New: Hoopoe Studio brings every channel into one calm workflow.See what’s new

Product Strategy

AI marketing team, governed like a control panel.

Hoopoe starts with the proven social-management foundation: connected channels, content calendar, publishing, inbox, analytics and approvals. Then it adds a controlled autonomous layer that can plan campaigns, produce content, respond to customers, capture leads, follow up, and improve results.

Comprehensive Plan

Build Metricool's foundation, then give it accountable autonomy.

The product should not become Metricool with extra buttons. It should become a governed operating loop: objective in, campaign out, conversations handled, leads followed up, results measured, and the next campaign improved.

Keep the proven foundation

Workspaces, brands, official channel connections, content calendars, inbox, analytics, approvals, and reports.

Add a controlled autonomous layer

Campaign planning, content production, approved publishing, response handling, lead capture, follow-up, and optimisation.

Make every action governable

Automation levels, approval thresholds, audit logs, policy rules, spend limits, and emergency pause controls.

Tie social work to revenue

Every comment, message, lead, appointment, proposal, sale, and report is connected to the campaign that created it.

Social-management foundation

First, match the central control panel customers already understand.

Hoopoe should use official APIs, authorised accounts, secure tokens, supported inbox permissions, scheduled jobs, and transparent platform limitations. That is the non-negotiable base for safe automation.

Workspace model

Account -> workspace -> brand -> channels, team, campaigns, CRM, automations

Connections

OAuth, account selection, encrypted tokens, permission checks, renewal, webhooks

Publishing

Platform adapters, media validation, queues, retries, reconciliation, manual handoff

Inbox

Permitted comments and messages collected through official APIs and webhooks

Analytics

Followers, reach, impressions, engagement, ad performance, best-time signals, exports

Approvals

Team and client review with dynamic gates created from policy rules

Autonomous Loop

Objective in. Revenue learning out.

Each step produces structured output, records why it acted, and routes higher-risk work through the policy engine before anything reaches a customer.

Observe

Read business profile, channels, audience patterns, CRM movement, and campaign history.

Example business objective

"I need 20 new catering customers in Nairobi during August with a budget of KSh 30,000."

Campaign theme

Back-to-office lunch plans

Approval

Draft review

Lead path

Instagram -> WhatsApp -> CRM

Optimisation

Move proven posts to 9:30

Systems Required

The product architecture is a set of specialised systems, not one giant chatbot.

Social Connection Hub

OAuth and account selection
Encrypted tokens
Webhook registration
Connection health

Brand Brain

Products and services
Voice, offers, prices
Approved and prohibited claims
FAQs and objections

Campaign Director

Audience segments
Campaign theme
Channel strategy
Performance targets

Production Agents

Strategy
Copy
Images
Platform adaptation

Policy Engine

Risk scoring
Approval routing
Autopublish rules
Escalation controls

Publishing Engine

Adapters
Validation
Idempotency
Rate-limit handling

Engagement Agent

Intent classification
Approved replies
CRM creation
Sensitive escalation

Revenue CRM

Contacts
Leads
Appointments
Sales attribution

Optimisation Engine

Best-time changes
Content learning
ROI analysis
Decision logs
Policy engine

The difference between useful autonomy and risky automation.

Hoopoe should decide whether it can act, needs approval, or must hand off to a person. The rules are readable, auditable, and tied to brand knowledge.

SignalConditionAction

Educational tip

No new claim or price

Publish automatically

Previously approved product post

Matches Brand Brain and channel rules

Publish automatically

New discount

Discount above 10%

Owner approval

Price change

Affects offer or quote

Business owner approval

Complaint

Negative sentiment or support risk

Human review

Quotation request

Buying intent detected

Create CRM lead and assign salesperson

Direct Comparison

Hoopoe keeps the control panel and adds the accountable AI team.

Account connections

Customer connects social accounts

Same, plus health monitoring and capability discovery

Content calendar

User plans content

AI can build the calendar from a business objective

Content generation

AI suggests text

AI creates complete campaigns and channel-specific variants

Publishing

Scheduled publishing

Policy-controlled autopublishing with manual handoff

Approval

Team or client approval

Risk-based dynamic approval by content, spend, and intent

Comments and messages

Unified inbox where APIs permit

Classify, reply, escalate, and convert to CRM

Analytics

Social and ad reports

Social, CRM, sales, revenue, CAC, and ROI analytics

Learning cycle

User interprets reports

Hoopoe recommends or executes approved improvements

What "all by itself" cannot mean

Access every personal account or private inbox
Publish every native content format or trending audio
Override social platform moderation or API limits
Continue after permissions are revoked
Avoid human approval for legal, sensitive, or high-risk issues
Guarantee sales, reach, or engagement outcomes

Accurate autonomy promise

Hoopoe autonomously performs every permitted and customer-authorised marketing action, while escalating restricted, sensitive, or unsupported actions for human completion.

Every optimisation records what changed, why it changed, the data used, the previous setting, the new setting, who authorised automation, and the result after the change.
Review checklist

The requested target is represented end to end.

This page now explains the plan, the systems needed, the autonomous operating loop, the compliance limits, the team approval model, and the commercial difference from a traditional social-media manager.

Metricool-style account and brand model
Official API and platform limitation language
AI campaign strategy and production loop
Risk-based dynamic approval examples
Unified inbox to CRM conversion
WhatsApp-first lead follow-up positioning
Revenue and ROI analytics connection
Three automation levels

Build the campaign once. Let Hoopoe run the approved work.

The target product is not just a scheduler. It is a governed marketing operator for businesses that want social activity connected to customers, follow-up, and revenue.