---
title: "AI impact"
canonical_url: "https://docs.getdx.com/reports/ai-impact/"
md_url: "https://docs.getdx.com/reports/ai-impact.md"
last_updated: "2026-06-18"
---

# AI impact
The AI impact report helps you measure and report on the impact of AI code assistants. To enable reporting, DX requires tagging users (i.e. using custom attributes and properties) for categorizing and tracking users' adoption levels:

- The "Group comparisons" and "Time savings" views require tagging users based on current AI usage—read more about [AI usage attributes](https://docs.getdx.com/reports/ai-usage-attributes/).
- The "Before vs after" and "Trend correlations" views require setting user properties for AI adoption dates—read more about [AI adoption dates](https://docs.getdx.com/reports/ai-adoption-dates/).

In addition, the "Time savings" view requires a completed Snapshot survey with the "AI time savings" workflow question enabled.

## Group comparisons

The "Group comparisons" report compares metrics for AI users versus others to provide insight into how AI usage is affecting developer productivity.

Below are a few tips:

- To control how your developer population is segmented, click the gear icon to access report settings where you can configure which cohorts should be treated as "Active AI users".
- To get more information about the metrics being displayed, read the tooltips displayed in the titles of each chart.
- Don't be alarmed if the metrics don't tell the story you were expecting—many factors can influence metrics.

## Before vs after

The "Before vs after" report analyzes developers' productivity metrics before and after they reach a specific level of AI usage. This view helps you understand how individual developer productivity and output changes as their AI usage increases.

## Trend correlations

The "Trend correlations" report displays how AI adoption changes over time, alongside productivity metrics. Use this view to spot trends — for example, you might see PR Throughput go up as AI adoption increases. Note that correlation does not prove causation—many factors can affect these metrics.

## Time savings

The Time savings tab provides data on time and cost savings, alongside industry benchmarks and potential unrealized savings.

There are two different calculations depending on when AI time savings data was collected:

- **After June 16th, 2025:** Responses are non-sampled. Metrics are calculated from the full set of responses.
- **Before June 16th, 2025:** Responses are sampled. Metrics are extrapolated from the sampled data.

**After June 16th, 2025**, metrics are calculated as follows:

1. DX multiples each response by 48 [weeks] to compute annualized hours saved. The sum equals the **total annualized time savings for the organization**.
2. DX divides the annualized time savings for the organization by the number of developers, calculating the **average annualized time savings per developer**.
3. DX multiplies annualized time savings for the organization by fully loaded FTE cost, divided by the number of work hours per year (2,080 hours), to calculate **total annualized dollar savings for the organization**.
4. Potential cost savings are calculated by comparing the current cost savings against the hypothetical state in which all developers are attaining the same amount of time savings as the 75th percentile of the active cohort.

**Before June 16th, 2025**, metrics are calculated as follows:

1. DX computes the average per-developer weekly time savings for each cohort (e.g., "Daily Active User", "Weekly Active User").
2. For each cohort, DX multiplies average weekly time savings by the number of developers in the cohort, then multiples that number by 48 [weeks] to compute the annualized time savings per cohort.
3. DX takes the sum of all cohorts' annualized time savings to calculate **total annualized time savings for the organization**.
4. DX divides the annualized time savings for the organization by the number of developers, calculating the **average annualized time savings per developer**.
5. DX multiplies annualized time savings for the organization by fully loaded FTE cost, divided by the number of work hours per year (2,080 hours), to calculate **total annualized dollar savings for the organization**.
6. Potential cost savings are calculated by comparing the current cost savings against the hypothetical state in which all developers are attaining the same amount of time savings as the 75th percentile of the active cohort.
---

## Sitemap

[Overview of all docs pages](/llms.txt)