research — competitor best-practice survey
Input: a feature or domain question. Output: .ai/research/{feature-slug}.md with
consensus patterns and a recommendation for Carbon. This skill only researches and
writes the findings file — it does not design, plan, or write code.
Announce at start: "Using the research skill — surveying how competitors
handle {feature}."
Step 1: Classify the domain
Match the request against this table. A feature may match multiple rows (e.g.
"shop floor inspections" → MES + Quality). Research every matching row.
| Domain |
Always |
Also research |
Keywords |
| Accounting |
SAP |
NetSuite |
GL, AP, AR, invoicing, payments, accruals, journal entries, period close, financial reports |
| Discrete manufacturing |
SAP |
Epicor |
job shops, work orders, BOMs, routing, job costing |
| MES / production tracking |
SAP |
Manufacturo, First Resonance |
real-time tracking, dispatch, machine monitoring, scheduling |
| CNC parts / sheet metal |
SAP |
Fulcrum, Paperless Parts |
quoting, estimating, nesting, RFQ |
| Quality |
SAP |
1factory, HighQA |
inspection, SPC, FAI, PPAP, GD&T, NCR |
| Inventory |
SAP |
NetSuite, Fishbowl |
valuation, lots, cycle counting, warehouses |
| Sales / CRM |
SAP |
NetSuite, Salesforce |
quotes, orders, pricing, commissions |
| Purchasing |
SAP |
NetSuite, Coupa |
POs, suppliers, receiving, procurement |
| Anything else |
SAP |
discover first |
search best <domain> software and pick the top 2 leaders before deep research |
SAP is always included — it is the reference for enterprise patterns even when
point solutions are more innovative.
Step 2: Write the research questions
Before searching, list 3–6 concrete questions the design will need answered.
Prefer data-model and workflow questions over UI questions:
- What entities and status lifecycles does the competitor use?
- What edge cases do they handle that we might miss?
- Is there an industry-standard term we should adopt instead of inventing one?
Step 3: Search
For each competitor, run searches shaped like:
<Competitor> <feature> documentation
<Competitor> <feature> how it works
<Competitor> <feature> best practices
Rules:
- Be specific:
SAP S/4HANA inventory valuation methods, not SAP inventory.
- Fan searches out to subagents (one competitor or one question per subagent) to
keep the main context clean. Each subagent returns findings + source URLs only.
- Focus on data models, workflows, and terminology — not screenshots or UI copy.
Step 4: Write the findings file
Save to .ai/research/{feature-slug}.md (kebab-case slug). Use exactly this
structure:
# {Feature} Research: Best Practices Survey
## Summary
One paragraph: what was researched, key findings.
## Competitors Surveyed
- **SAP S/4HANA** — {why relevant}
- **{Competitor}** — {why relevant}
## Key Consensus Patterns
### 1. {Pattern name}
- **SAP**: {how SAP does it}
- **{Competitor}**: {how they do it}
- **Rationale**: {why this is the standard}
## Answers to Research Questions
1. {Question} — {answer, citing which competitor}
## Competitor-Specific Details
### {Competitor}
{notable configuration options, terminology, unique approaches}
## Recommended Approach for Carbon
1. {Recommendation with rationale, naming the competitor pattern it follows}
## Sources
- {URL}
Done when
Next step
Hand the findings file to /spec-writing (design + spec) or cite it from an
existing spec. Do not start implementation from research alone.
1---2name: research3description: Survey how best-in-class ERPs and point solutions implement a feature, and produce a findings file at .ai/research/{slug}.md. Use when designing a new feature, evaluating approaches, or answering "how do competitors / SAP / the industry do X". Required before designing any ERP-domain feature (accounting, costing, tax, inventory valuation, RMA, MRP, etc.). Do not use for UI styling questions (copy an existing Carbon screen instead) or for bugs (use root-cause).4---56# research — competitor best-practice survey78Input: a feature or domain question. Output: `.ai/research/{feature-slug}.md` with9consensus patterns and a recommendation for Carbon. This skill only researches and10writes the findings file — it does not design, plan, or write code.1112**Announce at start:** "Using the research skill — surveying how competitors13handle {feature}."1415## Step 1: Classify the domain1617Match the request against this table. A feature may match multiple rows (e.g.18"shop floor inspections" → MES + Quality). Research every matching row.1920| Domain | Always | Also research | Keywords |21|--------|--------|---------------|----------|22| Accounting | SAP | NetSuite | GL, AP, AR, invoicing, payments, accruals, journal entries, period close, financial reports |23| Discrete manufacturing | SAP | Epicor | job shops, work orders, BOMs, routing, job costing |24| MES / production tracking | SAP | Manufacturo, First Resonance | real-time tracking, dispatch, machine monitoring, scheduling |25| CNC parts / sheet metal | SAP | Fulcrum, Paperless Parts | quoting, estimating, nesting, RFQ |26| Quality | SAP | 1factory, HighQA | inspection, SPC, FAI, PPAP, GD&T, NCR |27| Inventory | SAP | NetSuite, Fishbowl | valuation, lots, cycle counting, warehouses |28| Sales / CRM | SAP | NetSuite, Salesforce | quotes, orders, pricing, commissions |29| Purchasing | SAP | NetSuite, Coupa | POs, suppliers, receiving, procurement |30| Anything else | SAP | discover first | search `best <domain> software` and pick the top 2 leaders before deep research |3132SAP is always included — it is the reference for enterprise patterns even when33point solutions are more innovative.3435## Step 2: Write the research questions3637Before searching, list 3–6 concrete questions the design will need answered.38Prefer data-model and workflow questions over UI questions:3940- What entities and status lifecycles does the competitor use?41- What edge cases do they handle that we might miss?42- Is there an industry-standard term we should adopt instead of inventing one?4344## Step 3: Search4546For each competitor, run searches shaped like:4748- `<Competitor> <feature> documentation`49- `<Competitor> <feature> how it works`50- `<Competitor> <feature> best practices`5152Rules:5354- Be specific: `SAP S/4HANA inventory valuation methods`, not `SAP inventory`.55- Fan searches out to subagents (one competitor or one question per subagent) to56 keep the main context clean. Each subagent returns findings + source URLs only.57- Focus on data models, workflows, and terminology — not screenshots or UI copy.5859## Step 4: Write the findings file6061Save to `.ai/research/{feature-slug}.md` (kebab-case slug). Use exactly this62structure:6364```markdown65# {Feature} Research: Best Practices Survey6667## Summary68One paragraph: what was researched, key findings.6970## Competitors Surveyed71- **SAP S/4HANA** — {why relevant}72- **{Competitor}** — {why relevant}7374## Key Consensus Patterns75### 1. {Pattern name}76- **SAP**: {how SAP does it}77- **{Competitor}**: {how they do it}78- **Rationale**: {why this is the standard}7980## Answers to Research Questions811. {Question} — {answer, citing which competitor}8283## Competitor-Specific Details84### {Competitor}85{notable configuration options, terminology, unique approaches}8687## Recommended Approach for Carbon881. {Recommendation with rationale, naming the competitor pattern it follows}8990## Sources91- {URL}92```9394## Done when9596- [ ] Every matching domain row was researched, SAP included97- [ ] Every research question from Step 2 has an answer (or is explicitly marked98 unanswered — carry it into the spec's Open Questions)99- [ ] The findings file exists at `.ai/research/{slug}.md` with all sections filled100- [ ] Every claim has a source URL in the Sources section101102## Next step103104Hand the findings file to `/spec-writing` (design + spec) or cite it from an105existing spec. Do not start implementation from research alone.