Purpose
By the end of this lesson, you will be able to use AI effectively to help turn raw numbers into a clear written narrative, while maintaining accuracy.
Lesson Explanation
AI can be genuinely useful for turning a table of numbers into readable narrative prose – but the numbers and the relationships between them need to come FROM the actual data, provided explicitly in the prompt, not generated by the AI as plausible-sounding filler.
A useful prompt for this task provides the actual figures directly: “Revenue was $2.1M in Q1, $2.4M in Q2, $2.2M in Q3, and $2.8M in Q4. Write two sentences summarizing this trend for a leadership audience” – rather than “write about our revenue trend,” which gives the AI nothing concrete to describe.
When asking AI to identify a trend or pattern in a small set of numbers, it remains essential to verify the claimed pattern against the actual numbers yourself – a claim like “revenue grew steadily throughout the year” is factually wrong if there was actually a dip in Q3, even if the sentence itself reads smoothly and confidently.
AI is also useful for adjusting the SAME underlying facts for different audiences: the same Q4 revenue increase might be described for a sales team (“we beat our Q4 target by 12%”) versus for a board audience (“Q4 results support the full-year growth trajectory”) – same facts, genuinely different framing for a different reader.
Practice Questions
1. A prompt asks an AI to “write about our revenue trend” with no actual figures provided. What is the problem with this request, per this lesson?
View Answer
This gives the AI nothing concrete to describe, risking generated content that sounds plausible but isn’t actually grounded in the real data.
2. Write a well-formed prompt providing actual quarterly revenue figures and asking for a two-sentence summary for a leadership audience.
View Answer
Something like: “Revenue was $1.8M in Q1, $2.0M in Q2, $1.9M in Q3, and $2.5M in Q4. Write two sentences summarizing this trend for a leadership audience.”
3. An AI produces the sentence “revenue grew steadily throughout the year” based on quarterly figures that actually show a dip in Q3. What does this lesson say should happen before using this sentence?
View Answer
The claimed pattern should be verified against the actual numbers – this specific claim is factually wrong given the Q3 dip, even though it reads smoothly and confidently.
4. The same Q4 revenue increase needs to be described for two different audiences: a sales team and a board of directors. Does this lesson suggest the same written description works well for both audiences?
View Answer
No – the same underlying facts can be reframed differently for each specific audience, even though the facts themselves stay the same.
5. Write a sales-team-appropriate framing of “we exceeded our Q4 revenue target by 12%,” and then a board-appropriate framing of the same fact.
View Answer
Sales team: “We beat our Q4 target by 12% – great work this quarter!” Board: “Q4 results of 112% of target support continued confidence in the full-year growth trajectory.”
6. A user provides an AI with actual monthly sales figures and asks for a narrative summary highlighting any unusual months. What makes this a well-formed request, per this lesson’s principles?
View Answer
It provides the actual concrete data directly, rather than a vague request, giving the AI real numbers to identify patterns within, rather than generating something disconnected from the actual figures.
7. An AI-generated narrative claims “Q2 was our strongest quarter,” but the actual data shows Q4 had higher revenue than Q2. What should happen with this narrative before it’s used?
View Answer
It should be corrected or verified against the real data before use, since this specific claim is factually inaccurate despite sounding confident.
8. Why does this lesson emphasize that numbers and relationships need to come FROM the actual data, rather than being AI-generated filler?
View Answer
AI-generated narrative can sound fluent and plausible while still containing invented or inaccurate claims about trends – the actual factual content needs to be grounded in real, provided data, not assumed by the AI.
9. A user provides five years of annual revenue figures and asks an AI to identify the single best-performing year and explain why it might be described that way. What kind of factual grounding does this request provide that a vague request would lack?
View Answer
Concrete, actual data points for the AI to analyze and describe, rather than requiring the AI to invent plausible-sounding figures or trends on its own.
10. A narrative describes customer satisfaction as “improving” based on scores that actually stayed flat over the measured period. What verification step from this lesson would have caught this error?
View Answer
Checking the claimed trend directly against the actual underlying scores before using the narrative, rather than trusting the fluent-sounding description.
11. Why might reframing the same facts differently for a sales team versus a board audience be a genuinely useful application of AI, rather than a form of factual distortion?
View Answer
The underlying facts remain identical and accurate in both versions – only the tone, emphasis, and framing change to suit what each specific audience needs or cares about, which is a legitimate communication adjustment, not a change to the facts themselves.
12. A finance team wants AI help turning a dense table of 50 line-item expenses into a readable narrative for a department head. What should be included in the prompt to ensure the narrative stays grounded in the actual data?
View Answer
The actual expense figures and categories themselves (or a clear, faithful summary of them), so the AI’s narrative is built directly from the real numbers rather than generic or invented content.