

Understanding AI’s Capabilities and Limitations

Before you can effectively “hire” AI, you need to understand what it’s good at and where it struggles.

AI excels at:

• Processing vast amounts of information quickly
• Recognizing patterns in data
• Generating content rapidly
• Providing decision support with data analysis

AI’s key challenges:

• Reliability issues and hallucinations
• Knowledge limitations and training cutoffs
• Reasoning gaps, especially with novel situations
• Integration and deployment complexity

With this SWOT analysis in place, you can identify tasks that leverage AI’s strengths while accounting for its weaknesses.

Jobs Where AI Can Add Value

There are two approaches to deploying AI in your organization:

Gap-filling approach: Look for resource constraints in your organization where AI could help meet demand (remember, this gap should be at a currently constrained resource, or you might just make the problem worse).

Best available talent approach: Think of AI as adding capability even if you don’t have an immediate perfect use case, but where the potential for future value is clear.

Four Key Job Categories for AI

Information processing jobs: Research, analysis, summarization Example: Instead of having analysts spend hours reading industry reports, use AI to summarize key findings, then have humans verify and contextualize the insights.

Pattern recognition jobs: Fraud detection, customer segmentation, anomaly detection Example: Use AI to flag unusual spending patterns in expense reports, with finance teams reviewing flagged items rather than checking every transaction.

Content generation jobs: First drafts, variations, formatting, templates Example: Have AI create initial product descriptions or email drafts, then have marketing teams refine and personalize the content.

Decision support jobs: Risk assessment, recommendation engines, scenario analysis Example: Use AI to analyze customer data and suggest next-best actions for sales teams, who then apply relationship knowledge to make final decisions.

Hiring Criteria Checklist

Before deploying AI for any job, ask these three questions:

☐ Is this job at a constrained resource? (Not enough people to meet current demand)

☐ Does this job involve processing existing information rather than creating entirely new insights?

☐ Is speed more important than perfection for this job?

If you can check all three boxes, you’ve likely found an excellent area to deploy AI.

Red Flags: Where AI Typically Fails

Avoid hiring AI for jobs that require:

• High-stakes decisions without human oversight
• Deep relationship management or emotional intelligence
• Creative problem-solving for truly novel situations
• Tasks where errors have serious legal, safety, or financial consequences

Next Steps

Start small and focused:

1. Pick one job category that aligns with your biggest constraint
2. Identify a specific, low-risk task within that category
3. Run a small pilot with clear success metrics
4. Learn from the results before scaling

Remember: AI is a tool to augment human capabilities, not replace human judgment. The most successful AI deployments happen when you clearly define the job you’re hiring AI to do and maintain appropriate human oversight.