How it works
The plain-language guide to what RAD actually does, hour by hour — what runs automatically, what waits for a human, and what a typical day looks like from the approval seat. Every outward action follows the same path:
RAD is two products on one architecture. RAD Business runs everything you do to find, win and keep customers; RAD Professional gives an executive a private chief-of-staff inside their own Google Workspace. Both work the same way: AI generates continuously in the background, and everything outward-facing stops at a human.
The AI drafts an email, a post, a reply or a calendar change, then stages it with the status Pending Approval. A person reviews it and flips the status to Approved — or rejects it. Only then does a separate sender pick it up and deliver it. Read-only work — looking up your agenda, research, answering questions from your documents — needs no approval and happens instantly.
For RAD Business, all pending items live in your RAD dashboard: log in, find the relevant queue (leads, social posts, campaigns, inquiries), and approve or reject. For RAD Professional, the approval happens right in the conversation — the assistant shows you the draft in Telegram or Slack and waits for your go-ahead.
Auto-posting to LinkedIn, X or Reddit would violate their Terms of Service — so RAD never posts to social platforms automatically. Approved social content is staged for a human to post manually.
RAD Business handles the full customer journey — from spotting potential leads online to capturing them, nurturing them with personalised emails, generating content, and syncing everything to your CRM. AI does the heavy lifting; your team approves before anything goes out. Every client gets the combination of features that fits their operation — a marketing team might focus on content and leads; a growth team might add bulk campaigns and CRM sync.
Two automated weekly reports that tell you what competitors are doing and which companies are looking to buy. Neither touches a customer, so neither needs approval.
Every Monday morning, the system visits each competitor's website, compares it to last week's snapshot, and writes a plain-English digest of what changed — readable as a private post in your team blog. No manual monitoring needed.
Every Tuesday, the system scans forums, community sites and the web for people expressing buying intent in your market, scores each signal with AI, and adds the best to a prospect list for your sales team.
Three ways leads enter your pipeline — all landing in the same queue for your team to review.
When a visitor chats with your website chatbot and expresses interest, their details are captured and they get an instant personalised acknowledgment. The lead is queued for your team to follow up after review.
Applicants to your programmes get an instant personalised confirmation, and the application lands in your dashboard. A follow-up email is staged for a team member to approve before sending.
Approval required: follow-up emailEvery Monday morning, new leads are pulled from your prospecting tool, deduplicated against existing contacts, enriched, and given a personalised outreach draft — queued for review before anything is sent.
Approval required: outreach emailEmails to your support address are captured as well. Once an hour the system reads new messages, classifies them, and drafts a reply for a human to review and approve before sending.
Keeps leads warm with personalised follow-ups on a cadence your team sets, and handles large-scale campaign sends.
Every morning, the system identifies leads due a follow-up touch, drafts a personalised email for each, and stages them for review. Nothing is sent until approved.
Checks throughout the day for emails your team has approved, sends them, and marks them as sent. Your team approves; the system delivers.
For large sends to subscriber lists: create a campaign in your dashboard, approve it, and the system dispatches it to your email platform automatically.
A content factory grounded in your actual product catalogue — so nothing it writes drifts from what you really offer.
Refreshes a structured view of your product catalogue from your website every week, keeping all AI-generated content factually accurate about what you actually sell.
On weekday mornings, generates a content pack for a product from your catalogue — a blog post plus per-channel social copy for LinkedIn, X and dev.to — all staged as drafts for review.
Request a blog post from a form on your site or the blog chatbot: provide a brief, and the AI produces a draft for you to review and publish.
When you publish a blog post, it's automatically rewritten as per-channel social copy, sized and styled for each platform. You approve every post in your dashboard before it goes anywhere.
| Channel | What happens after approval |
|---|---|
| Blog | Publishes automatically |
| dev.to | Publishes automatically |
| Staged for a human to post manually (platform ToS) | |
| X / Twitter | Staged for a human to post manually (platform ToS) |
Background plumbing that keeps your records straight — no AI, no approval gates, just data kept in order.
New leads are mirrored into your CRM as contact records throughout the day. Your lead dashboard stays the source of truth, and re-runs never create duplicates.
Watches your calendar for new appointment bookings from your scheduling page. Each booking is recorded and linked back to the originating lead — the full journey from first contact to booked meeting.
The instant someone completes a survey (NPS, CSAT), the response lands in your dashboard with the score and answers recorded. Nothing is ever sent outbound — a pure listening post that keeps feedback in one place. Optional add-on; runs automatically.
A private "ask the company" system — separate from the public chatbot on your website. The chatbot answers customer questions from your product docs; this answers internal questions from your team's own documents.
Every night, documents are fetched from your internal wiki and approved sources into the knowledge base. Re-ingesting a document updates it — no duplicates — and your team can trigger a manual sync any time.
Ask a question, get an answer citing exactly which internal document it came from — retrieval from your own content, not general AI guesswork.
