Five product concepts are ranked below against 460 research findings gathered from dozens of independent sources: 300 on missed calls, after-hours demand, estimates and no-shows; 80 on procurement and industrial installed-base revenue; 80 on preferred-language buying and conversational commerce. The evidence mixes primary, academic, vendor and secondary claims — each row below is labeled — and the 1–10 scores are structured judgments assuming a small US software/AI founding team, not measured facts.
x: execution ease for a small team (average of MVP feasibility and founder execution ease, 1–10) · y: opportunity (average of ROI directness, urgency, whitespace and sales-cycle attractiveness, 1–10) · bubble size: defensibility · colour: recommended first build vs the rest. Scores are explicit assumptions, not measurements.
Higher is better on every row, including execution ease (10 = easiest for a small team). Assumptions: small software/AI founding team, selling in the United States, no proprietary distribution. These scores rank relative attractiveness; they are not measured facts.
Customer: HVAC, plumbing, electrical, roofing and restoration firms with high job values. Hidden pool: after-hours and missed calls that defect to whoever answers next, plus unsold estimates sitting in the CRM. Workflow: AI voice/SMS answers every call, qualifies, books into the schedule, revives dormant estimates, and reports verified incremental gross profit against a holdout. Why blue ocean: answering services sell minutes, CRMs sell seats, chatbots sell deflection — nobody sells attributable booked gross profit. Pricing: modest base + fee per qualified booking or share of attributed recovered revenue. 30-day MVP: call forwarding + human-in-the-loop agent for 3–5 design partners; automate only what repeatedly books jobs. Primary risk: attribution disputes — solved by holdout design from day one. Defensibility: proprietary revenue-event data per trade and metro; outcome pricing competitors with seat-based models cannot copy without cannibalizing themselves.
Customer: medspas, dental and specialty clinics, salons in one metro. Hidden pool: predictable cancellations, no-shows and empty chair-hours — median utilization sits ~20–30 points below top performers. Workflow: predict dissolving slots, manage deposits and waitlists, offer freed slots to the best-fit nearby customer within minutes. Why blue ocean: schedulers manage the calendar; nobody runs a demand-and-capacity exchange that monetizes only filled capacity. Pricing: fee on filled slots only. 30-day MVP: manual concierge refilling cancellations for a handful of medspas. Primary risk: thin demand density before the network exists — start where one vertical and one metro are dense. Defensibility: cross-business local demand pool that a single-tenant scheduler cannot assemble.
Customer: small businesses in one vertical — cybersecurity, environmental services or medical supplies. Hidden pool: set-aside goals (23% of federal prime contracts to small business; 5% WOSB; 3% HUBZone) that many eligible firms never pursue because proposals are too costly to produce. Workflow: match only winnable opportunities, auto-build compliance matrices, reuse past evidence, recommend partners, support pricing. Why blue ocean: proposal tools sell writing; this sells wins, charged partly on wins. Pricing: base + success fee. 30-day MVP: concierge bid/no-bid service for ten firms in one NAICS cluster. Primary risk: long procurement cycles delay the win signal. Defensibility: win/loss corpus per vertical compounds.
Customer: clinics, legal intake, education, travel, property services, cross-border sellers facing multilingual demand. Hidden pool: ~1 in 5 US residents speaks a non-English language at home; most consumers prefer buying in their native language; 69% of small exporters have no foreign-language site. Workflow: web, WhatsApp, SMS and voice inquiries converted to completed purchases or booked appointments in the buyer's language. Why blue ocean: translation tools translate words; chatbots deflect tickets; this completes transactions and is paid per completed sale or qualified appointment. Pricing: per completed transaction. 30-day MVP: Spanish-language intake for 3 clinics or law firms in one metro. Primary risk: generic AI-agent platforms adding languages fast — defensibility is the weakest of the five. Defensibility: vertical conversation-to-revenue data and payer/compliance integrations.
Customer: mid-size industrial OEMs and distributors. Hidden pool: installed equipment that pays twice — support ran ~36% of Lam Research revenue over FY2019–26 (secondary source), spare parts can be 40–50% of profits, and aftermarket margins run 2–3x original equipment. Workflow: connect ERP, service, warranty, quote and product data; emit prioritized revenue plays (replacement due, dormant quote, consumable reorder, cross-sell) instead of generic CRM alerts. Why blue ocean: CRMs record activity; this manufactures aftermarket pipeline. Pricing: base + share of play-attributed revenue. 30-day MVP: one distributor's quote and warranty files mined by hand into a ranked play list. Primary risk: the hardest integration lift and the longest enterprise sales cycle of the five — biggest moat, slowest start. Defensibility: the revenue graph itself: entity-resolved installed-base data no competitor can rebuild quickly.
Candid caveat: this water is not empty — answering services, AI receptionists and ServiceTitan itself crowd the space. The blue-ocean move is the unit of sale (verified incremental gross profit, with a holdout) and the pricing, not the phone-answering itself. Choose one trade and one workflow, prove the attribution loop, then expand.
| Idea | Finding | Value | Source type | Source |
|---|
Method: 460 structured research findings (300 on missed calls, after-hours demand, estimates and no-shows; 80 on procurement and industrial installed-base revenue; 80 on preferred-language buying and conversational commerce), collected 2024–2026 across ~60 independent URLs. The table shows 28 representative rows; the rest were cut for space. Scores are 1–10 structured judgments under stated assumptions (small US software/AI team), not measurements; values are quoted as published, and vendor, secondary and illustrative figures are labeled as such.