WorkflowBack to dashboard

Flight Reservation Modification Flight Cancellation Refund Other Cabin Class Upgrade Downgrade New Flight Booking Booking Conflict Resolution Flight Delay Compensation

Metric scope

Workflow metrics are projected; evidence stays tied to analyzed threads

Projected workflow runs226 / moThis workflow represented 68 of 300 analyzed threads.

Analyzed workflow sample68 threadsFindings, recommendations, and evidence cards are still anchored to the normalized workflow sample.

Projection factor3.3xApplied to this workflow's spend, savings, runs, and token totals. Confidence: medium.

Source pool996 sessionsThe full source-pool population used by the dashboard projection.

derived-flight-cancellation-refund · Based on 68 threads · medium confidence

Flight Cancellation Refund

Derived primarily from user-authored prompts across a 300-thread slice. Full-slice prompt clustering ran on every thread, and Claude consolidated the major workflow types from cluster exemplars because the slice exceeds the non-sampling threshold.

Projected spend / mo$2.49sample $0.75

Projected savings / mo$0.63sample $0.19 · Could cut spend by ~25%

Projected runs / mo226sample 68

Projected total tokens587.4Kavg 2.6K per run

Projected input / output125.7K / 461.7K

This workflow

Token and spend trend

14 hour buckets

Dec 4 4AM

Dec 4 5AM

Dec 4 6AM

Dec 4 7AM

Dec 4 10AM

Dec 4 1PM

Dec 4 11PM

Dec 5 12AM

Dec 17 4PM

Dec 17 5PM

Dec 17 6PM

Dec 17 7PM

Dec 17 8PM

Dec 17 9PM

Input tokensOutput tokens

Model mix

Tokens and spend by model

6 models

Tokens by model

input + output

gemini-3-pro-preview109.6K tokens

325 calls31.2% of total

Kimi-K275.6K tokens

322 calls21.5% of total

Claude Opus 4.573.3K tokens

153 calls20.9% of total

GPT-5.269.3K tokens

224 calls19.7% of total

GPT-5.1-CODEX14.1K tokens

48 calls4% of total

Qwen3-Coder9.6K tokens

44 calls2.7% of total

Spend by model

estimated cost

Claude Opus 4.5$1.53

61.7% of total

GPT-5.2$0.62

24.8% of total

gemini-3-pro-preview$0.13

5.1% of total

Kimi-K2$0.12

5% of total

GPT-5.1-CODEX$0.08

3.2% of total

Qwen3-Coder$0

0.2% of total

Opportunities

2 opportunities for this workflow

$0.63 projected

Tool misuse

Failed benchmark outcomes are still paying the full workflow cost

The imported outcome labels show a high failure rate after the workflow has already spent tokens and tool calls, which points to missing early exits or weak preflight checks.

$0.33projected / month savedsample $0.10/mo

high riskmedium confidence

Recommended changes

Compare passing and failing traces for this workflow and add an early gate before the expensive tool loop starts.
Use the imported outcome label as an evaluation dimension so regressions are ranked by wasted spend, not just by raw failure count.

Evidence (1)

StepImported failing outcome

Imported benchmark outcome ended with failure

Tool misuse

Tool loops are dense enough to need batching or early stopping

tool dominates repeated tool activity, so the workflow is likely doing incremental calls where batching, caching, or tighter stop conditions would reduce churn.

$0.30projected / month savedsample $0.09/mo

medium riskmedium confidence

Recommended changes

Batch or cache repeated tool calls where the inputs overlap across adjacent steps.
Add a per-run tool budget and stop condition so failed runs do not keep exploring after the likely answer is already unreachable.

Prompt composition

Input token breakdown

37.7K tokens

user37.7K · 100%

Tool signals

How this workflow runs

Retries0

How often steps had to re-run.

Delegated subtasks0

Tasks handed off to sub-agents during the workflow.

Documents retrieved0

Total documents pulled in across all tool calls.

Median step latency0 ms

Typical time each step takes to finish.

