Fact-Checking and Avoiding 'Botsitting'

The Friday Afternoon Dilemma

It’s 3:00 PM on Friday. You have a high-priority client update due in two hours. You use an AI co-pilot to draft the email from your notes.

The result? It's fast, but it’s wrong. The dates are mixed up, the tone is aggressive, and it missed the main point.

Imagine it’s Friday afternoon, and the clock is ticking toward a 5:00 PM deadline. You ask your AI co-pilot for a quick draft, but as you read the three paragraphs it produced, your heart sinks. It mixed up the project phases and hallucinated a deadline that doesn't exist. Thirty minutes later, you're still editing. You realize that if you had just written it yourself, you’d be done by now.

• AI can generate drafts in seconds but often lacks context.
• Fixing a bad draft can sometimes take longer than writing from scratch.
• Recognizing the 'time-sink' early is a critical workplace skill.

What is 'Botsitting'?

Botsitting is the hidden labor of supervising AI. It includes re-prompting, feeding missing context, and cleaning up 'confident-but-wrong' answers.

We call this 'Botsitting'—the hidden labor of supervising AI. Recent studies show that many professionals lose more than six hours a week simply cleaning up poor AI outputs. This creates a productivity paradox: you feel like you're working faster, but you're actually stuck in an endless loop of 'fixing the bot' instead of moving on to the next task.

• Professionals lose an average of 6 hours a week 'babysitting' AI.
• The Productivity Paradox: Feeling fast, but staying stagnant in output quality.
• Botsitting happens when we supervise the machine more than we lead the task.

The 'Scrap or Save' Framework

To remain the pilot in command, use the 5-Minute Audit to decide if a draft is worth your time.

• Check Core Facts: Verify dates and numbers immediately.
• Evaluate 'Sanding' Effort: Estimate the fix time.
• The 50% Rule: If fixing takes >50% of writing time, scrap it.

To stay in control, you need a framework. First, perform a 5-minute audit. Start by checking core facts—dates, names, and numbers—against your source documents. Next, evaluate the 'sanding' effort. If the tone is off or logic is tangled, estimate the fix time. Here is the rule of thumb: if fixing the draft will take more than half the time it would take to write it from scratch, scrap it immediately.

• Never trust an AI-generated stat without a primary source.
• If 'sanding' the draft takes too long, start over.
• Scrapping a bad draft is often faster than fixing it.

Decision Point: Fix or Scrap?

Scenario: You asked AI to summarize a 20-page legal contract. The summary is 2 pages long but uses terms not in the original and missed the 'Termination' clause entirely.

What is your strategy?

Let's put the framework to the test. You have a summary that missed a critical clause and added fake legal terms. Type your plan: would you fix this line-by-line, or scrap it and try a different approach? Explain why.

• Identify major hallucinations or omissions.
• Apply the 50% rule to complex documents.
• Learn to re-prompt or go manual when AI fails the core task.

The Pilot's Final Checklist

Before hitting 'send' on any AI-assisted work, shift your mindset from proofreading (grammar) to verification (truth).

As you conclude this lesson, remember: proofreading looks for typos, but verification looks for truth. You are the pilot in command. Use AI to get your wheels off the ground, but never let it fly the plane alone. Always cross-reference AI claims with your trusted internal data before you share your work.

• Proofreading = Grammar/Typos.
• Verification = Accuracy/Logic.
• You are the Pilot; the AI is just the co-pilot.