Developer Intelligence

Measure engineering performance. Prove AI’s impact.

Connect Engineering Output, code quality, agent effectiveness, and AI spend. See what your team delivers and where to improve.

Try Now Get Pricing

Engineering Overview

Example · Four complete weeks

Engineering Output

Complexity-adjusted work delivered

1,920 pts

↗ 20%

Compared with the previous four weeks

Engineering Output points

HumanAI-assistedAgent-authored

1,220 points from AI-assisted and agent-authored work.

63.5%

AI Output Share

18h

Median PR cycle time

Quality & Reliability

5.2%

30-day code rework

Change failure rate2.1%

Agent Effectiveness

81 / 100

Across scored sessions

Claude Code: 84
Codex: 81
Cursor: 78

Agent Readiness

8 / 12 repos

At Level 3 or above

4 repos below Level 3

AI Economics

67.6 pts

Engineering Output / $1K

AI spend$28.4K

Estimated net ROI4.8×

What’s limiting our Engineering Output?Explore with Ask AI

Start with review wait. Payments has the longest review queue in this example. Compare review wait, session verification, and repository readiness to decide what to improve next.

Explore Ask AI

Trusted by AI-forward enterprises

AI Impact

Connect AI spend to Engineering Output, delivery speed, and code quality.

Engineering Output / $1K · Engineering Output points per $1K of AI spend

Quality · AI Quality Score and defect rate on AI-assisted code

Velocity impact · PR cycle time, AI-assisted against your own baseline

AI ROI · estimated net value divided by invoiced AI spend

Explore AI Impact

AI impact, team by team

Example · Four complete weeks · Engineering

67.6 pts

Engineering Output / $1K of AI Spend

AI Quality Score, AI-assisted code

−38%

Velocity impact, median PR cycle

4.8×

Estimated net ROI

Team	AI Code Share	Engineering Output / $1K	Quality	Velocity	AI ROI
Platform	72%	80.0 pts	94 ↗	14h	5.6×
Payments	58%	80.0 pts	92 →	17h	5.1×
Infra	46%	60.0 pts	90 ↗	19h	4.4×
Growth	31%	44.4 pts	88 →	26h	2.9×

Engineering Output / $1K is Engineering Output points per $1K of AI spend. Velocity is median PR cycle time. Team level by default; individual views are coaching context for the engineer.

Larridin Router

Choose models for the task. Compare cost, speed, and quality.

Intelligent routing · match requests to models using your routing policy

Model choices · compare model tiers and cache usage across teams

Routing policies · set the cost and quality priorities for your coding workflows

See the Router

Router impact

Example · Last 12 weeks

$18,420

saved, 31% below pinned cost

4.6

PRs per engineer per week

2.1%

defect rate, routed work

Where AI coding work ran

Routed 64%Pinned 17%Off-router 19%

Jun 1Jun 22Jul 13Aug 3Aug 17

Routed vs pinned, same tasks

Output per engineer4.6 vs 4.4+4.5% ↗

Code quality87 vs 86within noise →

Defect rate2.1% vs 2.4%−0.3 pts ↘

Cost per task$2.68 vs $4.64−42% ↘

Agent Effectiveness

Review coding-agent sessions, outcomes, and cost.

Session evidence · review captured steps, token usage, and estimated cost

Comparable scores · compare similar sessions using the same scoring rubric

Practices to improve · find examples of effective prompting, steering, and verification

Explore Agent Effectiveness

Agent Traces

App: All

All · 1,529 Merged PR · 80 Needs attention · 166

ScoreSessionAppPRsEnded ↓

53Document router-prompt scoringClaude Code1Aug 31

77Larridin brand voice guidelinesClaude Code0Aug 30

57Optimize dbt CI pipelineClaude Code0Aug 29

80Git worktree setup automationClaude Code0Aug 29

67HubSpot lead generation syncClaude Code0Aug 29

70AI Impact dashboard filtersClaude Code0Aug 29

60GCP dbt job log analysisClaude Code0Aug 29

63Plugin and authentication flowClaude Code0Aug 28

57Local persistence for draftsClaude Code0Aug 28

67Spend Mapping UI migrationCodex0Aug 28

37Performance impact of the routerClaude Code0Aug 28

70AI Spend Intelligence exportCursor0Aug 27

Prompts

30h 23m open

Engaged

2h 40m

4 re-warms

Cost

$37.54

Opus 5 · Fable 5

Tokens

76.5M

99.5% cached

EvaluationRecommendationsDetails

Overall score

81 / 100 ▲ 8 vs team's August average of 72

A strong session. The engineer opened with an explicit goal and asked for a plan, held scope with five targeted corrections, and the agent verified its work before reporting completion.

Show full reasoning →

16 citations · 9 prompts, 7 agent turns

Themes

team average

Prompt clarity77

Asks are scoped with an explicit session goal, constraints and an upfront plan request.

Prompt quality82

Each prompt combines the goal, a performance target and constraints for the CI speed-ups.

Session steering91

Catches drift quickly with targeted corrections and clear scope boundaries.

Engineering Performance

Measure delivery from code to production.

Measure how engineers and AI agents deliver software, from code changes through production, with visibility into velocity, quality, and reliability.

AI Assisted Code Share Engineering Output Defect Rate Reliability AI Quality Score

Explore Engineering Performance

Telemetry across the SDLC

Prompt to production · Last 12 weeks

Prompt

Code

Review

Merge

Deploy

Production

Sessions captured

412

this quarter

Claude Code · Codex · Cursor · Copilot

AI Assisted Code Share

41%

+6 pts ↑

of merged lines written with AI

AI Quality Score

+4 ↑

durability, review load, reverts

Engineering Output

1,920

+18% ↑

Output points, four complete weeks

Reliability

99.7%

flat →

change failure rate 3.2%

Defect Rate

2.1%

−0.4 pts ↓

bugs per merged PR, 30 days

3 of 4

teams above the illustrative comparison on AI Assisted Code Share

Growth

is the outlier: 33% code share, 38% of sessions close without a test run

Anonymized by default.

