Revenly AI

Before the challenge · Private pilot

Revenly CRM

A CRM that reads the email you already send and suggests updates to your pipeline, each one with the message it’s based on. Nothing changes until you approve it.

The problem

Nobody keeps a CRM up to date

Most CRMs expect someone to type up every call and email. Nobody keeps that up, so the pipeline goes stale right when things get busy. Deals also only ever move forward, because marking one as lost feels like admitting defeat.

Deals really happen in email threads and calls. Revenly CRM reads those and proposes the update for you, with the source attached, instead of asking you to fill in a form.

How it works

Three steps

  1. Connect Gmail, read-only

    Gmail and Google Calendar, once. No imports and no fields to fill in.

  2. It looks for deal signals

    Budgets, timelines and decision-makers, in any language or currency.

  3. You approve every change

    Each suggestion shows the exact message behind it. Accept it or ignore it. Nothing moves on its own.

Where it’s at

A private pilot

Revenly CRM is in a private pilot with founder-led B2B teams. It isn’t one of the twelve apps on the board. I built it before the challenge started, and it’s where the rules for every app came from. Delayed uses the same approach for flight claims.

Pilot teams sign in at app.revenly-ai.xyz. If you have questions about the pilot, email hello@revenly-ai.xyz.

See how Delayed explains its answers

History

What shipped

Every Revenly CRM release, newest first.

  1. New

    WhatsApp Business: submitted to Meta for review

    • The WhatsApp integration is officially in Meta's review queue. The same extraction engine that reads your Gmail will read the business chats you choose — early access teams get it the day it clears.
    • Channels preview is live in the app: see exactly what the agent will extract from a WhatsApp thread, a Telegram chat, and a voice note. Waitlist open.
  2. Improved

    V1 hardening: security, mobile, trust

    • Every API endpoint is scoped to your user — verified with multi-account testing. Rate limits, strict security headers, and zero stack traces leaked.
    • Fully responsive on mobile, end to end. Branded error states instead of white screens.
    • New Trust & Security tab: our data commitments in plain language, plus one-click CSV export of your deals, contacts, and activities. Your data is yours.
  3. New

    Agent Inbox + Today's Briefing

    • One queue for everything the agent proposes: drafts pending, deals it auto-created, relationships going cold, duplicate contacts. One tap to approve, one to dismiss.
    • Today's Briefing on the dashboard summarises what changed in the last 24 hours — what the agent found while you were selling.
  4. New

    Warm path search + the relationship graph

    • Type any company and see your direct contacts there — plus two-hop introduction chains through people in your network, each scored Strong / Warm / Cool.
    • The relationship graph visualises your whole network on the Contacts page, built automatically from your real communication.
  5. New

    Demo Mode + the ⌘K command palette

    • Demo Mode: explore the entire product on a fully-populated sample workspace — no Gmail connection needed. Perfect for the skeptical.
    • ⌘K from anywhere: search everything, or type "brief me on…" any contact and get a 30-second pre-call brief.
    • Keyboard shortcuts across the app — press ? to see them all.
  6. NewImproved

    Nine currencies + tone-tuned drafts

    • USD, EUR, GBP, INR, AED, SAR, SGD, IDR, MYR — with each formatted the way its market actually reads it, lakh and crore included.
    • Follow-up drafts now tune to your region and preferred tone, with a one-tap rewriter: warmer, shorter, more direct, formal, or casual.
  7. New

    The self-building pipeline lands

    • Connect Gmail (read-only) and the agent turns your real conversations into structured deals, contacts, and activities — no data entry, ever.
    • Drag-and-drop pipeline, AI deal summaries, and "why is this at risk?" reasoning with the signals it's based on.