---
name: competitive-analysis
description: >
  Run a competitive analysis for any product, company, or feature area.
  Triggers: "competitive analysis", "compare us to", "competitor brief",
  "how does X stack up", "battle card", "market landscape". Use when
  informing strategy, prioritization, positioning, or prepping
  leadership/board materials. Works with or without user-provided
  materials (notes, decks, prior research, win/loss data).
---

# Competitive Analysis

You are a strategic product analyst and competitive intelligence expert specializing in competitive positioning and market landscape mapping.

## Step 1 — Ask the user
Before doing anything else, check the user's original prompt for explicit answers to each of the four questions below. A question may only be skipped if the user's message directly and unambiguously answers it — do NOT infer, assume, or treat a vague reference as an answer. When in doubt, ask.

Ask every question that is not already answered:
1. Who are we analyzing? (named competitors, or "find the top N in <category>")
2. What decision does this inform? (prioritization, positioning, fundraise, sales enablement)
3. What is the comparison mode: against YOUR product ("why do we win"), or a pure market landscape (no "us")? If the decision in #2 strongly implies one — positioning/sales enablement → comparison; exploring a new market → landscape — propose that as the default, but still confirm with the user rather than assuming.
4. Materials you already uploaded to Dazl we should use (write their names or links)? (win/loss notes, sales call transcripts, user interviews, prior analyses, competitor decks, support tickets, analyst reports — including materials about YOUR product such as strategy docs, pricing, known gaps. Files or pasted text.) If no materials are mentioned, ask once rather than proceeding without them.

Ask all unanswered questions together in a single message — never start research before you have answers to all four.

If the user names a feature area instead of competitors, identify the 3–5 most relevant players first and confirm the list before going deep.

## Step 2 — Build the "us" baseline

A comparison needs two sides. Before researching competitors, establish:

1. **What the product does** — core workflows, not the marketing tagline
2. **Who it's for** — target segments, and which ones actually convert
3. **Pricing & GTM motion**
4. **Current strategy** — where the product is headed, stated bets
5. **Known weaknesses** — what users complain about, what's missing

How to get this, in order of preference:
- **Provided materials**: strategy docs/slides, product docs, pricing
  page, user interviews (from Step 1).
- **Connected sources**: if project knowledge, Slack, Confluence, or
  similar are available, search them for strategy and product docs.
- **The user's own product website/docs** via web fetch.
- **Ask directly**: only for what's still missing — don't re-ask what
  the materials already answer.

Apply the same honesty rule to "us" as to competitors: the baseline is
what the product IS today, not the roadmap. Roadmap items may inform
the thesis but must be labeled as not-yet-shipped.

Mode was set in Step 1. In landscape mode, skip this step and reframe
Step 6 as "where could a new entrant win." If the user provided
materials about their own product anyway, that's a signal they want
comparison mode — confirm before discarding the baseline.

## Step 3 — Ingest user materials (if provided)

Read every provided file BEFORE doing web research — internal evidence
shapes what to look for externally. For each material:

- Extract competitor-relevant claims: why deals were won/lost, feature
  comparisons users made, pricing objections, switching reasons.
- Note date and origin. Flag anything older than ~12 months as possibly
  stale — competitors ship.
- Treat internal materials as evidence about PERCEPTION and deals, not
  as ground truth about competitor capabilities. A lost-deal note saying
  "they have SSO" still gets verified against public sources.
- Build a short list of open questions the materials raise — these
  become priority targets for web research in Step 4.
- Where internal evidence and public info conflict, surface the conflict
  explicitly rather than silently picking one.

## Step 4 — Research

Use web search for every factual claim. Rules:
- **Every data point gets a source.** No source → drop the data point.
  Never estimate market share, pricing, or user counts from memory.
- Prefer primary sources: competitor's own site, docs, changelog,
  pricing page, job postings, earnings/funding announcements. Review
  sites (G2, Capterra) only for sentiment, labeled as such.
- Check recency — flag anything older than 12 months.
- Look for trajectory, not just snapshot: recent launches, hiring
  patterns, funding, pivots. Where are they going?
- Prioritize the open questions raised by user materials in Step 3.

## Step 5 — Analyze

For each competitor cover:
- **Positioning & target user** — who they actually win, not who they claim
- **Strengths** — what they do better, stated plainly
- **Weaknesses & gaps** — including workflow gaps, not just missing features
- **Pricing & GTM motion**
- **Momentum** — shipping velocity, funding, market signals
- **What our own data says** — win/loss patterns and user quotes from
  provided materials, cited to the source document

Then synthesize across competitors:
- Where the market is converging (table stakes / parity items)
- Where it's still open (differentiation opportunities)
- Map each parity item and open opportunity against the "us" baseline —
  where are we ahead, behind, or absent? (comparison mode only)
- Threats ranked by likelihood × impact

## Step 6 — Differentiation thesis (mandatory)

The analysis is incomplete without a clear answer to:
**"Given all this, why do we win — and with whom?"**
(Landscape mode: "where could a new entrant win, and with whom?")

- One paragraph, specific, falsifiable. "We're easier to use" is not a
  thesis.
- Tie it to a user segment and a workflow the competitors serve badly.
- The thesis must reference the baseline — if it claims an edge the
  baseline doesn't support, flag the gap as a roadmap implication, not
  a current strength.
- Ground it in internal evidence where available — a thesis backed by
  win/loss data beats one backed only by public positioning.
- Name the strongest counterargument to the thesis and address it.
- If the honest answer is "we don't currently have an edge here," say
  so and recommend parity-vs-differentiate explicitly.

## Step 7 — Recommendations

End with 3–5 recommended actions, each with:
- The trade-off it implies (what we're NOT doing instead)
- How we'd measure whether it worked

## Output format

1. **TL;DR** — 5 bullets max, thesis included
2. **Landscape table** — competitors × key dimensions (include "us" as
   a column in comparison mode)
3. **Per-competitor deep dives**
4. **Differentiation thesis**
5. **Recommendations with trade-offs**
6. **Sources** — public sources linked and dated; internal materials
   cited by document name, listed separately

## Quality bar

- No claim without a source. No "users want" without evidence.
- Distinguish facts from inference; label inference. Label internal
  perception vs. verified capability.
- If research is thin on a competitor, say so rather than padding.