Stage order

Typical workflow path

3 steps

1
Loop×2
Loop: respond → plan → tool — repeats 2 times.
Latency unavailable762 tok avg
1. 1
  RespondKimi-K2
  Respond step in the workflow.
  Latency unavailable225 tok avg
2. 2
  PlanKimi-K2
  Plan the next steps in the workflow.
  Latency unavailable78 tok avg
3. 3
  Tooltool
  Tool step in the workflow.
  Latency unavailable78 tok avg
2
Loop×97
Loop: plan → tool — repeats 97 times.
Latency unavailable15.1K tok avg
1. 1
  PlanKimi-K2
  Plan the next steps in the workflow.
  Latency unavailable78 tok avg
2. 2
  Tooltool
  Tool step in the workflow.
  Latency unavailable78 tok avg
Verify
Verify step in the workflow.
Latency unavailable132 tok avg

Threads

Pick a thread to see what happened

68 threads

Cost per run$0.06

Monthly runs1

Monthly cost$0.06

Operation path43 named tool/models

1
RespondRespond
claude-opus-4-5claude-opus-4-5
7 tok · $0.00
2
RespondRespond
claude-opus-4-5claude-opus-4-5
52 tok · $0.00
3
PlanPlan
claude-opus-4-5claude-opus-4-5
24 tok · $0.00
4
toolTool
Run tooltool
172 tok
5
PlanPlan
claude-opus-4-5claude-opus-4-5
41 tok · $0.00
6
toolTool
Run tooltool
166 tok
7
toolTool
Run tooltool
166 tok
8
RespondRespond
claude-opus-4-5claude-opus-4-5
218 tok · $0.01
9
PlanPlan
claude-opus-4-5claude-opus-4-5
94 tok · $0.00
10
toolTool
Run tooltool
1.3K tok
11
PlanPlan
claude-opus-4-5claude-opus-4-5
68 tok · $0.00
12
toolTool
Run tooltool
1.3K tok
13
RespondRespond
claude-opus-4-5claude-opus-4-5
357 tok · $0.01
14
PlanPlan
claude-opus-4-5claude-opus-4-5
29 tok · $0.00
15
toolTool
Run tooltool
5 tok
16
PlanPlan
claude-opus-4-5claude-opus-4-5
129 tok · $0.00
17
toolTool
Run tooltool
5 tok
18
PlanPlan
claude-opus-4-5claude-opus-4-5
50 tok · $0.00
19
toolTool
Run tooltool
56 tok
20
PlanPlan
claude-opus-4-5claude-opus-4-5
52 tok · $0.00
21
toolTool
Run tooltool
14 tok
22
PlanPlan
claude-opus-4-5claude-opus-4-5
129 tok · $0.00
23
toolTool
Run tooltool
179 tok
24
PlanPlan
claude-opus-4-5claude-opus-4-5
129 tok · $0.00
25
toolTool
Run tooltool
195 tok
26
PlanPlan
claude-opus-4-5claude-opus-4-5
129 tok · $0.00
27
toolTool
Run tooltool
181 tok
28
PlanPlan
claude-opus-4-5claude-opus-4-5
129 tok · $0.00
29
toolTool
Run tooltool
242 tok
30
PlanPlan
claude-opus-4-5claude-opus-4-5
129 tok · $0.00
31
toolTool
Run tooltool
12 tok
32
RespondRespond
claude-opus-4-5claude-opus-4-5
241 tok · $0.01
33
PlanPlan
claude-opus-4-5claude-opus-4-5
213 tok · $0.01
34
toolTool
Run tooltool
14 tok
35
PlanPlan
claude-opus-4-5claude-opus-4-5
110 tok · $0.00
36
toolTool
Run tooltool
16 tok
37
RespondRespond
claude-opus-4-5claude-opus-4-5
269 tok · $0.01
38
PlanPlan
claude-opus-4-5claude-opus-4-5
103 tok · $0.00
39
toolTool
Run tooltool
217 tok
40
PlanPlan
claude-opus-4-5claude-opus-4-5
119 tok · $0.00
41
toolTool
Run tooltool
5 tok
42
RespondRespond
claude-opus-4-5claude-opus-4-5
55 tok · $0.00
43
dataset evaluationTool
Run dataset_evaluationdataset_evaluation
100 tok
44
Imported benchmark outcomeVerify
Imported benchmark outcome
235 tok

The old plan/tool string was the normalized span order. Rows above use imported operation records; when a tool name is missing, the source only provided the normalized stage and operation label.

Snapshots

full_transcriptSnapshot 1 · imported

Hi! How can I help you today? I need to cancel my upcoming flights for reservations XEHM4B and 59XX6W. I'd be happy to help you cancel those reservations. First, I need to verify …