Coaching context for the engineer.

Ask AI

Ask an engineering question. Follow the evidence.

Ask about engineering metrics from your connected assistant. The MCP connector is in Beta, with access based on your organization and role. See the MCP setup guide for supported clients.

Explore Ask AI

Claude larridin connected

We rolled Copilot out to the three mobile teams in July. Has it paid for itself yet?

larridin query_larridin_analytics domain: ai_coding metric: cost_summary 12 weeks Running 0.8s

Two of the three are ahead. iOS and Android return 2.3× and 1.9× on seat cost. Wearables sits at 0.6× and carries the slowest review time of the three.

Team	Seats	AI spend	PRs per eng / wk	Review time	Return on seat cost
iOS	14	$4,180	3.9	19h	2.3×
Android	19	$5,640	3.6	22h	1.9×
Wearables	8	$2,390	2.1	34h	0.6×

Sources: ai_coding/cost_summary and velocity/summary, 12 weeks, 41 engineers across 3 teams.

Which of my teams get the least out of agents, and what should each one fix first?

larridin query_larridin_analytics domain: agent_effectiveness metric: overview 8 weeks Running 1.1s

Three teams sit below the org median of 71. Each is losing effectiveness somewhere different, so the first fix is different too.

Wearables

Require a test run before the agent closes. 44% of its sessions end unverified, and those produce 2.7× more follow-up fixes.

Growth

Open every session with a stated goal. 63% of its sessions start without one. Sessions that state a goal take 31% fewer iterations.

Billing

Name the files in scope up front. Its diffs run 2.4× larger than the teams above the median, and clear review a day later.

Sources: agent_effectiveness/overview and team_distribution, 8 weeks, scoped to your org chart.

Ask about adoption, spend, velocity, quality or agent effectiveness

Works in

Claude Claude Desktop Cursor VS Code Compatible remote MCP clients

Security

Enterprise-ready security.

SOC 2 compliant

SOC 2 Type II certified with audited controls across access, infrastructure, and data protection.

GDPR compliant

Clear data retention, deletion, and classification policies with customer control over personal data lifecycle.

Audit logs

Full audit trails across access, system activity, and changes, supporting investigation and compliance.

Secure AI

Model-provider retention controls are separate from Larridin session storage. Review data handling and retention terms in the Trust Center.

Fine-grained RBAC

Organization and role-based access controls support your data-access policies. Review enabled tools and scopes with your administrator.

SSO & SCIM

Secure authentication via SSO and automated provisioning with SCIM across your identity provider.

Review the Larridin Trust Center

AI ROI by department

Example · Last quarter · Company

3.2×

Blended AI ROI, all departments

$2.4M

Net value, annualized

Departments measured

Department	Weekly active	Hours back	AI ROI
Engineering	84%	6.1h	4.8×
Customer Support	71%	4.4h	3.6×
Marketing	66%	3.8h	3.2×
Sales	58%	3.2h	2.9×
Finance	43%	1.9h	1.8×
Legal	31%	1.2h	1.4×

Hours back is per person per week. Net value is hours returned at loaded cost plus measured revenue effects, over invoiced AI spend. Adoption from observed usage, department by department.

Beyond engineering

Measure AI impact across your company.

Connect engineering results to the company view.

Track AI adoption, fluency, and impact across Support, Sales, Marketing, Finance, and Legal. Review each department’s results alongside its AI spend.

AI AdoptionWhich AI tools each department actually uses, by team and role, and what they cost. AI Opportunity DiscoveryAI-touched workflows mapped from observed behavior, so leaders know where to invest next. AI FluencyThe skills behind the adoption numbers. Find the power users and close the gaps. AI ImpactProductivity, sales performance and cost control for every function, tied back to AI spend.

Explore the Larridin platform

Developer Intelligence

Measure developer productivity across the SDLC

The complete picture

What does Developer Intelligence measure?

Developer Intelligence connects engineering performance, AI coding activity, and AI spend. Follow the work from an agent session through merged code and production outcomes, then decide what to improve.

Engineering leaders

Review Engineering Output, quality, reliability, and AI investment across consistent teams and time periods.

Engineering managers

Investigate delivery changes, repository readiness, and coding-agent practices with the evidence behind each signal.

Engineers

Use session feedback and repository guidance to improve how you work with agents. Read the changes behind an aggregate metric.

Questions about Developer Intelligence

How is developer productivity different from AI coding ROI?

Developer productivity concerns engineering work delivered and the conditions that help teams deliver it. AI coding ROI compares an estimated benefit with AI spend. Engineering Output, cycle time, quality, and reliability provide complementary evidence; no single metric establishes business value.

Does Developer Intelligence measure both human and agent work?

Engineering Performance measures eligible merged work and separates AI-assisted, agent-authored, and human contributions. Agent Effectiveness examines coding-agent sessions, while WorkGraph shows team-level patterns in captured work.

Where should we start?

Start with your question: Engineering Performance for delivery outcomes, AI Impact for spend and ROI, Agent Effectiveness for session practices, or Agent Readiness for codebase gaps. Compare the same teams, repositories, and complete weeks.

How should I interpret the product examples?

The product views on this page use illustrative data to explain the metrics and workflows. For metric definitions, sample calculations, and assumptions, see our measurement methodology.

See what your engineering team delivers.

Start with your repositories and coding agents, or get pricing for your organization.

Get Pricing Try Now