Skip to main content

CURRENT CHAPTER · 03

Building Falcoscan, an AI competitive intelligence platform for SaaS operators.

Solo build. Live in production. 300+ active users indexing 6,700+ AI tools across 29 markets.

THE OPPORTUNITY

Operators discover AI tools through Twitter threads and noisy directories. Comparison data is stale within weeks. Founders ship without knowing what already exists.

Across 27 user interviews, the pattern held: 3 to 5 hours per week, per operator, lost to manual tool research.

The cost is larger than time. It's competitive blindness. Operators learn about a competitor's tooling edge in retrospect, not at decision time.

$14.2B

TAM · AI tools discovery & intelligence, 2026

Grand View Research, 2025

$2.4B

SAM · US mid-market SaaS, 2K to 5K employees

Internal sizing from Built In + LinkedIn data

$48M

SOM · 240 named accounts × $20K ACV

2-year capture target

WHAT THE 27 INTERVIEWS REVEALED

Each card composites 9 interviews in that segment. Quotes are verbatim from research, with names and identifying details changed.

Marcus K., Founder · 18-person B2B SaaS

PERSONA 01 · COMPOSITE

Marcus K.

Founder · 18-person B2B SaaS · Austin, TX

I waste four hours a week scrolling for tools my competitors already ship with.

GOALS

  • Stay current on AI tooling without reading newsletters
  • Spot competitive moves early
  • Cut research from hours to minutes

PAIN POINTS

  • Twitter is noisy, Product Hunt is stale
  • Newsletters arrive after decisions are made
  • No single source for "what shipped this week"

WILLINGNESS TO PAY: $49 to $99/mo

Priya S., Growth lead · 220-person B2B SaaS

PERSONA 02 · COMPOSITE

Priya S.

Growth lead · 220-person B2B SaaS · New York, NY

My team asks me what tools to try. I want one answer I trust, not a 40-tab Chrome session.

GOALS

  • Vet AI tools faster for team adoption
  • Forward category-specific briefs to teammates
  • Track which tools competitors actually use

PAIN POINTS

  • Tool sprawl across departments
  • No comparison data on lesser-known entrants
  • IT and security questions block adoption

WILLINGNESS TO PAY: $199/mo team plan

David L., Analyst · Seed-stage AI fund

PERSONA 03 · COMPOSITE

David L.

Analyst · Seed-stage AI fund · San Francisco, CA

I need to see the AI tooling layer move in real time. Annual reports are a year late.

GOALS

  • Track funding-stage to product-stage signals
  • Identify duplicate ideas in the pipeline
  • Build a private intel layer over public data

PAIN POINTS

  • Crunchbase shows funding, not product
  • Most directories are pay-to-list
  • No structured comparison of model providers

WILLINGNESS TO PAY: $499/mo individual

Research journey map

BEFORE FALCOSCAN

  1. Scroll Twitter

    Sunday evenings, 90 min

  2. Open 6 tabs

    Tool homepages, no comparison

  3. Manual spreadsheet

    Updated once, never again

  4. Forget to update

    Data goes stale by Friday

  5. Ship blind

    Decision made without intel

  6. Learn too late

    Competitor shipped 3 months ago

WITH FALCOSCAN

  1. Open Falcoscan

    One tab, one place

  2. Describe in plain English

    Semantic search v2

  3. Review weekly brief

    Auto-delivered Monday 6AM

  4. Save to watchlist

    Alerts on changes

  5. Ship informed

    Decision made with intel

STRATEGY & VISION

POSITIONING · MOORE FRAMEWORK

For
SaaS founders, growth leads, and AI-curious operators
Who
need to evaluate the live AI tool landscape before shipping
Falcoscan is
an AI competitive intelligence platform
That
surfaces the right tools in plain-English search and a weekly brief
Unlike
Product Hunt directories or G2 reviews
We
index 6,700+ AI tools in real time and write the analysis for you

NORTH STAR METRIC

WBCU

Weekly Briefs Consumed per active User

  1. 01Captures both reach (active users) and depth (briefs read).
  2. 02Tied directly to retention. A user who reads briefs stays.
  3. 03Survives a paid model. Vanity metrics like signups don't.

PRODUCT ROADMAP

Certainty decreases left to right. NOW is committed and dated. LATER is directional.

