Licensed AI Agent Partner · Professional Tier

Imbila.ai × 2nth.ai
Your AI practice.
Our infrastructure.

Imbila operates as 2nth.ai's in-house AI advisory practice — building, deploying, and managing custom AI agents for enterprise clients, powered by Cloudflare-native edge AI infrastructure.

Explore the platform → Live demo: Wellness → Give feedback
∞
Agents
∞
Clients
70%
Margin retained
6
Platform modules
Edge
Cloudflare native
Partner Programme

The Licensed Partner tier

How Imbila sits in the 2nth.ai ecosystem — and what that unlocks versus other partner types.

Channel Partner
Builder
e.g. Proximity Green · Self-serve skills
✓Co-branded platform
✓9 pre-built skills
✓Web + WhatsApp channels
—Agent builder
—Client registry
—API access
—Revenue dashboard
—Custom model routing
—Knowledge studio
White-label
Agency
e.g. VibeCrafters · Custom brand deployment
✓Full brand customisation
✓All skills
✓Multi-channel
✓Basic agent config
✓Client sub-accounts
—Agent builder
—Knowledge studio
—Revenue dashboard
—Custom model routing
Licensed AI Partner
Professional
Imbila.ai · Full practice platform
◆Everything in lower tiers
◆Agent Studio — build any agent
◆Knowledge Studio — per-client RAG
◆Delivery Suite — engagement tools
◆Revenue Dashboard — full billing
◆Custom model routing per client
◆Unlimited agents + clients
◆Full REST API access
◆Sub-partner management
Platform Modules

Six-module AI practice platform

Everything Imbila needs to run a professional AI advisory practice — from intake to delivery to billing.

🏠
Practice Hub
The command centre. Active engagements, deployed agents, client portfolio, token ledger, and team activity — all in one view.
DashboardAnalyticsD1
🤖
Agent Studio
Build custom AI agents for any use case. Configure system prompts, tool access, knowledge sources, deployment channels. Version and roll back.
Agent BuilderKV PromptsVersioning
🏢
Client Portal
Each client gets their own AI workspace. Deployed agents, deliverables library, usage stats, and direct feedback — white-labelable per client.
Sub-accountsWhite-labelD1
📚
Knowledge Studio
Upload documents, connect to know.imbila.ai, build per-client RAG indexes on Vectorize. BGE-large embeddings, semantic search, automatic chunking.
VectorizeR2BGE Embed
📋
Advisory Suite
Client intake, scoping workflows, delivery milestones, and AI-generated reports. Track engagements from discovery to signed-off deliverable.
EngagementsAI ReportsR2
💰
Revenue Dashboard
Token consumption per client, engagement billing, margin reporting. Imbila retains 70% gross margin. Full transparency on infra cost vs client revenue.
Billing70/30 SplitPaystack
Platform Preview

What it looks like

Interactive mockups of each module. These become real in Phases 2–6.

