AI Tool Adoption Report

Overview

The AI Tool Adoption page helps you understand how AI coding assistants are being adopted and used across your engineering organization. Track adoption trends, measure the impact of AI-generated code, and identify opportunities to maximize the value of your AI investments.

Location: Navigate to AI Transformation → Adoption in the main sidebar.

What This Report Measures

The AI Tool Adoption report provides comprehensive insights into:

• Adoption rate: Percentage of developers actively using AI coding assistants

• Usage frequency: How often developers use AI tools in their daily work

• Code volume: Amount of AI-generated code being accepted and merged

• Tool preferences: Which AI coding assistants your team uses most

• Adoption trends: Whether AI usage is growing, plateauing, or declining

• Team patterns: How adoption varies across different teams and individuals

This report helps you answer critical questions about your AI tool investment: Are developers using the tools? How much impact are they having? Where should you focus adoption efforts?

Report Structure

The AI Tool Adoption report is organized into two main sections:

Adoption

Tracks who's using AI tools and how frequently:

• Active user counts (weekly and monthly)

• Adoption percentages by team and individual

• Usage frequency patterns

• Tool-specific adoption rates

What Counts as "Active Use":

• Developer generated AI suggestions

• Developer accepted AI-generated code

• Developer was actively coding (not on vacation/OOO)

Key Filters Applied:

• ✅ Excludes developers marked as out-of-office

• ✅ Requires evidence of coding activity during the period

• ✅ Only includes classified coding contributors

Impact

Shows the effect of AI tool usage on code delivery:

• Volume of AI-generated code accepted

• Percentage of codebase from AI

• Correlation with productivity metrics

• Impact summary with key highlights

Supported AI Coding Tools

Span tracks usage data for these AI coding assistants:

Code Generation Tools

• GitHub Copilot (Business and Enterprise plans)

• Cursor (Pro and Team plans)

• Claude Code (Team and Max plans)

• Augment Code

• Codex (CLI and Cloud)

• Gemini CLI

• Devin

• Async Copilot

• Codegen

Core Metrics Explained

1. Active Users (Weekly/Monthly)

Definition: Number of developers who actively used an AI tool within the selected time window.

Measurement Windows:

• Weekly Active Users: Used AI tool within the last 7 days

• Monthly Active Users: Used AI tool within the last 30 days

Display Formats:

• Raw count: "15 developers used AI tools"

• Percentage: "42% of active developers used AI tools"

What counts as "active use":

• Developer generated AI suggestions

• Developer accepted AI-generated code

• Developer was actively coding (not on vacation/OOO)

Use this to: Track adoption penetration across your organization.

2. Adoption Rate

Definition: Percentage of your active developer population using AI tools.

Formula: (Developers using AI ÷ Total Active Developers) × 100%

Example:

• 30 active developers

• 18 used AI tools this month

• Adoption Rate: 60%

Benchmarking:

• < 25%: Early adoption phase

• 25-50%: Growing adoption

• 50-75%: Majority adoption

• > 75%: Mature adoption

Use this to: Set adoption goals and track progress toward them.

3. AI Days Per Week

Definition: Average number of days per week that a developer uses AI coding tools.

Formula: Total AI Usage Days ÷ Active Weeks

Interpretation:

• 0-1 days/week: Occasional use

• 2-3 days/week: Regular use

• 4-5 days/week: Heavy, integrated use

Example:

• Developer A: 4.5 AI days/week (using AI most days)

• Developer B: 1.2 AI days/week (occasional use)

Use this to: Understand depth of adoption, not just breadth.

4. Accepted Lines of AI Code

Definition: Total volume of AI-generated code that developers accepted into their work.

What's included:

• Code suggestions accepted from AI tools

• Multi-line completions

• Whole function generations

• Code across all branches (not just merged code)

What's excluded:

• Rejected AI suggestions

• Manually written code

• Generated code from non-AI tools

Important caveat: This metric includes code from development branches that may not be merged yet, so numbers can be higher than actual merged code.

Use this to: Measure overall AI code generation volume.

5. AI Code Percentage

Definition: What percentage of your total merged codebase comes from AI coding assistants.

Formula: (AI-Accepted Lines ÷ Total Merged Lines) × 100%

Example:

• Total merged lines: 50,000

• AI-accepted lines: 8,000

• AI Code Percentage: 16%

Typical ranges:

• < 10%: Limited AI impact

• 10-25%: Moderate AI contribution

• 25-40%: Significant AI acceleration

• > 40%: Heavy AI reliance

Note: Can appear higher than actual AI-generated code because developers may accept AI code in work branches that gets refactored before merge.

