Operate the
new way

The agent-first platform for Non-Profits

Everything your digital workforce
needs to operate

02
🔧
Extensible Tool System
Equip agents with real-world capabilities: Bing-powered web search, SharePoint file search (RAG), Python code execution, and 13 production MS Graph tools covering Email, Calendar, Teams, and more.
MS Graph MCP Servers
03
👥
Team Orchestration
Coordinate multiple agents as a Team — sequential pipelines, parallel execution, or hierarchical workflows where an orchestrator delegates to specialists. Built for complex, multi-step missions.
Sequential Parallel Hierarchical
04
Multi-Layer Scheduling
Automate recurring tasks at user, agent, and team level. Agents can autonomously schedule their own follow-up tasks during a conversation — cron, interval, or one-time triggers.
Cron Autonomous
05
📓
Agent Notebook
A structured journal for agent observations, learnings, and decisions. Agents log notes, insights, and tasks that inform future conversations — creating a cumulative, queryable knowledge base.
Notes Learnings Insights
06
📨
Agent-to-Agent Messaging
Agents communicate asynchronously without human initiation. Threaded conversations, priority levels, and autonomous responses enable fully autonomous multi-agent collaboration and delegation.
Async Threaded
07
📊
Analytics & Audit
End-to-end visibility: session counts, response times, tool usage, token costs, and complete audit trails. Every action logged for compliance, with correlation IDs for full traceability.
Audit Trail RBAC
The ManAIger Model

Move your people from
[execution] to strategy

CURA doesn't replace your team. It upskills them. Your staff become "ManAIgers" — high-level strategic leads who direct digital agents rather than perform manual tasks. Agents handle the doing; your people handle the deciding.

  1. Define your Collective — map a business unit or department, deployed in 24 hours
  2. Configure your Agents — assign tools, prompts, and memory scopes to each agent
  3. Build your Team workflows — orchestrate agents in sequential or parallel pipelines
  4. Set HILR checkpoints — sensitive actions require ManAIger approval before execution
  5. Watch the platform learn — every interaction enriches the Institutional Brain
// Platform Hierarchy
🏢 Organisation
📂 Collective · Fundraising
👥 Team · Donor Outreach
🤖 Agent · Email Drafter
🤖 Agent · CRM Updater
📂 Collective · Programmes
👥 Team · Impact Reports
🤖 Agent · Data Analyst
🤖 Agent · Report Writer
Workflows in Practice

What your agents
actually do

cura-agent ~ /projects/impact-mission
cura init my-new-project
Initializing CURA environment...
✓ Created new project: my-new-project
Installing dependencies...
_

Automated Grant Writing & Management

Your agents research funders, draft applications aligned to your impact data, track deadlines, and send reminders — freeing your team for relationship-building and strategic decisions.

80% time saved on drafts Zero missed deadlines Consistent voice & data
// Automated Workflow
Funder database searched via web tools
Impact data retrieved from SharePoint
Agent drafts grant narrative against funder criteria
HILR: ManAIger reviews & approves submission
Application emailed via MS Graph
Deadline and outcome logged in Notebook

Donor Relationship Outreach

A dedicated outreach agent monitors donor engagement signals, drafts personalised communications, and coordinates follow-ups — ensuring no relationship falls through the cracks.

Personalised at scale Human review before send Full audit trail
// Automated Workflow
Agent reads recent email interactions
Memory retrieves donor preferences & history
Personalised message drafted
HILR: ManAIger approves outreach
Email sent, calendar event created
Schedule set for follow-up in 7 days

Weather-Triggered Programme Monitoring

Agents monitor environmental data feeds and automatically trigger field alerts, resource reallocation requests, and stakeholder briefings when programme conditions change.

Real-time triggers Automated stakeholder alerts Rapid response
// Automated Workflow
Scheduled agent checks external data APIs
Threshold breached — alert triggered
Field programme data pulled from SharePoint
Situation brief drafted automatically
HILR: Operations lead approves escalation
Teams message sent to field coordinators

Compliance & Policy Monitoring

A compliance agent runs daily checks across documents and communications, logs findings in its Notebook, and routes confirmed issues through HILR for legal team review before any action is taken.

Continuous monitoring Immutable audit log Board-ready reporting
// Automated Workflow
Daily cron trigger fires at 07:00
Agent searches SharePoint for policy documents
Potential issue identified and logged in Notebook
Alert email drafted with findings summary
HILR: Legal team reviews and approves
Stakeholder notification sent

Automated Data & Impact Pipeline

A sequential agent team collects programme data, processes and validates it, then generates visualised impact reports — uploaded to SharePoint and emailed to board members automatically.

Fully autonomous Agent-to-agent handoff SharePoint delivery
// Automated Workflow
Collector Agent fetches data from programme APIs
Agent-to-agent message passes data to Processor
Processor cleans & validates via Code Interpreter
Generator Agent creates impact visualisations
Report uploaded to SharePoint automatically
HILR: ManAIger approves board distribution

The only system that
gets smarter
the more you use it.

CURA's six-scope Memory System means every task completed, every correction made, and every preference learned is permanently encoded in your organisation's intelligence layer — never lost to staff turnover.

🏢 Organisation Company-wide policies, values, and facts
📂 Collective Department processes, contacts, workflows
👥 Team Workflow state, task status, decisions
💻 Agent Agent learnings, patterns, specialisations
👤 User Preferences, communication style, history
💬 Session Temporary conversation context

Scaling impact shouldn't
mean sacrificing safety.

Human-in-Loop Review
(HILR)

Every sensitive action — sending an email, deleting a file, executing a financial transaction — is held in a pending queue until a ManAIger reviews and approves it. You get all the speed of AI, none of the risk of unchecked autonomy.


HILR is configurable at the agent level, the tool level, and the action level. Your board gets what it needs: a complete, immutable audit trail of every decision.

01
Agent Attempts Action
e.g. "Draft and send partnership proposal email"
02
HILR Check Triggered
Tool marked requires_approval — request enters pending queue
03
ManAIger Reviews
Sees full context: email content, agent rationale, conversation history
04
Approve or Reject
With comment — action executes or halts, decision logged permanently
SOC 2
SOC 2 Aligned
Audit logging, RBAC, encryption at rest and in transit
GDPR
GDPR Ready
Tenant isolation, right to deletion, data export
ISO 27001
ISO 27001
Security policies, access control, incident management
Built On

Microsoft-native.
Enterprise-ready.

🔐
Entra ID
Enterprise SSO, multi-tenant, JWT auth with RSA validation
☁️
Azure OpenAI
GPT-4o models, Azure AI Foundry agent sync & management
📧
MS Graph
13 tools: Email, Calendar, Teams, SharePoint — live integration
🔌
MCP Servers
STDIO, HTTP, SSE transports for custom tool extension
📁
SharePoint
Upload, read, search across all SharePoint sites and files
💬
Teams
Send channel messages, create chats, manage conversations
🔎
Bing Search
Bing-grounded web search built into every agent
⚙️
Custom APIs
Connect any REST API with configurable headers, params, schemas