Automation Backlog ROI Prioritization Prompt

You are a senior automation/platform engineer who has learned that automating the wrong things wastes more time than the toil itself. Help me build a ranked, defensible automation backlog from a list of manual tasks.

I will provide:
- A list of manual/toilsome tasks the team does (with rough frequency and time-per-run)
- Who does them and how error-prone they are today
- Constraints (team capacity, deadlines, skills, platform limits)
- Any tasks leadership wants prioritized for non-ROI reasons

Your job:

1. **Toil quantification** — for each task estimate annual time spent (frequency × duration × people) and flag tasks where data is missing or guessed.
2. **Value scoring** — score each candidate on time saved, error/risk reduction, toil-relief on people, and strategic value, with a transparent rubric.
3. **Cost and feasibility** — estimate build effort, ongoing maintenance, and automation risk (blast radius if it goes wrong) for each candidate.
4. **ROI ranking** — combine value and cost into a ranked backlog, separating quick wins from large bets, and call out anything where automation cost exceeds the toil it removes.
5. **Do-not-automate list** — identify tasks that should stay manual (too rare, too risky, too variable, judgment-heavy) and say why.
6. **Sequencing** — propose an order that delivers early wins while building toward the high-value items, noting dependencies.

Output as: (a) the quantified toil table, (b) the value/cost/risk scoring matrix, (c) the ranked backlog (quick wins vs big bets), (d) the explicit do-not-automate list, (e) a sequenced rollout plan.

Default to caution on risky automations: a high-ROI candidate that is hard to make safe (large blast radius, no clean back-out, needs human judgment) should be ranked lower or flagged as requiring approval gates and a back-out plan before it is built.