Logical

What do they actually do

Logical is a desktop AI copilot you install on your Mac (macOS 14+; Windows listed as coming soon) and is currently in a private/beta rollout. It watches the windows and files you’re working on to build context and then helps without long prompts, by answering questions, drafting content, suggesting actions, and performing small automations across apps (homepage, context awareness, cross‑app understanding, integrations).

It can proactively monitor your screen to surface timely suggestions (Sentry Mode), automatically capture action items as to‑dos, and let you attach local knowledge bases so answers incorporate your docs. The company says data remains on‑device and nothing is stored on Logical’s servers, with only minimal data sent to third‑party LLMs when a model call is required (Sentry Mode, to‑do identification, knowledge bases, privacy).

Who are their target customer(s)

• Product managers and program leads: They juggle meetings, specs, and cross‑team threads and need fast meeting notes, follow‑up drafts, and reliable capture of action items without switching tools or writing prompts (to‑do identification, context awareness).
• Sales and customer‑facing reps: High volumes of email and Slack make timely follow‑ups hard; they want quick, context‑aware reply drafts, reminders, and light automations like CRM or record updates (integrations, to‑do identification).
• Spreadsheet‑heavy analysts and finance users: Broken formulas and data errors slow them down; they need something that flags issues in real time and suggests fixes without manual debugging (Sentry Mode, integrations).
• Individual makers and small‑team knowledge workers: Working across many apps forces them to re‑explain context to chatbots; they want context transferred automatically across windows and documents (cross‑app understanding, context awareness).
• IT/security/ops and procurement evaluators: They must approve software that requests screen‑recording and accessibility permissions and need assurance that sensitive data remains local or tightly controlled (privacy / permissions).

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

• First 10: Pull early users from the invite/waitlist and YC network, do concierge onboarding (install, permissions walkthrough), and run weekly feedback loops to tune Sentry Mode and to‑do capture, collecting in‑product testimonials (homepage, YC listing, Sentry Mode).
• First 50: Outreach to power users in niche communities (PM, sales, ops, analysts) and short, use‑case demos (meeting notes, email drafts, spreadsheet fixes) to convert to small paid pilots; emphasize on‑device data handling during onboarding to address IT concerns (to‑do identification, integrations, privacy).
• First 100: Scale pilots into team trials and channel partnerships (accelerators, VC portfolio programs, complementary tools). Provide standardized security/deployment materials to speed procurement and enable multi‑seat trials that justify deeper integrations (knowledge bases, integrations, privacy).

What is the rough total addressable market

Top-down context:

The addressable market is desktop knowledge workers on macOS and Windows who want AI assistance embedded in daily workflows; the emergence of OS‑ and suite‑integrated copilots (e.g., Microsoft Copilot) signals a broad, mainstream category (Microsoft Copilot).

Bottom-up calculation:

Near‑term, if Logical reaches 500,000 Mac power users at $25/user/month, that’s ≈$150M ARR. Longer‑term, cross‑platform reach to 5M desktop workers at $20/user/month implies ≈$1.2B ARR.

Assumptions:

• Mac power users and early adopters are the initial reachable segment before Windows GA and enterprise rollouts.
• Willingness to pay is $20–$30 per user per month for a proactive desktop copilot.
• Cross‑platform expansion and enterprise readiness unlock a multi‑million‑user ceiling over several years.

Who are some of their notable competitors

• Rewind: Local‑first desktop “memory” app that records activity for searchable recall. Overlaps on privacy and context, but it centers on searchable recordings rather than cross‑app, proactive suggestions and actions (site, privacy).
• Microsoft Copilot (Windows / 365): OS‑ and suite‑integrated copilot with cross‑app assistance and enterprise controls; tightly coupled to Microsoft identity and apps, unlike a third‑party desktop agent monitoring arbitrary windows (Windows, What is Copilot).
• Raycast: Mac launcher with AI features and automations. Similar in helping power users act quickly, but Raycast is command/extension‑driven rather than a background, screen‑aware copilot (AI features, AI privacy).
• Glean: Enterprise search and assistant that indexes company tools with permissions‑aware search and connectors; focuses on centralized knowledge and IT integrations rather than on‑device screen monitoring and real‑time suggestions (assistant, workplace search).
• Mem: AI notes and “memory” app for individuals/teams that organizes meetings and research; proactive in notes context but does not monitor live desktop windows or perform cross‑app automations like a desktop copilot (site, help guide).