Your platform team handles the technical setup. As the marketing or growth lead, here's the path to being up and running:
Most of the system runs itself. Your team's main interaction is the approval dashboard. Here's what runs without you:
| When | What happens | What you see |
|---|---|---|
| Monday mornings | Competitor websites checked for changes | Digest post in your team blog |
| Monday mornings | New leads fetched, enriched, outreach drafted | New leads in dashboard (Pending) |
| Tuesdays | Web scanned for companies showing buying signals | Prospect signals list updated |
| Every night | Internal knowledge base refreshed | Updated answers in KB queries |
| Every morning | Nurture emails drafted for leads due a touch | New emails in dashboard (Pending) |
| Weekday mornings | Product spotlight content generated | Blog draft + social posts (Pending) |
| Throughout the day | Approved emails sent | Status updates to Sent |
| Throughout the day | Support inbox checked and classified | New inquiries in dashboard (Pending) |
| Throughout the day | New leads synced to CRM | CRM contact records updated |
| Continuously | Calendar watched for new bookings | Bookings logged in dashboard |
RAD Professional is a personal AI assistant for a single executive or business owner — a private chief-of-staff, not a shared team tool. You talk to it through Telegram or Slack. It reads your Gmail, manages your calendar, drafts documents, does research and captures tasks — but it never acts in your name without your explicit approval on each action. Your data stays in your own Google Workspace.
Every action that creates or changes something in your Workspace passes through you first: you ask in the chat, the assistant proposes a draft or an action plan, you approve in-thread, the action happens within minutes, and it's logged in the audit trail.
Sending emails, creating or updating calendar events, creating Google Docs, adding tasks — anything that writes to your Google Workspace. The assistant always shows the proposed action and waits for your explicit go-ahead.
Reading your calendar, summarising emails, web research, querying your knowledge base — any read-only action is answered directly, no confirmation step.
Always on, instant, and conversational — through Telegram private messages or Slack. It understands plain English; there are no commands to learn. Just describe what you need. One prerequisite: your platform team must allowlist your Telegram or Slack user ID before first use, so the assistant responds only to you — until then, all messages are ignored.
Every hour, reads new Gmail messages, classifies them by type and priority, and surfaces the ones needing attention in your chat. Routine items are logged without interrupting you. Reads only — never sends.
For emails that need a response, the assistant drafts a reply and presents it. Approve in the chat and it sends from your Gmail. Or ask "draft a reply to [person]" any time.
"What's on my calendar this week?" gets an immediate, detailed answer — read-only, so no approval step.
Ask it to schedule, propose a time, or reschedule. It drafts the calendar action and waits for your approval before touching your Google Calendar.
Every morning in your timezone: the day's meetings, emails needing attention, open tasks, and relevant news — delivered to your briefing log.
"Prep me for my 3pm with [Name]" — background on attendees, relevant email history, previous notes, and research on the topic, before you walk in.
Ask for an email, memo or proposal. The assistant drafts it and shows you; refine it in the chat if you like. On approval it sends from your Gmail or creates the doc in your Drive and shares it with whoever you specify.
Your Drive and Docs are indexed nightly into a private knowledge base. Ask "what does our partnership agreement say about IP?" and get cited answers from your actual documents.
"Research [topic]" — the assistant searches through your own private search instance, reads the results, and returns a cited report. Queries never go to external search providers.
Mention an action item in conversation and the assistant proposes a task entry. Approve, and it lands in your Google Tasks with the right due date and notes.
You'll need a Google Workspace account (Gmail, Calendar, Drive) and Telegram or Slack. Your platform team handles deployment — as the executive user, here's your part:
Seeing a "Forbidden" error when the assistant sends an email or updates the calendar? Contact your platform team — a Google API may need enabling in your account's project. It's a one-time fix on their side.
| When | What happens | What you see |
|---|---|---|
| Every morning, 7am | Daily brief compiled | Agenda, priority email, tasks and news in your briefing log |
| Nightly | Knowledge base refreshed from your Drive/Docs | Updated KB answers |
| Every hour | Gmail inbox checked and classified | Priority items surfaced in your chat |
| Every 10 minutes | Approved actions executed | Confirmation message in your chat |
"What's on today?" → instant agenda summary. The 7am brief arrives with more context, and you can respond to any briefed item right in the thread.
"Draft a reply to [Name]'s email about the contract." → the assistant shows a draft. "Looks good, send it." → sent from your Gmail within 10 minutes.
"Research [Competitor] — what are they announcing this week?" → it searches, reads the results, and returns a cited summary.
"What does our SLA with [Client] say about response times?" → it searches your indexed Drive documents and returns the relevant clause, cited.
"Schedule a 30-minute catch-up with [Name] next Tuesday afternoon." → it proposes time, title and invitees. Approve → the event is created.
"Prep me for my 2pm with [Company]." → background on the attendees, recent email history, and relevant research, delivered before the meeting.
Whether RAD is nurturing your leads or running an executive's inbox, the line is drawn in the same place: AI drafts at full speed, a human approves, and only then does anything go out under your name — with every decision recorded in an immutable ledger. That's the whole system. The rest is detail you just read.
AI does the work; your people own the decision.
Takes a minute — we'll email you back to confirm a time.