Caddy

What do they actually do

Caddy is a desktop voice assistant for work that is currently in early access. The public site routes people to a waitlist, and YC’s profile directs interested users to request access. The product aims to let you speak a command and have Caddy carry out actions across your existing apps, and to use dictation to replace typing where you enter text today (site, YC profile).

In demos, you run Caddy on your computer, speak naturally, and it tries to understand your screen and context to perform multi‑step tasks like creating a Linear issue and sharing it in Slack, scheduling meetings, or sending messages—so you avoid switching tabs and copy/paste. The team frames this as “Siri for work” with action and dictation modes and is recruiting early users via the waitlist and founder outreach (YC profile, Forbes batch coverage).

Who are their target customer(s)

• Startup founders / solo operators: They bounce between Slack, issue trackers, email, and calendars, wasting time copying context between apps. They want a faster way to handle common cross‑app tasks without window switching and manual copy/paste.
• Product managers and engineers: They frequently create and link tickets and post updates across tools after meetings or code reviews. They want fewer manual steps when turning what they just discussed or saw on screen into tasks and status messages.
• Customer support and operations reps: They pull details from emails, update CRM/ticketing systems, and send follow‑ups across multiple tools for each case. They need quicker, more reliable ways to populate forms and send templated responses with less typing.
• Sales and account executives: They draft outreach and follow‑ups, update CRM records, and schedule demos while juggling call notes. They want an easier way to turn spoken notes into calendar invites, emails, and CRM entries without losing context.
• Productivity power users / voice‑first adopters: They prefer speaking over typing for speed or accessibility, but current dictation and automation are inconsistent across apps. They want a single, dependable interface that turns natural speech into accurate text and safe cross‑app actions.

How would they acquire their first 10, 50, and 100 customers

• First 10: Founder‑led concierge pilots with users from the waitlist and YC/founder networks. Hands‑on onboarding where a founder observes real tasks, configures integrations, and ships quick fixes; capture short testimonials and recordings for learning.
• First 50: Invite small cohorts of PMs, engineers, support, and sales at startups via communities/LinkedIn/cold outreach. Run short, free pilots with role‑specific workflow templates in exchange for structured feedback, usage data, and permission to publish anonymized case notes or quotes.
• First 100: Ship a self‑serve installer, vetted workflow templates, and guided setup while maintaining a light concierge playbook for teams that want help. Add a referral benefit and target look‑alike teams (same app stacks) with short demo clips and customer stories.

What is the rough total addressable market

Top-down context:

Caddy sits at the intersection of speech/voice recognition and cross‑app automation. Recent estimates put speech/voice recognition at roughly $8.5–12.6B in 2024, growing rapidly through 2030–2032, while RPA/automation software ranges from ~$3.2B software revenue in 2023 to ~$18.2B broader market value in 2024, both growing double‑digits (MarketsandMarkets, Gartner, Fortune Business Insights).

Bottom-up calculation:

If priced at $25–$40 per user per month and focused first on tech/startup roles (PM, Eng, Support, Sales, Ops), a practical near‑term TAM could be 300k–600k seats among early adopters globally, implying ~$90M–$288M in annual spend. Longer‑term, if a voice‑first assistant reaches a few million knowledge‑worker seats, TAM expands into the low single‑digit billions at similar pricing.

Assumptions:

• Initial targetable users are concentrated in teams already using tools like Slack, Gmail, Linear/Jira, and modern CRMs.
• Willingness to pay aligns with productivity tooling at $25–$40 per seat per month.
• Adoption starts with power users and specific roles, then broadens as reliability and integrations mature.

Who are some of their notable competitors

• Microsoft 365 Copilot / Windows Copilot: Assistants built into Office apps and Windows that draft/edit content and answer questions from organizational data. Strong inside the Microsoft stack; less focused on a universal desktop voice layer across third‑party apps.
• Dragon NaturallySpeaking (Nuance): Long‑standing desktop dictation with high‑accuracy speech‑to‑text and voice commands for editing/app control. Competes for users who primarily need reliable dictation and in‑app voice commands.
• Otter.ai: Meeting AI focused on transcription, summaries, and action items that can sync to other tools. Overlaps when turning spoken content into tasks, but centered on meetings rather than live desktop command execution.
• Zapier: Automation platform connecting thousands of apps for reliable, repeatable workflows. Functional competitor for cross‑app task automation, though not voice‑first or ad‑hoc from on‑screen context.
• Raycast: Keyboard‑centric macOS launcher with extensions and automations for power users to run commands across apps. Competes for the same productivity audience via fast keyboard commands and scripts rather than voice.