New ideaAgent services
GetNinjas-style copycat
An agent-native quote network for local services in Brazil: consumers' AI agents post a structured job (cleaning, electrician, plumbing, moving), verified local pros reply with quotes on WhatsApp, and the agent returns options for the owner to approve; pros pay per lead, GetNinjas-style.
Working title: GetNinjas-style copycat for AI agents in Brazil (working name)
Summary
BriefingSelf-reportedHypothesis
New / agent-services idea (hunch). Benchmark: GetNinjas (https://www.getninjas.com.br; listed on B3, now ARND3 after a corporate reorganization): Self-reported (company filings via CredCapital and Arandu IR) — 2025 total revenue R$70.9M; Q2 2025 net revenue R$21.5M with 93% gross margin; Self-reported via Folha (31 Dec 2025) — more than 2.6M requests and 5M quotes in 2025.…
Read the rest
Pros pay "coins" to unlock leads; the platform does not intermediate the service payment. Existence check: Not a global gap: agent-native quote tools already exist — Cleartrade (structured supplier quotes for agents), Operator.fyi (MCP quotes and bookings), AllServicePros (request_quote MCP), Haulest (movers MCP), ProCompare (Nigeria, MCP). Agent messaging: HatchLoop, AgentCell. No Brazil or Portuguese-language agent-native services network was found in this run; human incumbents are GetNinjas, Triider and Parafuzo (names only). X: 1,766 posts/7d (Down 39% Wed→Mon (full UTC days: Wed 14 290 → Mon 19 176); peak Wed 14 290; Mon 19 is 39% below peak; window 2026-10-13 10:30 UTC → 2026-10-20 10:00 UTC (Tue 13 Oct and Tue 20 Oct are partial UTC days; full days Wed 14 → Mon 19)). Revenue hypothesis: ~$3K–$8K/month at month 12 (HYPOTHESIS: 3,000 agent requests/month × 3 lead unlocks × ~$0.40–$0.90).
DescriptionVerified
An agent-native quote network for local services in Brazil: consumers' AI agents post a structured job (cleaning, electrician, plumbing, moving), verified local pros reply with quotes on WhatsApp, and the agent returns options for the owner to approve; pros pay per lead, GetNinjas-style.…
Read the rest
Who: Primary users: consumer AI agents (Meta Muse, ChatGPT, Gemini, Hark and agent builders) acting for Brazilian households; supply side: independent pros and small service firms who already work through WhatsApp. Pain: Agents can browse and buy goods but stall on local services: there is no structured way to request, compare and book quotes from pros who only answer on WhatsApp, and human marketplaces are not built for agents. Monetization: Pros pay per unlocked lead (GetNinjas "coins" mechanic) and optional monthly plans; agent builders get free API access with usage caps; labeled sponsored pros in results. No approved price. Build: 3–5 agent-weeks (MCP + REST job schema, WhatsApp Business API relay to pros, pro onboarding and verification flow, lead wallet with Pix, owner-approval callbacks); stack: Next.js on Vercel; MCP server; WhatsApp Business Platform via an official provider; Postgres; Pix provider; Stripe. GTM: Listings in MCP registries and agent app directories; free API for agent builders; pro acquisition through WhatsApp groups and SEO/GEO pages per service and neighbourhood; faceless content only.
Target customer
Primary users: consumer AI agents (Meta Muse, ChatGPT, Gemini, Hark and agent builders) acting for Brazilian households; supply side: independent pros and small service firms who already work through WhatsApp.
Pain
Agents can browse and buy goods but stall on local services: there is no structured way to request, compare and book quotes from pros who only answer on WhatsApp, and human marketplaces are not built for agents.
Why this score
Why it scored lower: Two-sided cold start and uncertain agent demand: consumer agents may not route service requests to third-party networks yet, and WhatsApp Business policies limit automated outreach to pros.
MoatHypothesis
Pro supply density + agent integrations: The moat is the number of responsive pros per neighbourhood and being the default tool in agent directories in Portuguese. Pros return for leads and agents prefer the network with the fastest quotes; both sides compound. Status: Hypothesis.
