Proxis

What do they actually do

Proxis runs AI email agents that connect to company inboxes and company context (calendar, docs, CRM). The agents read inbound mail, generate replies, and can either draft for human review or auto‑send within guardrails the customer configures homepage/pricing. It’s sold as seat‑based SaaS with plans that unlock features like multi‑inbox support, custom rules, and auto‑send.

A typical setup is: sign up, connect Google/Outlook inboxes plus sources like calendar, Notion, and Google Drive (and, on Enterprise, CRMs like Salesforce/HubSpot), then configure reply guardrails and templates. The system ingests email history and context to draft replies; teams can keep it in draft‑only mode or enable auto‑send with limits and monitoring, including enterprise controls and SLOs on higher tiers pricing/features.

They emphasize safety and privacy: customers can set explicit reply rules/guardrails, and their policy states data pulled from Google Workspace APIs is not used to train general models homepage, privacy.

Who are their target customer(s)

• Inbound enterprise sales reps (AEs/SDRs): High volumes of inbound emails and repetitive follow‑ups slow response times and cost opportunities; they want fast, consistent replies without retyping similar messages YC page, homepage.
• Sales operations / revenue‑ops leaders: Need consistent messaging, CRM syncing, and scalable workflows; manual processes and fragmented tools create errors and overhead. They look for CRM integrations, policy controls, and auditability pricing/features.
• Account managers / customer success teams: Have to pull context from calendars, docs, and CRM to answer complex threads; slow context gathering or mistakes hurt retention and upsells. They need agents that surface the right context and draft accurate replies pricing/features.
• IT, security, and compliance leaders at larger companies: Require data controls, audit trails, and SLAs before allowing auto‑send; worry about customer data being used to train external models. They need clear guardrails and privacy assurances pricing/features, privacy.
• Small sales teams and startups scaling inbound: Limited headcount and budget; want low‑cost drafting and optional human approval before anything is sent. Seat‑based plans and draft‑first workflows help them scale without hiring pricing.

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

• First 10: Run high‑touch pilots via YC/founder warm intros to revenue or support teams with heavy inbound email; offer short, low‑risk pilots with white‑glove setup and clear privacy docs to ease procurement YC, privacy.
• First 50: Turn pilot outcomes into short case studies and referrals; target similar verticals via outbound and add a reseller/consultancy channel with CRM integrators (Salesforce/HubSpot), while opening self‑serve seat plans for small teams pricing/features.
• First 100: Segment GTM: productize onboarding/admin/audit logs for IT signoff, staff AEs for enterprise deals with SLAs, and drive leads via partners plus targeted LinkedIn/content highlighting cost/latency benefits from model‑serving work pricing/features, distillation guide.

What is the rough total addressable market

Top-down context:

AI for sales & marketing is forecast around $58B in 2025; Proxis’s niche (email agents for sales/revops) is a subset of that. Email‑marketing/automation software itself is around ~$10B near‑term, suggesting a multi‑billion umbrella with Proxis addressing a focused slice MarketsandMarkets, TBRC email market.

Bottom-up calculation:

Using Proxis’s listed seat pricing (e.g., Pro at $33/month = $396/year) and BLS’s ~13.38M U.S. “sales and related” roles, assume 10–20% are enterprise B2B sales/AM/CS seats (= ~1.34–2.68M seats). At $396/year, full penetration implies ~$0.53–$1.06B in U.S. ARR; 1–5% capture yields roughly ~$5–$53M ARR pricing, BLS.

Assumptions:

• 10–20% of U.S. sales roles are B2B/inbox‑heavy targets (excludes retail).
• ARPU approximates the Pro plan at $33/month; enterprise deals may differ.
• Estimate is U.S.‑only; global expansion would scale figures.

Who are some of their notable competitors

• Front: Shared‑inbox and support platform that drafts replies, routes, and automates across channels; strong for support/omnichannel ticketing vs. Proxis’s sales‑inbox agent focus Front.
• Outreach: Sales‑engagement platform with sequences, analytics, and AI assists centered on outbound cadences and CRM workflows rather than autonomous multi‑inbox inbound reply agents Outreach.
• Lavender: AI email coach that scores and helps write outbound emails; focused on coaching/improving human drafts, not autonomous agents that auto‑send under enterprise guardrails Lavender.
• Gmelius: Gmail‑native shared inbox and workflow tool with AI drafting and automations; strong Gmail integration, less positioned as an enterprise agent platform or model‑serving provider Gmelius.
• Zendesk: Enterprise support platform with AI agents/Copilot for suggested replies, triage, and automation; oriented to ticketed support operations vs. sales‑inbox automation and custom LLM serving Zendesk.