Use this to: Understand AI's contribution to your codebase.

6. Total Active AI Days

Definition: Cumulative count of distinct days each developer used an AI tool.

Calculation: Counts unique calendar days with AI usage activity.

Filters applied:

• Only developers (not other roles)

• Only during active coding periods

• Excludes OOO/vacation days

Use this to: Identify your most consistent AI tool users.

How Data is Collected

Tool Integration Method

Span gathers AI adoption data through direct API integrations with each AI tool provider:

1. Daily Sync: Span connects to each tool's API daily

2. Usage Tracking: Logs which developers used which tools

3. Code Volume: Captures lines of AI-generated code accepted

4. Activity Filtering: Cross-references with developer active status

5. Data Aggregation: Combines data across all connected tools

Key Use Cases

1. Track AI Investment ROI

Measure whether your AI tool licenses are being utilized effectively.

Example: "We're paying for 50 GitHub Copilot licenses. Only 32 developers (64%) are actively using them—opportunity to improve adoption or optimize licensing."

2. Executive Reporting & Board Updates

Demonstrate AI adoption progress to leadership and stakeholders.

Example: "AI tool adoption grew from 35% in Q1 to 68% in Q2. AI-generated code now represents 22% of our merged codebase, accelerating feature delivery."

3. Team Benchmarking

Compare AI adoption across teams to identify leaders and laggards.

Example: "Team A has 85% adoption while Team B has only 30%. Let's learn what Team A is doing right and replicate it."

4. Identify Champions & Support Needs

Find developers who are AI power users (to champion adoption) and those who need support.

Example: "Sarah uses AI 4.8 days/week—let's have her share best practices. John hasn't used AI in 6 weeks—let's check if he needs training or has blockers."

5. Tool Selection & Optimization

Compare usage patterns across different AI tools to inform purchasing decisions.

Example: "GitHub Copilot has 70% adoption while Cursor has 15%. Should we consolidate on one tool or maintain both?"

6. Measure Adoption Growth

Track whether adoption is accelerating, plateauing, or declining over time.

Example: "Adoption growth has plateaued at 60% for 3 months. Time to launch a training initiative to reach the remaining 40%."

7. Correlate with Productivity

Analyze whether AI adoption correlates with changes in delivery metrics.

Example: "Teams with >70% AI adoption show 18% higher PR merge volume and 12% faster cycle times than teams with <30% adoption."

8. Ensure Equitable Access

Verify that all team members have access to and are using AI tools.

Example: "Junior developers show 45% adoption vs. 75% for senior developers. Need to ensure juniors have equal access and training."

How It Relates to Other Metrics

AI Tool Adoption works best when analyzed alongside complementary metrics:

AI-Related Metrics

Productivity Metrics

Team Health Metrics

Powerful Analysis Combinations:

AI Adoption Rate + AI Code % + PRs Per Week
= Complete AI impact picture

Team Adoption Rate + Individual Usage Frequency
= Identify champions and training opportunities

AI Code Volume + Code Quality Metrics
= Ensure AI doesn't sacrifice quality

Insights You Can Gain

Adoption Velocity

• How fast is adoption growing? Month-over-month % change

• Will we reach target adoption? Project timeline to adoption goals

• Are we plateauing? Identify when growth stalls

• What drives adoption spikes? Correlate with training or events

Penetration Analysis

• What % of developers use AI? Current adoption rate

• Who are the non-adopters? Identify teams/individuals not using tools

• Is adoption evenly distributed? Compare across teams and levels

• Are there access issues? Ensure all developers have licenses

Usage Depth

• Casual vs. heavy users? AI days/week distribution

• Integrated into daily workflow? High days/week indicates deep integration

• Tool stickiness? Are users coming back regularly?

• Usage patterns? When and how often do developers use AI?

Tool Preferences

• Which tools are most popular? Usage by tool type

• Are paid tools worth it? Compare usage vs. cost

• Should we consolidate tools? Identify underutilized tools

• Tool switching patterns? Are developers trying multiple tools?

Volume Impact

• How much code comes from AI? AI code percentage

• Is AI volume growing? Trend in AI-accepted lines

• Volume vs. adoption alignment? High adoption with low volume may indicate issues

• Productivity correlation? Do high AI volumes correlate with higher output?

