82Healthy
- Tool adoption Strong
- Policy findings Reviewed
- Machine coverage On track
- Decisions reviewed Current
Tracelet / Leadership
Understand your engineering AI rollout.
Review tool adoption, policy findings and machine coverage in one conversation with engineering and security. Then see what the spend bought, and how much of it rests on a single vendor.
TRACELET.AILeadershipIllustrative exampleLeadership / overview
Last 30 daysThe engineering AI rollout, in one view.
$186Knet value, illustrative
1,420 hours saved against a $41K token spend this period.
Policy decisions Recorded Open findings Review Machine coverage Tracked
Teams Projects People
Unregistered MCP servers to reviewTeams with machines not yet enrolledBlocked actions awaiting review
Illustrative example: adoption, findings, coverage and fleet-wide value are available signals today.
Three questions for the next rollout conversation.
Every answer comes from enrolled machines rather than a survey: adoption from recorded AI activity, findings from policy decisions, coverage from what is actually reporting. Values are illustrative.
Benchmarks are agreed with each team, so the tick on every row is your bar, not ours. The rollout stage underneath places the organisation on the adoption guide, so the conversation starts from where you actually are.
01
Which teams are using AI coding tools?
Tool adoption
62%
Enrolled machines with recorded AI activity · illustrative02
What did policy find?
Policy findings
91
Findings evaluated, severity-weighted, 28 days · illustrative03
How much of the fleet is covered?
Machine coverage
72%
Machines enrolled and reporting · illustrativeThe fourth question: who does the rollout depend on?
Model and vendor strategy reads your own AI-provider spend and shows how much of it sits with one vendor, what a price change there would cost each month, and how much already has a same-tier alternative. Developer AI-tool spend gets its own cost-by-model table, kept apart from the provider bill so the two never blur.
Concentration is not a failure; it is a negotiating position. The leverage numbers exist so the next vendor conversation starts with figures rather than a feeling.
38of 100
Single-vendor
- Top vendor
- 71% of spend
- 10% price move there
- $2.9K a month
- Frontier models Primary vendor$29.1K71%
- Frontier models Second vendor$6.6K16%
- Open-weight models Self-hosted$3.3K8%
- Embeddings and small models Two vendors$2.0K5%
Leverage
The three numbers to take into the next vendor conversation.
$6.40per M tokens
Weighted across every provider and tier in use.
$7.2Ka month
If routine workloads moved to the standard tier.
58%of spend
Has an active same-tier alternative today.
Illustrative example: provider spend, concentration and leverage over the same 30-day period as the overview above.
How far along is the rollout, really?
Benchmarks and maturity place the organisation on the same five stages as the scorecard above, with the composite score behind the stage and the four pillars that feed it. The gap to the next stage is named, so the conversation is about the one thing to lift rather than the whole curve.
Reference bands are configurable placeholders agreed with your teams, not measured peer data. Values are illustrative.
TRACELET.AIBenchmarks & maturityIllustrative exampleLast 30 days52of 100 · Scaling
8 points to Optimising. Security posture is the drag: 42 against a 70 bar.
- Tool adoption Target 75%62%13 pts under
- Code keep rate Target 70%58%12 pts under
- Security score Target 704228 pts under
- Spend per active developer Band up to $250$220Inside band
Health pillars
The four scores behind the composite, each against its own bar.
62of 100 · target 75
4 points up on the prior 30 days
58of 100 · target 70
2 points up on the prior 30 days
46of 100 · target 60
3 points down on the prior 30 days
42of 100 · target 70
5 points up on the prior 30 days
Illustrative example: composite maturity, core metrics against their targets and the health pillars behind them.
What is the fleet exposed to?
Risk and compliance rolls policy findings, violations, secrets and destinations into one fleet score, with the enforcement mode beside it so a low number reads as exposure recorded rather than damage done. Break-glass activity is the same keycards engineering approves on the governance side, rolled up for the board.
Every count comes from policy decisions on enrolled machines, over the same 30 days as the rest of the page. Values are illustrative.
42of 100
At risk
- Enforcement
- allow_all · recorded, not blocked
- Weighting
- Severity and age of open findings
4
of 34 critical + high findings
23
29 medium alongside
6
across the 30-day window
9
of 138 distinct domains contacted
+5over the last 30 days · 28 below the 70 bar
- Critical11
- High23
- Medium29
- Low28
- Open
- 31
- Resolved
- 60
- Total
- 91
RuleSeverityTotalOpen
Unapproved destinationMedium218
Sensitive path accessHigh146
Blocked tool callHigh93
Secrets exposureCritical62
Dangerous commandCritical52
- Requests
- 18
- Granted
- 72%
- Pending
- 2
Time-boxed elevated access. The same keycards engineering approves on the governance side, rolled up for the board.
Illustrative example: fleet score, findings, violated rules and break-glass activity in one board-grade view.
Where are spend and return heading?
Strategic forecast projects the cycle from today's run-rate: spend, hours saved and net return, with headcount and unit-price levers to test the plan before the budget round or the next vendor conversation.
Run-rate projections are for planning, not commitment. Values are illustrative.
TRACELET.AIStrategic forecastIllustrative exampleDay 18 of 30$44.2K
$1.47K a day run-rate
$26.5K
12 days left in the cycle
$198K
1,540 hours saved, projected
186 / 300
62% adoption
Actual, days 1–18 Projected, days 19–30 Today
Headcount
−20%Flat+20%+50%
Unit price
−10%Flat+25%
- Scenario monthly spend
- $44.2K Matches baseline
- Scenario net ROI
- $198K Matches baseline
Assumes a linear run-rate for the rest of the cycle at today's blended rate. A planning scenario, not a commitment.
Illustrative example: run-rate, projected spend and return, and a scenario planner over the current billing cycle.
Design partners