Reuse opportunity by workload type
Which agent workloads actually have reusable structure?
Not every workload benefits. This suite breaks down opportunity ratio and median Workload Reuse Score by agent category, so teams can self-assess before instrumenting anything.
What the corpus shows.
| Workload | Opportunity ratio | Median WRS | Typical action | Note |
|---|---|---|---|---|
| Coding agents | 0.74 | 78 | Tune ordering, then route | Repo instructions + tool registries recur heavily. |
| Support automation | 0.61 | 58 | Provider tuning + BYOK review | - |
| Research agents | 0.49 | 47 | Improve prompt construction | - |
| RAG platforms | 0.42 | 41 | Pack documents before infra | - |
| Consumer chat | 0.18 | 16 | Do not optimize for reuse | - |
Takeaways
- Coding agents are the strongest reuse case by a wide margin.
- A high opportunity ratio with low WRS usually means weak retention locality.
- Consumer chat rarely justifies any of this - opportunity is too low.
Methodology
Workloads are classified by traffic shape (system/tool/context/turn composition). We report the median opportunity ratio and median WRS per class.
How we grade evidenceGet these numbers for your traffic.
A diagnostic runs this analysis on your own workload and attaches an evidence level to every figure.