Demand signals
X posts (7 days)
1766
Signal trend
Down 39% Wed→Mon (full UTC days: Wed 14 290 → Mon 19 176); peak Wed 14 290; Mon 19 is 39% below peak
Web demandVerifiedSelf-reported
Benchmark revenue above (Self-reported). Agent-commerce signal: Hark's free-pizza promo ran 203,500 Handoff browser sessions and delivered nearly 20,000 pizzas, with checkout and login failures (Axios, 20 Oct 2026; Brett Adcock on X) — 372 X posts in 7 days, peaking Sat 17 Oct. Brazilians seldom ask for pros on X (47 posts), so demand lives on WhatsApp and Instagram.…
Read the rest
The first X query is noisy (it also matches estate agents and other "agents"). Other X counts (same window, Verified X API): 47 posts: (diarista OR eletricista OR encanador OR pedreiro OR marido de aluguel OR chaveiro) (orçamento OR indicação OR indica OR "alguém conhece") lang:pt -is:retweet | 372 posts: (hark OR @harkai OR "hark ai") (pizza OR pizzas OR doordash OR handoff) -is:retweet
Market and competitors
TAM
Not sized this run — Scope: Local home services lead generation in Brazil. Method: No market report fetched. (Basis: n/a)
SAMHypothesis
~$36K–$100K / yr (hypothesis) — Scope: Agent-originated requests in São Paulo, two categories. Method: 3,000 requests/month × 3 unlocks × $0.40–$0.90 × 12 (HBV assumption). (Basis: HBV assumption)
SOMHypothesis
~$36K–$96K ARR (hypothesis) — Scope: Month-12 run-rate. Method: as SAM. (Basis: Low (hypothesis))
Competitors
Agent-native: Cleartrade, Operator.fyi, AllServicePros, Haulest, ProCompare; agent messaging: HatchLoop, AgentCell; agent food ordering: Agent Menu, Soldi Now, LOKAL; agent deals: Minty, Haggl, Kindred, McClipFace, AgentPromoCode; Brazil human incumbents: GetNinjas, Triider, Parafuzo.
Business model
Monetization
Pros pay per unlocked lead (GetNinjas "coins" mechanic) and optional monthly plans; agent builders get free API access with usage caps; labeled sponsored pros in results. No approved price.
Revenue hypothesis (monthly)Hypothesis
~$3K–$8K/month at month 12 (HYPOTHESIS: 3,000 agent requests/month × 3 lead unlocks × ~$0.40–$0.90).
Jobs to be done
When my owner asks me to hire a local service pro, I want to post a structured job and get comparable quotes from available, rated pros, so I can return two or three options for approval and book the chosen one.
HXC
The agent of Rafael, 34, São Paulo (illustrative) · Rafael uses Meta Muse and ChatGPT for errands; his agent manages his calendar and shopping. Buying signal: Rafael asks his agent to "find a diarista for Saturday under R$200"; today the agent can only search the web, so a structured quote network with owner approval would finish the job.
Build and go-to-market
Tech stack
Next.js on Vercel; MCP server; WhatsApp Business Platform via an official provider; Postgres; Pix provider; Stripe
GTM channels
Listings in MCP registries and agent app directories; free API for agent builders; pro acquisition through WhatsApp groups and SEO/GEO pages per service and neighbourhood; faceless content only
Risks
Main risk
Two-sided cold start and uncertain agent demand: consumer agents may not route service requests to third-party networks yet, and WhatsApp Business policies limit automated outreach to pros.
All risks
Value: Unvalidated — Agent demand for services in Brazil unproven | Usability: Attention — Pros must answer structured requests on WhatsApp | Feasibility: Attention — WhatsApp Business policy limits | Business Viability: High — Two-sided cold start
Next validation step
Recruit 100 pros in two São Paulo categories (diarista, eletricista) and wire one agent integration; kill if fewer than 50 agent-originated requests in 30 days.
Sources
- credcapital.net
- aranduinvestimentos.com.br (2T25 release)
- www1.folha.uol.com.br
- blog.getninjas.com.br
- axios.com
- getcleartrade.com
- operator.fyi
- allservicepros.net
- procompare.ng
- X API counts (window above)
ideaagent-servicesmcplocal-serviceslead-genwhatsapppixbrazilhunchbenchmark-getninjas
Want the next one the day it is scored?
The score, the sources, the risks and the next step, by email.
Related ideas
X2 Agent Gateway (Hosted)
A hosted pub/sub message hub where every agent session (Claude Code, Codex, Cursor, Grok Bot, HARK and others) registers on start, so one orchestrator session can coordinate work across all of them; unregistered sessions are woken via cloud APIs or an optional local bridge.
63/100
Why it scored lower: Free, funded competitors (Reason Network, Agent Relay YC, BAND) and vendors shipping native cross-vendor messaging; per-platform wake reliability.
Open ProductBots.ai →Watchable BYO-subscription agent cloud devbox
A watchable, SSH-able agent cloud devbox billed on bring-your-own Claude/ChatGPT subscription — wedge vs Aether/Cursor black-box agents.
59/100
Why it scored lower: Aether/Cursor category entry; unclear BYO wedge durability
Open ProductBots.ai →Orbit / Orbitagents
Pre-built AI agents that find leads, do cold email, LinkedIn and Reddit outreach and SEO on a schedule, sharing one memory that outside AI tools reach through MCP.
58/100
Why it scored lower: Crowded agent-GTM category with funded players; revenue concentrated on a few large purchases; X demand Unknown.
Open ProductBots.ai →
Research, not investment advice. Numbers keep the label they were published with.