# AI Automation Opportunity Score — eCommerce

Companion to **AI agent ROI for eCommerce: how to decide what to automate first**.

This is a **prioritization worksheet**, not a guaranteed ROI model. Do not treat a high score as a forecast of ticket-cut or GMV. Platform cost silhouette to **model against** (not as savings): support-style AgentCore **~$791/mo at 50K sessions** — [pricing calculator](https://www.factualminds.com/tools/amazon-bedrock-agentcore-pricing-calculator/).

## Score

Each factor is **1–5**. Score = product (higher = automate sooner, still subject to blast-radius veto).

```text
OpportunityScore =
  BusinessVolume × Frequency × ManualEffort × BusinessImpact × AutomationFeasibility
```

| Factor | 1 | 3 | 5 |
| ------ | - | - | - |
| **Business volume** | <20 cases / week | 50–200 / week | 400+ / week |
| **Frequency** | Monthly / seasonal | Daily | Continuous (chat / webhooks) |
| **Manual effort** | <3 min, one screen | ~15 min, two systems | 30+ min, three+ systems |
| **Business impact** | Internal nicety | CX or associate load | Money, inventory, trust |
| **Automation feasibility** | No API; HTML only | Reads exist; writes messy | Named tools + Policy path |

**Blast-radius veto:** if a wrong write is irreversible (payment capture, ATP, live price, account takeover), cap week-one autonomy at Recommend / Draft even if the product score is high. See [`autonomy-spectrum-by-action.md`](autonomy-spectrum-by-action.md).

## Simplified priority table (fill yours)

Volume / Impact / Complexity are a **communication** view for leadership. Complexity is roughly the inverse of feasibility (High complexity → do not start there).

| Use case | Volume | Impact | Complexity | Priority |
| -------- | ------ | ------ | ---------- | -------- |
| WISMO | High | Medium | Low | **Very high** |
| Returns eligibility | High | High | Medium | **Very high** |
| Support policy lookup | High | Medium | Low | High |
| Daily ops brief | Medium | Medium | Low | High |
| Inventory risk brief | Medium | Very high | High | High |
| Purchase orders | Medium | High | High | High |
| Order exceptions | Medium | High | High | Medium |
| Catalog validation | Medium | High | Medium | Medium |
| Cart recovery (no invented codes) | High | High | Medium | Medium |
| Review intelligence | Low | Medium | Low | Low |
| Pricing / discount issuance | Low | Very high | High | Low (HITL forever) |
| Account / PII changes | Low | Very high | High | Do not automate |

## Worked example (demo tags — replace)

A merchant scores **their** helpdesk tags. These weekly counts are a **worksheet illustration**, not a FactualMinds client measurement.

| Use case | V | F | E | I | Feas. | Product | First 1–3? |
| -------- | - | - | - | - | ----- | ------- | ---------- |
| WISMO lookup + delay notice | 5 | 5 | 3 | 3 | 5 | 1125 | **Yes — #1** |
| Returns recommend | 4 | 4 | 4 | 5 | 4 | 1280 | **Yes — #2** (HITL on refund) |
| Daily ops brief | 3 | 5 | 3 | 3 | 5 | 675 | **Yes — #3** (internal) |
| Inventory risk brief | 3 | 4 | 4 | 5 | 3 | 720 | Later (read-only first) |
| PO create | 3 | 3 | 4 | 5 | 2 | 360 | Not week one |
| Auto-refund delivered | 4 | 4 | 2 | 5 | 1 | 160 | **Veto** |

Returns can outscore WISMO on product and still ship **second** if refund writes are not ready. Feasibility 1 + veto beats a pretty table.

## How to pick the first 1–3

1. Pull **your** volume (helpdesk tags, OMS exception queue, PIM tickets). Do not use this sheet’s demo counts as yours.
2. Strike rows a deterministic Flow / webhook already closes.
3. Apply the blast-radius veto (payment, ATP, live price, account).
4. Prefer **read-heavy** high scores (WISMO, policy lookup, ops brief) before write-heavy high scores (PO, inventory adjust).
5. Model sessions on the [AgentCore pricing calculator](https://www.factualminds.com/tools/amazon-bedrock-agentcore-pricing-calculator/). If you cannot name a workload that would even **notice** a **~$791/mo** platform floor at 50K sessions, you are buying a demo, not an operation.
6. One owner for evals + cost alarm. Three automations max this quarter unless the first has goldens.

## Related

- Fifteen-area map: [`15-automations-matrix.md`](15-automations-matrix.md)
- Autonomy: [`autonomy-spectrum-by-action.md`](autonomy-spectrum-by-action.md)
- Monday gate: [`monday-checklist.md`](monday-checklist.md)