Organizational Patterns

• Which teams lead adoption? Team-by-team comparison

• Level-based patterns? Junior vs. senior adoption rates

• Geographic differences? Adoption by location

• Role-based usage? Frontend vs. backend developer adoption

Common Scenarios & Interpretations

Scenario 1: High Adoption but Low AI Code Volume

What you see: 70% adoption rate but only 8% AI code

Possible causes:

• Developers are trying tools but not accepting many suggestions

• Heavy editing of AI suggestions before merge

• Using AI for exploration/learning, not production code

• AI suggestions not meeting quality expectations

• Detection model not identifying AI patterns

Actions:

• Survey developers about AI tool quality and usefulness

• Provide training on effective AI prompt engineering

• Review which types of work AI helps with most

• Consider if different tools might perform better

Scenario 2: Growing Adoption but Declining Volume

What you see: More developers using AI but less AI code being merged

Possible causes:

• Increased scrutiny/editing of AI suggestions

• Shift from simple to complex work (less AI-suitable)

• Improved developer judgment about when to use AI

• Quality controls reducing low-quality AI code

• Initial "novelty effect" wearing off

Interpretation: Not necessarily negative—may indicate mature, thoughtful AI usage.

Scenario 3: Plateaued Adoption

What you see: Adoption stuck at 50-60% for several months

Possible causes:

• Early adopters maxed out, harder to reach laggards

• Lack of training or support for non-users

• License constraints limiting access

• Some developers skeptical or prefer traditional methods

• Tool limitations for certain work types

Actions:

• Launch targeted training for non-adopters

• Identify and address specific barriers

• Have power users mentor non-users

• Share success stories and best practices

• Ensure licenses available to all who want them

Scenario 4: Uneven Adoption Across Teams

What you see: Team A at 90% adoption, Team B at 25%

Possible causes:

• Different tech stacks (AI works better for some)

• Different work types (features vs. infrastructure)

• Leadership support varies by team

• Access or licensing limitations

• Cultural differences (experimentation vs. conservatism)

Actions:

• Learn from high-adoption teams

• Provide team-specific training and support

• Address tech stack or tooling barriers

• Ensure managers encourage adoption

• Create cross-team sharing sessions

Scenario 5: High Volume from Few Users

What you see: 20% of developers account for 80% of AI code

Possible causes:

• Power users generating lots of boilerplate/repetitive code

• Uneven work distribution (some do AI-suitable tasks)

• Skill differences in using AI effectively

• Some developers more comfortable with AI

• Certain roles/projects more AI-amenable

Actions:

• Learn from power users about effective practices

• Provide mentorship from heavy to light users

• Share specific use cases where AI excels

• Recognize power users as champions

• Ensure AI-suitable work distributed fairly

Best Practices

1. Set Realistic Adoption Goals

Don't expect 100% adoption immediately:

Typical adoption curve:

• Month 1-2: 10-20% (early adopters)

• Month 3-6: 30-50% (early majority)

• Month 6-12: 60-80% (late majority)

• Beyond 12 months: 80-90% (mature adoption)

Healthy target: 70-80% adoption within 12 months of tool availability.

2. Combine Quantitative and Qualitative Data

Metrics tell you "what" but not "why":

• ✓ Pair adoption metrics with developer surveys

• ✓ Conduct interviews with high and low adopters

• ✓ Gather feedback about tool quality and usefulness

• ✓ Understand barriers to adoption

3. Focus on Meaningful Usage, Not Just Adoption

High adoption with low usage frequency may indicate superficial adoption:

Quality indicators:

• AI days/week consistently > 3

• AI code volume growing alongside adoption

• Positive developer feedback about value

• Productivity metrics improving

4. Identify and Empower Champions

Find developers who are AI power users:

• Recognize them publicly

• Have them share best practices

• Create "office hours" or mentoring programs

• Document effective AI use patterns

• Build a community of practice

5. Address Non-Adopters Thoughtfully

Don't mandate AI usage, but understand barriers:

Common barriers:

• Lack of training or confidence

• Concerns about code quality

• License availability issues

• Tool doesn't support their tech stack

• Philosophical opposition to AI

Supportive approach:

• Provide training and resources

• Share success stories

• Address specific concerns

• Ensure access isn't the issue

• Respect informed decisions not to use AI

6. Monitor AI Code Quality

High AI volume doesn't equal high value:

Quality checks:

• PR revert rate for AI-heavy PRs

• Test coverage in AI-generated code

• Review comments on AI code

• Bug rates correlated with AI usage

• Developer satisfaction with AI quality

7. Track Both Adoption and Impact

Adoption alone doesn't guarantee value:

Balanced scorecard:

• ✓ Adoption rate (are people using it?)

• ✓ Usage frequency (how often?)

• ✓ Code volume (how much?)

• ✓ Productivity impact (does it help?)

• ✓ Developer satisfaction (do they like it?)

• ✓ Code quality (is quality maintained?)

