What the AI found
“AI found that clients consistently reported elevated stress levels (avg. 7.2/10) on days where their average glucose variability exceeded 2.5 mmol/L, a pattern previously overlooked due to manual data aggregation.”
Before
Disjointed client data, manual weekly reviews
After
Integrated weekly metabolic reports, AI-driven insights
The same system, three states — real screens, not a screenshot
| Client | K. Janssen |
| Date | 2024-05-18 |
| Avg Glucose | 5.8 mmol/L |
| Glucose StdDev | 1.9 mmol/L |
Prompt
Here is anonymised daily client data for the last month, correlating average glucose, glucose standard deviation, reported stress levels (1-10), and total carbohydrate intake (grams). Identify any significant correlations, particularly between stress and glucose variability. Data: [Pasted 30 days of client data: date, avg_glucose_mmol, glucose_stddev_mmol, stress_level_1_10, carb_grams_total]
Here is anonymised daily client data for the last month, correlating average glucose, glucose standard deviation, reported stress levels (1-10), and total carbohydrate intake (grams). Identify any significant correlations, particularly between stress and glucose variability. Data: [Pasted 30 days of client data: date, avg_glucose_mmol, glucose_stddev_mmol, stress_level_1_10, carb_grams_total]
AI
Analysis reveals a notable pattern: on days when reported stress levels were 7 or higher, the average glucose standard deviation increased to 2.8 mmol/L, compared to 1.7 mmol/L on lower stress days. Specifically, 8 out of 10 days with glucose variability exceeding 2.5 mmol/L coincided with a reported stress level of 7.2 or above.2.1 mmol/L
Avg. Glucose Variability (last 7 days)
6.5/10
Avg. Stress (last 7 days)
8
High Variability Days (last 30)
Streamlining Metabolic Insight: A Practitioner’s New Workflow
A practitioner moves from fragmented data collection to an integrated, AI-assisted client review process for metabolic health.
A nutritionist running a small, busy EU practice, focusing on client metabolic health.
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
Before All-Access, the nutritionist faced a common challenge: client data scattered across multiple platforms. Continuous glucose monitor (CGM) readings lived in one app, food diaries in another, and activity logs elsewhere. Preparing for weekly check-ins meant laborious manual synthesis, often missing subtle patterns.
| Client | K. Janssen |
| Date | 2024-05-18 |
| Avg Glucose | 5.8 mmol/L |
| Glucose StdDev | 1.9 mmol/L |
| Meals Logged | 3 |
Working state
All-Access, doing its job
To address this, the nutritionist began exporting raw data from client CGMs and food tracking apps into a Google Sheet. She then used a prompt in Gemini to analyse correlations between glucose metrics, reported stress, and macronutrient intake. This allowed for quick, quantitative insight generation.
Prompt
Here is anonymised daily client data for the last month, correlating average glucose, glucose standard deviation, reported stress levels (1-10), and total carbohydrate intake (grams). Identify any significant correlations, particularly between stress and glucose variability. Data: [Pasted 30 days of client data: date, avg_glucose_mmol, glucose_stddev_mmol, stress_level_1_10, carb_grams_total]
Here is anonymised daily client data for the last month, correlating average glucose, glucose standard deviation, reported stress levels (1-10), and total carbohydrate intake (grams). Identify any significant correlations, particularly between stress and glucose variability. Data: [Pasted 30 days of client data: date, avg_glucose_mmol, glucose_stddev_mmol, stress_level_1_10, carb_grams_total]
AI
Analysis reveals a notable pattern: on days when reported stress levels were 7 or higher, the average glucose standard deviation increased to 2.8 mmol/L, compared to 1.7 mmol/L on lower stress days. Specifically, 8 out of 10 days with glucose variability exceeding 2.5 mmol/L coincided with a reported stress level of 7.2 or above.Use case implemented
The finished system, running on its own
Now, each Sunday, the nutritionist pulls updated client data into a pre-formatted Google Sheet. Gemini automatically analyses the new entries, highlighting notable trends or anomalies related to metabolic markers. This streamlines her client preparation, ensuring she focuses on actionable insights during consultations.
2.1 mmol/L
Avg. Glucose Variability (last 7 days)
6.5/10
Avg. Stress (last 7 days)
8
High Variability Days (last 30)
What an outside observer would notice
from 45 to 15 minutes
Weekly review time reduced
increased by 30%
Client insight depth
up 20%
Client engagement in data review
The stack — build it yourself
Its ability to quickly process tabular data and identify nuanced patterns from a natural language prompt was essential.
Provided a universally accessible and easy-to-update central location for disparate client data streams.
These applications are standard for continuous glucose monitoring, providing the foundational metabolic data.
These are the tools used in this story. Any can be swapped for an equivalent you already trust.
Go deeper
Do this yourself
Read the practitioner’s full workflow
This story runs on All-Access. The tools and prompts above are the real build — swap any tool for your own equivalent and follow the same steps.