Skip to content
Talk to usSign inJoin the preview
Comics · Stories · Episode 2

Say It Differently: The Work Before the Work

Masitus · 8 Nov 2026 · 14 pages
Cover. Say It Differently. The Work Before the Work: What AI Asks of You, Episode 2. Neha and Priya sit side by side, backs to us, each looking at the same email thread on her screen: “Re: Escalation, Brightwell.”
Page 1 of 14

Neha keeps her prompt short. Priya rewrites hers eight times. Before the 11:00 leadership review, neither answer catches the one thing that matters. The second episode of What AI Asks of You.

Read the text of every page
  1. Cover. Say It Differently. The Work Before the Work: What AI Asks of You, Episode 2. Neha and Priya sit side by side, backs to us, each looking at the same email thread on her screen: “Re: Escalation, Brightwell.”
  2. Monday, 9:30. Both are invited to a leadership review today at 11:00, on the Brightwell escalation. Both inboxes show the same thread, “Re: Escalation, Brightwell order sync”: Brightwell Retail, “Orders still aren’t syncing to our warehouse…”; Support, “Logged as P1. Investigating now.”; Platform team, “Root cause looks like a sync timeout.”; Neha, “We’ll have it fixed by Thursday.”; Brightwell Retail, “Thanks. We’re holding shipments until then.”; and 18 more messages. Same email thread. Same meeting. Same AI.
  3. Neha attaches the 23-message thread to an AI chat and types: “summarize this.” The AI replies: “The customer reported issues. The team is investigating.” Neha asks simply. The answer is vague.
  4. Priya attaches the same thread and writes: “You’re preparing a brief for the 11:00 leadership review. Audience: VP of Customer Success and the eng leads. Format: five bullets max, then a table of owners and dates. Include: current status, root cause, an owner for every fix, committed dates, open risks and customer sentiment. Tone: calm, factual, no jargon. Don’t speculate. Flag anything unresolved. Keep it under 200 words. Use only the thread below. If two messages disagree, say which one is newer. Don’t repeat the customer’s wording. Lead with what leadership needs to decide…” Priya writes a long, detailed prompt.
  5. 10:15. The brief comes back with status, owners and dates. Priya frowns: “Wait… where did the risks go?” She adds to her prompt, “Always include open risks,” thinking “Fine. I’ll spell it out.” Tap tap tap. The next brief is a solid block of text: “Risks are back… but now it’s a wall of text?”
  6. She adds “Use short bullets.” The brief is now four bullets, owners not included: “Short now. But where did the owners go?” She adds “Include owners, dates, risks, everything.” The brief runs to three pages of background, timeline, status and root cause: “Way too long…” At her desk, the clock nearly at eleven: “Good enough. It’s 10:50.” Change one word, and the answer changes.
  7. 11:00, Neha’s turn. On the screen, her summary: “The customer reported issues. The team is investigating.” The leader across the table asks: “So what do we actually do?” Neha thinks: “Um… I’m not sure.” Neha’s answer is too vague to use.
  8. 11:00, Priya’s turn. Her leadership brief: team has investigated, sync timeout; fix lands Monday; owner, Platform team. Priya says: “We’ve investigated. The fix lands Monday.” The leader frowns: “We told them Thursday. Now it’s Monday. Have we told them about the delay?” Priya thinks: “I missed it. The Thursday promise was buried in the thread.” The customer was promised Thursday. Nobody told them about the delay.
  9. What went wrong? Neha’s summary is stamped Too vague; she thinks, “I couldn’t answer a single question.” Priya’s brief is stamped Missed the Thursday promise; she thinks, “The AI missed the promise we made to them.” Neha’s answer said nothing. Priya’s answer missed a promise.
  10. That evening. Neha clicks to close the AI chat: “Forget it. I’ll write it myself.” Priya looks at version 8 of her prompt, which now adds “Flag any promise we made to the customer”: “What should I have asked to catch that promise?” Neha gives up. Priya wonders how else she could have written the prompt.
  11. Rewind. Same Monday, 9:30. Same thread, same meeting. This time, with Masitus. Both screens show Masitus: “Good morning, Neha. Your 11:00 pre-read is ready,” and “Good morning, Priya. Your 11:00 pre-read is ready,” each with an Open pre-read button.
  12. Before 11:00, both open the meeting pre-read for the leadership review. It flags a conflict on the Brightwell escalation: “Customer was promised Thursday, but the fix lands Monday. Nobody has told them yet. Update them before 11:00.” Tags: Promised Thu; Fix Mon; Customer not told. A button: Draft customer update. Masitus caught the broken promise. No prompt needed.
  13. 11:00, with Masitus. The leadership brief: root cause, sync timeout; fix lands Monday; new date communicated to Brightwell. The leader smiles: “Awesome work, team. Really great that the customer already knows about the delay.” No prompt. Just what the meeting needs.
  14. Close. The Masitus mark over red rays, with Neha and Priya below. You shouldn’t have to learn to talk to AI. New to AI or living in it. Join: masitus.com.