8. Celebrate Milestones

Recognize adoption achievements:

• First team to 50% adoption

• Organization-wide 25%, 50%, 75% milestones

• Individuals who effectively use AI

• Teams showing productivity improvements

• Innovation in AI usage patterns

9. Continuously Educate

AI tools evolve rapidly—keep training current:

• Regular "AI tips" sharing sessions

• New feature announcements

• Best practice documentation

• Use case libraries

• Cross-team learning forums

Setting Up AI Tool Adoption Tracking

Requirements

To use this report, ensure you have:

• ✓ AI coding tool licenses (GitHub Copilot, Cursor, etc.)

• ✓ AI tool integrations connected in Span

• ✓ VCS integration active (for code analysis)

• ✓ Developers assigned to teams in Span

• ✓ Active contributor classifications configured

Setup Steps

1. Connect AI Tool Integrations

   • Navigate to Settings → Integrations

   • Connect GitHub Copilot, Cursor, or other tools

   • Authorize Span to access usage data

   • Verify connection is active and syncing

2. Verify VCS Integration

   • Ensure GitHub/GitLab/ADO connected

   • Confirm PR data syncing properly

   • Check that code analysis is enabled

3. Configure Team Structure

   • Define teams and groups

   • Assign developers to teams

   • Set team hierarchies

4. Establish Baseline

   • Review current adoption rate

   • Note which teams/individuals are already using AI

   • Set realistic adoption goals

5. Set Up Monitoring

   • Add AI adoption to leadership dashboards

   • Schedule regular review cadence (monthly recommended)

   • Define adoption milestones and targets

Frequently Asked Questions

Q: What's a "good" AI adoption rate?
A: It depends on your goals and timeline:

• 6 months post-launch: 40-60% is strong

• 12 months post-launch: 70-80% is excellent

• Mature adoption: 80-90% (some developers may choose not to use AI)

Focus on growth trajectory and meaningful usage, not just hitting a number.

Q: Should every developer use AI tools?
A: Not necessarily. Some reasons not to use AI are valid:

• Work type doesn't benefit from AI (architecture, design)

• Tech stack not well-supported by AI tools

• Personal preference for traditional methods

• Philosophical concerns about AI-generated code

Aim for broad adoption but respect informed decisions.

Q: Why is AI code percentage higher than I expected?
A: The metric includes code accepted in development branches, which may be heavily refactored before merge. Additionally, developers may use AI as a starting point but significantly edit suggestions.

Q: How do I know if AI is actually helping productivity?
A: Look beyond adoption metrics:

• Compare PR merge rates pre/post AI adoption

• Check PR cycle time for AI users vs. non-users

• Survey developers about perceived value

• Monitor story points completed

• Track developer satisfaction scores

Q: Should I mandate AI tool usage?
A: Generally no. Forced adoption often leads to superficial usage without real value. Instead:

• Provide training and resources

• Share success stories

• Remove barriers to adoption

• Create positive incentives

• Let value drive organic adoption

Q: Can I track which specific AI tools are most effective?
A: Yes! Use the tool-specific breakdowns to compare:

• Adoption rates by tool

• Code volume by tool

• Usage frequency by tool

• Developer preferences

This helps optimize your tool portfolio.

Q: Why do some developers have high AI days/week but low code volume?
A: Possible reasons:

• Using AI for research, documentation, or learning

• Generating suggestions but not accepting many

• Heavy editing before accepting

• Using AI chat features vs. code completion

• Working on exploratory or architectural tasks

Check the specific use patterns with those developers.

Q: How does AI detection (Telltale) differ from tool usage tracking?
A:

• Tool usage: "Developer used Copilot 4 days this week" (API data)

• AI detection: "This code looks AI-generated" (ML analysis)

They're complementary—tool usage shows intent, detection shows actual AI code in the codebase.

Q: What if adoption is growing but productivity isn't improving?
A: This suggests adoption quality issues:

• Developers may not be using AI effectively

• AI suggestions may be low quality for your domain

• Team might need better training on effective usage

• Work type may not be AI-suitable

• Need more time for developers to develop AI workflows

Survey developers and iterate on training.

Q: Should junior or senior developers use AI more?
A: Both benefit differently:

• Juniors: Learn patterns, accelerate common tasks, build confidence

• Seniors: Automate boilerplate, focus on architecture, increase throughput

Both should be encouraged to adopt based on their needs.

Need Help?

For additional support with the AI Tool Adoption report:

• Visit the Span Help Center

• Contact your Customer Success Manager

• Email support@span.app

This documentation reflects Span's platform capabilities as of the current version. Features and calculations are subject to updates.