NOW

4 items

COMMITTED · 0 TO 6 WEEKS

NEXT

5 items

PLANNED · 6 TO 12 WEEKS

LATER

6 items

EXPLORING · 3 TO 6 MONTHS

HIGH CERTAINTYLOW CERTAINTY

PRD · WEEKLY BRIEF

PRDWEEKLY BRIEF
STATUSSHIPPED · MAR 2026
OWNERSHEM NYACHIEO
REVIEWERSSOLO BUILD
LINKEDLINEAR FS-142 · FIGMA · PR #88

Problem

Users sign up and run one or two searches, then leave. Without a recurring trigger that returns them to the product, retention drops below 30% by week four. The weekly brief is the recurring trigger, a personalized digest of tools in the user's saved categories, delivered to inbox and in-app.

User stories

IDStory
US-01As a founder, I want one weekly summary of new AI tools in my categories, so that I stop manually scrolling Twitter on Sundays.
US-02As a growth lead, I want briefs grouped by use case (analytics, content, sales), so that I can forward the right section to the right teammate.
US-03As a VC analyst, I want to mark briefs as read and starred, so that I can build my own intel layer over time.

Functional requirements

IDRequirement
FR-01System must regenerate every user's brief every Monday at 06:00 local time.
FR-02Brief must contain at minimum 5 and at maximum 15 tools per user.
FR-03Tools must be ranked by relevance score (semantic match to the user's saved categories).
FR-04User must be able to mark a tool as "not relevant" and have it suppressed for 90 days.
FR-05Brief must include a one-line analysis per tool, generated from the tool's metadata.

Non-functional requirements

IDRequirement
NFR-01Brief generation per user must complete in under 8 seconds at p95.
NFR-02Brief delivery must achieve 99.5% on-time delivery (Monday 06:00 ± 15 min).
NFR-03One-line analyses must achieve >85% factual accuracy (sampled monthly against tool homepages).

Acceptance criteria

GIVENa user has saved at least one category
WHENMonday 06:00 local time arrives
THENa brief is delivered to their inbox and visible in-app
ANDthe brief contains 5 to 15 tools ranked by relevance
ANDeach tool has a one-line analysis under 200 characters

Success metrics

WBCU TARGET

≥ 0.6

60% of actives read at least one brief

OPEN RATE

≥ 45%

email opens, 7-day window

"NOT RELEVANT" RATE

≤ 8%

per-brief suppression rate

TIME TO FIRST READ

≤ 8 hrs

median, from delivery to first open

SYSTEM DESIGN

INGESTIONPIPELINESERVING200K+ records/week · ~4¢ per recordClaude for analysis · GPT for embeddings6,700+ sourcesRSS, sitemapsPublic APIsCrunchbase, GitHubGitHub releasesReal-time webhooksFirecrawl scraperHTML normalizationApify actorsDynamic JS pagesNormalizerTypeScript serviceDual-model enricherClaude · GPTSupabase Postgrespgvector embeddingsNext.js appApp Router, RSCSupabase authJWT + RLSPostHog analyticsFunnels, retentionVercel edgep95 < 200ms global

TELEMETRY

WEEKLY ACTIVE USERS

312

↑ 18% WoW

WBCU

0.71

↑ above 0.6 target

TOOL INDEX SIZE

6,742

↑ 412 added this week

WEEK-4 RETENTION

54%

→ flat, working on it

Public launchSemantic search v1Ben's Bites featureWeekly Brief shipsPricing experiment liveW0W8W16W24W34

Hover or tab through the line to inspect weekly active users and key launch moments.

GO-TO-MARKET

IDEAL CUSTOMER PROFILE

  • B2B SaaS companies
  • 2,000 to 5,000 employees
  • US-based
  • Already using 3+ AI tools
  • Growth, product, or ops leader as champion

TARGET LIST

240

named accounts sourced from Built In, LinkedIn Sales Navigator, and Crunchbase filtered for ICP fit.

WEDGE

Free Weekly Brief as the front door. Convert to $99/mo Pro after 4 briefs read.

Go-to-market timeline

Q1 · DESIGN PARTNERS

10 partners

  • $0 ARR
  • Validate brief value
  • 5 cycles of in-depth feedback

Q2 to Q3 · PAID CONVERSION

50 paid seats

  • $59K ARR run-rate
  • Prove willingness to pay
  • Pricing settled

Q4 · SCALE

$775K to $1M ARR

  • 150+ paid seats
  • Self-serve motion live
  • First enterprise close

NEXT CHAPTER →

Coca-Cola.

A structural operating-model case study about why Coke's margin problem in North America had nothing to do with the brand.