Notsy - AI Agent that takes and makes phone calls, and more.

What the company appears to be: Notsy is an AI agent platform that enables business owners to delegate administrative tasks via WhatsApp chat and voice calls. The platform integrates with Google Workspace and supports automation of tasks like scheduling, email sending, document lookup, invoicing, and customer service. It also includes Instagram and Facebook integration for community management.

What changed: During Build Week (July 15–16), the team focused on improving voice call performance by reducing initial latency from ~15s to ~3s, implementing A/B testing between voice engines, unifying the voice runtime, and launching a new affordable voice plan for small businesses. These changes were driven by using Codex + GPT-5.6 Sol to diagnose and fix technical issues.

Single most important open question: Is there evidence of traction or revenue generation beyond the developer-built prototype? The description states no such data exists.

The description states that Notsy is an AI assistant platform for business owners that allows users to perform administrative tasks through WhatsApp chat and voice calls. It connects with Google Workspace and supports actions like creating PDFs, sending emails, scheduling meetings, making calls, looking up documents, and sending invoices.

The product also includes Instagram and Facebook integrations for community management, including analyzing metrics, suggesting content ideas, and performing paid ad campaigns based on organic post performance.

Notably, the platform has a voice tool that enables businesses to delegate voice tasks such as reservations and customer service via phone calls. The system is described as being able to read and write in Google Workspace documents and respond to incoming leads through WhatsApp.

Evidence: Self-reported by author; no independent verification provided.

The description states that Notsy was inspired by the problem of entrepreneurs spending too much time on administrative tasks, with 60% of phone calls never answered. The platform aims to solve this by automating various business functions.

Initially focused on WhatsApp chat automation for administrative tasks, the product evolved during Build Week to emphasize voice capabilities for tasks like reservations and customer service. It also expanded into social media management for Instagram and Facebook, including content research and ad campaign suggestions.

The author claims that Notsy helps entrepreneurs who lack experience in critical business areas by providing intelligent agents to handle important tasks.

Evidence: Self-reported; no external validation or data on positioning evolution.

The description states that Notsy targets small and medium-sized businesses (SMBs), particularly Mexican SMBs, with a focus on entrepreneurs who lack experience in critical business areas. It is positioned as helping business owners perform tasks they might otherwise struggle with due to lack of knowledge or time.

It also mentions a specific use case for "Mexican SMBs" and includes pricing in MXN ($499/60 min), suggesting localization toward this market segment.

Evidence: Self-reported; no data on actual customer base or segmentation beyond stated intent.

The description states that Notsy offers a voice plan called "Esencial" priced at MXN $499 per 60 minutes. This plan includes usage alerts at 70/90/100% thresholds (dashboard + WhatsApp with email fallback), plan upgrades within the same Stripe subscription with proportional credit for unused minutes, and idempotent server-side endpoints for billing.

The author claims this pricing model was designed to be affordable for small businesses and implemented end-to-end across multiple repositories and surfaces.

Evidence: Self-reported; no independent verification of pricing structure or revenue generation.

During Build Week, the team made several technical improvements:

• Reduced call latency from 14.9s to ~3s by eliminating dead air at start of calls.
• Implemented A/B testing between voice engines (Deepgram + OpenAI + ElevenLabs vs Grok).
• Unified voice runtime with sticky engine per project/call and atomic idempotent minute metering.
• Added async voicemail detection for instant human connection.
• Shipped a new affordable voice plan with Stripe integration.

The author notes that these improvements were achieved using Codex + GPT-5.6 Sol, which helped diagnose and resolve complex technical issues such as telephony latency and billing correctness.

Evidence: Self-reported; no independent verification of delivery or technical performance.

Not evidenced.

The description does not provide any data on revenue, customer adoption, user growth, or product maturity beyond the developer-built prototype. The author explicitly states that no traction or revenue data is available beyond what they have described.

Evidence: Absence of evidence; no data provided.

Not evidenced.

The description does not mention competitors or competitive positioning. It focuses solely on internal development and improvements rather than market analysis or differentiation from existing solutions.

Evidence: Absence of evidence; no competitive context provided.

1. No traction or revenue data: The product is described as a prototype built during a hackathon, with no evidence of actual customers or monetization.
2. Unverified claims: All statements are self-reported and unverified; there is no third-party validation of functionality or performance.
3. Limited team size: Only one team member (HEUSAI Goch) is mentioned, raising questions about scalability and execution capability.
4. Dependency on AI tools: Heavy reliance on Codex + GPT-5.6 Sol for development raises concerns about long-term sustainability if these tools change or become unavailable.
5. Market localization: Focus on Mexican SMBs may limit broader market appeal unless there's a clear path to expansion.

Evidence: Self-reported; no external validation of risk factors.

1. What is the current stage of product development beyond the hackathon prototype?
2. Are there any early adopters or pilot customers using the platform?
3. How does the team plan to scale beyond a single developer?
4. What are the key assumptions behind the pricing model and target market?
5. Can you provide evidence of voice call performance improvements beyond the Build Week fixes?
6. Is there a roadmap for expanding beyond WhatsApp, Google Workspace, Instagram, and Facebook?
7. How is the team addressing potential regulatory or compliance issues in handling business communications?

Inference: These questions are based on the lack of traction, unclear scalability, and unverified claims in the description.