Practice Hub
Agent Studio
Client Portal
Knowledge Studio
Advisory Suite
Revenue
imbila.2nth.ai / practice-hub
Practice Overview — April 2026
Active Engagements
12
3 in delivery
Deployed Agents
28
across 9 clients
Tokens This Month
4.2M
↑ 18% vs last month
ClientEngagementAgentsStatusNext milestone
Standard BankCredit Risk AI4LiveQ2 review
GrowthpointFM Intelligence3BuildingUAT Apr 15
Tiger BrandsSupply Chain AI6LivePhase 2 scoping
NedbankCompliance Copilot5TestingGo-live Apr 22
PPC CementOps AI Layer2DiscoveryScoping Apr 8
imbila.2nth.ai / agent-studio
Agents (28)
⚖️
ComplianceAI
Nedbank · Compliance
📊
RiskRadar
Std Bank · Credit
🏗️
FacilityOS
Growthpoint · FM
🔗
SupplyIQ
Tiger · Supply Chain
🏭
OpsLayer
PPC · Operations
Agent name
System prompt (v3 — active)
Deployment targets
Web chat ✓ API ✓ WhatsApp
Config
Modelclaude-sonnet-4-6
Knowledgenedbank-compliance
RAG threshold0.76
Prompt versionv3 (active)
Token budget800K / mo
Toggles
Citation required
Escalation alerts
Audit logging
PII masking
Streaming
imbila.2nth.ai / clients / nedbank
Nedbank
Client since Jan 2026 · Compliance Copilot
Active
Agents deployed
5
Tokens used
1.1M
Open findings
3
AgentDaily queriesSatisfactionStatus
ComplianceAI142
91%		Live
POPIA Advisor87
88%		Live
FSCA Reporter34
76%		Testing
imbila.2nth.ai / knowledge-studio
nedbank-compliance-kb
Nedbank Compliance
847 chunks · 1024d · cosine
FAIS Act · FICA · Basel III · POPIA
growthpoint-fm-kb
Growthpoint FM
312 chunks · 1024d · cosine
OHS Act · SANS standards · Asset register
imbila-know-base
know.imbila.ai (shared)
2,400+ chunks · synced daily
29 domains · AI methodology · SA sector
+
New knowledge base
imbila.2nth.ai / advisory / engagements
EngagementClientStageAgentsValue (ZAR)Progress
Credit Risk AIStandard BankDelivery4R480,000
FM IntelligenceGrowthpointBuilding3R220,000
Supply Chain AITiger BrandsLive6R650,000
Compliance CopilotNedbankTesting5R380,000
Ops AI LayerPPC CementDiscovery0R180,000
imbila.2nth.ai / revenue
Monthly client revenue
R142K
↑ 22% MoM
Imbila margin (70%)
R99.4K
retained this month
2nth.ai infra (30%)
R42.6K
platform + AI inference
ClientTokensBilled (ZAR)Infra costMargin
Standard Bank1.4MR48,000R6,200R41,800
Tiger Brands1.8MR62,000R8,100R53,900
Nedbank1.1MR38,000R4,800R33,200
Growthpoint0.6MR22,000R3,200R18,800
Architecture

Connects to the existing Imbila stack

imbila.2nth.ai acts as the management and delivery layer — bridging your existing infrastructure with 2nth.ai's edge AI platform.

Existing Imbila infrastructure
🔌
api.imbila.ai
Workers API · existing clients
🖥️
platform.imbila.ai
Main client platform
📚
know.imbila.ai
29+ domain knowledge base
🤖
claude / openai / google
Existing model integrations
📝
blog.imbila.ai
Thought leadership · content
◆
imbila
.2nth.ai
⟷
2nth.ai edge infrastructure
🗄️
D1 · imbila-beta
Clients, agents, engagements
🧠
Vectorize · imbila-knowledge
Per-client RAG indexes
🗂️
R2 · docs + deliverables
Documents, reports, assets
⚡
Workers AI · model router
Llama / Claude / GPT4 / Gemini
🔑
KV · sessions + prompts
Auth, prompt versioning
Build Plan

Six-phase delivery

Phase 1 is live. The Imbila team steers the rest via feedback below.

Live
Phase 1
Partner Config + Proposal
  • CF resources created (D1, KV, R2, Vectorize)
  • Partner record + CLAUDE.md
  • Proposal site deployed
  • imbila-platform.pages.dev live
Next
Phase 2
Core Platform
  • D1 schema — clients, engagements, agents
  • Magic link auth
  • Practice Hub dashboard
  • Client registry
Phase 3
Phase 3
Agent Studio
  • Agent builder CRUD
  • KV prompt versioning
  • Model router (Workers AI / Claude / GPT / Gemini)
  • Deployment targets
Phase 4
Phase 4
Knowledge Studio
  • Document upload → R2
  • Auto-chunking + BGE embedding
  • Per-client Vectorize indexes
  • know.imbila.ai sync pipeline
Phase 5
Phase 5
Advisory Suite + Delivery
  • Engagement intake + scoping
  • Milestone tracking
  • AI-generated delivery reports
  • Client-facing portal
Phase 6
Phase 6
Revenue + Integration
  • Token billing per client
  • 70/30 margin dashboard
  • api.imbila.ai bridge
  • Sub-partner management
Imbila Team Feedback

Steer the build.
Your platform.

This is Phase 1. What should Phase 2 prioritise? What's missing? What needs rethinking? Your feedback directly shapes what gets built next.

◆
What's the primary use case?
Enterprise client platform, internal practice tool, or both?
◆
Which module to build first?
Agent Studio, Client Portal, Knowledge Studio — or something else?
◆
Integration priorities
Which existing imbila.ai infrastructure should connect first?