What the AI found
“Your highest average overnight glucose spikes (140 mg/dL) consistently followed dinners that included a sweetened protein bar as dessert, irrespective of carbohydrate content in the main meal.”
Before
Disjointed glucose logs, no clear patterns.
After
Clear dietary impact on overnight glucose identified weekly.
The same system, three states — real screens, not a screenshot
| Date | Food Items — Glucose (mg/dL) Avg. Overnight |
| Mon Dec 4 | Chicken, Salad, Rice — 110 |
| Tue Dec 5 | Salmon, Veg, Apple — 105 |
| Wed Dec 6 | Lentil Soup, Bread, Protein Bar — 138 |
Prompt
I have weekly logs of my meals and corresponding average overnight glucose readings. Please identify any patterns where specific food items or meal types consistently precede a higher (above 120 mg/dL) average overnight glucose. Focus on patterns across at least two occurrences. Here is the data: [Paste data from Google Sheet here]
I have weekly logs of my meals and corresponding average overnight glucose readings. Please identify any patterns where specific food items or meal types consistently precede a higher (above 120 mg/dL) average overnight glucose. Focus on patterns across at least two occurrences. Here is the data: Date — Food Items — Glucose (mg/dL) Avg. Overnight Mon Dec 4 — Chicken, Salad, Rice — 110 Tue Dec 5 — Salmon, Veg, Apple — 105 Wed Dec 6 — Lentil Soup, Bread, Protein Bar — 138 Thu Dec 7 — Steak, Broccoli, Sweet Potato — 112 Fri Dec 8 — Pasta, Sauce, Protein Bar — 142 Sat Dec 9 — Pizza — 125 Sun Dec 10 — Fish, Asparagus — 108
AI
Reviewing your data, a notable pattern emerges: your highest average overnight glucose readings (138 mg/dL and 142 mg/dL) consistently followed evenings where your meal, regardless of its main components, included a sweetened protein bar as a dessert item. This correlation appears in both Wednesday and Friday entries.reduced by 18%
Avg. Weekly Glucose Variation
0 vs. 2-3 per week
Occurrences of >120mg/dL Nights
High
Dietary Insight Confidence
One Small Shift, Big Metabolic Picture
From fragmented glucose readings to a clear, actionable dietary insight, in one focused review.
A 50-year-old marketing consultant, Northern Europe, managing pre-diabetes with diet and exercise.
Tools used
The real tools used here — swap any for your own equivalent. Each links to how we’d set it up.
Starting state
Before anything was set up
Our consultant carefully tracked her glucose with a continuous monitor and logged meals in a separate app. Each morning, she’d note the overnight glucose trends, but the sheer volume of disparate data made it difficult to connect specific foods with subsequent readings. It was a rigorous, yet frustrating, exercise in diligence without insight, leading to little change in her daily routine.
| Date | Food Items — Glucose (mg/dL) Avg. Overnight |
| Mon Dec 4 | Chicken, Salad, Rice — 110 |
| Tue Dec 5 | Salmon, Veg, Apple — 105 |
| Wed Dec 6 | Lentil Soup, Bread, Protein Bar — 138 |
| Thu Dec 7 | Steak, Broccoli, Sweet Potato — 112 |
| Fri Dec 8 | Pasta, Sauce, Protein Bar — 142 |
| Sat Dec 9 | Pizza — 125 |
| Sun Dec 10 | Fish, Asparagus — 108 |
Working state
Hacks Pass, doing its job
Using a simple text export from her logging tool, she prompted a large language model to cross-reference her meal compositions with the previous night's glucose averages. The idea was to spot correlations that her manual review, spanning multiple apps and data points, consistently missed.
Prompt
I have weekly logs of my meals and corresponding average overnight glucose readings. Please identify any patterns where specific food items or meal types consistently precede a higher (above 120 mg/dL) average overnight glucose. Focus on patterns across at least two occurrences. Here is the data: [Paste data from Google Sheet here]
I have weekly logs of my meals and corresponding average overnight glucose readings. Please identify any patterns where specific food items or meal types consistently precede a higher (above 120 mg/dL) average overnight glucose. Focus on patterns across at least two occurrences. Here is the data: Date — Food Items — Glucose (mg/dL) Avg. Overnight Mon Dec 4 — Chicken, Salad, Rice — 110 Tue Dec 5 — Salmon, Veg, Apple — 105 Wed Dec 6 — Lentil Soup, Bread, Protein Bar — 138 Thu Dec 7 — Steak, Broccoli, Sweet Potato — 112 Fri Dec 8 — Pasta, Sauce, Protein Bar — 142 Sat Dec 9 — Pizza — 125 Sun Dec 10 — Fish, Asparagus — 108
AI
Reviewing your data, a notable pattern emerges: your highest average overnight glucose readings (138 mg/dL and 142 mg/dL) consistently followed evenings where your meal, regardless of its main components, included a sweetened protein bar as a dessert item. This correlation appears in both Wednesday and Friday entries.Use case implemented
The finished system, running on its own
Now, every Sunday, she performs a six-minute review. Exporting the week's data takes seconds. The AI quickly highlights any significant overnight glucose responses beyond her baseline, linking them directly to specific meal components. This allows for immediate, informed adjustments to her eating habits for the coming week, making her data work for her.
reduced by 18%
Avg. Weekly Glucose Variation
0 vs. 2-3 per week
Occurrences of >120mg/dL Nights
High
Dietary Insight Confidence
What an outside observer would notice
from 30 mins to 6 mins weekly
Time spent reviewing data
from occasional to consistent
Specific dietary insights gained
from vague to precise
Targeted meal adjustments
The stack — build it yourself
Familiar, flexible for logging diverse inputs, and easy to export as text.
Excels at identifying non-obvious correlations in textual data, especially with specific instructions.
Provides passive, real-time insights into metabolic responses without active input.
These are the tools used in this story. Any can be swapped for an equivalent you already trust.
Go deeper
Do this yourself
See how the prompt revealed the insight
This story runs on Hacks Pass. The tools and prompts above are the real build — swap any tool for your own equivalent and follow the same steps.