Your AI Agent Operating System

Build, deploy, and operate AI agents that run on your nonprofit's actual data. Set them up with a guided wizard or a detailed custom build - every agent has access to all platform tools across your grants, finance, compliance, board, and operations. Open standards let them work with Claude, Cursor, and other external AI tools, and even connect to agents at other organizations.

Start Free View Integrations

Agents that think, decide, and act

Each agent runs an ongoing AI conversation with access to your organization's data through permission-controlled platform tools. The AI reads your compliance deadlines, grant opportunities, financial data, and website health - then takes action or surfaces recommendations.

Multi-Model AI

Agents run on leading AI models. Choose the right one for each task - a fast model for routine checks and a powerful model for grant writing. Switch models any time without rewriting anything.

Persistent Memory

Agents remember context across runs. An agent that monitors your grants pipeline remembers which opportunities it already flagged, what your team decided, and what changed since last time.

Human-in-the-Loop

Sensitive actions require human approval before execution. Your team stays in control - agents propose, humans approve. Configurable autonomy levels from observe-only to fully autonomous.

How It Works

From template to running agent in minutes

Pick a recipe, template, or build from scratch

Deploy in 30 seconds with a guided recipe, choose from 26 pre-built templates, or define a fully custom agent with your own instructions and tool access.

Configure triggers and permissions

Set a schedule (daily, weekly, or a custom time), define condition-based triggers, or start agents automatically from your other tools. Choose which areas of your platform data the agent can access - monitoring, compliance, grants, finance, people, assets, or support.

Agent runs and takes action

The AI reads your organization's data - only what you've allowed it to see - reasons about what it finds, and either takes action (creates tickets, saves opportunities, updates credentials) or surfaces recommendations for your team to review.

Monitor health, costs, and impact

A 5-factor health score tracks each agent's reliability. Per-agent cost tracking and spend limits prevent surprise bills. Outcome attribution measures real-world value when recommendations lead to actions, and a weekly digest summarizes your fleet's wins.

Real Impact

What agents do for your organization

Catch every compliance deadline

A Compliance Monitor agent checks your credential vault daily, flags upcoming expirations, and generates renewal preparation documents - so your team acts weeks before a deadline, not days after.

Reads credentials, deadlines, and renewal status
Generates renewal prep docs with AI
Routes urgent findings to your notification feed

Discover funding you would have missed

A Grant Scout agent searches across federal, state, and foundation sources, scores opportunities against your organization's mission and eligibility, and saves the best matches to your pipeline - before the deadline passes.

Searches grants, forecasts, and financing programs
AI-powered matching with explanation scores
Saves top opportunities directly to your pipeline

Orchestrate multi-step workflows

Chain multiple agents together into workflows with dependency graphs and input mapping. One agent researches, another drafts, a third reviews - all coordinated automatically with cost rollup across the entire workflow.

Visual workflow builder with dependency graphs
Step-by-step execution with result passing
Aggregate cost tracking across all workflow steps

Autonomous Development Director

The Development Director agent reviews your entire donor portfolio every week. It flags at-risk donors, identifies major gift prospects, drafts personalized outreach in Content Studio, and schedules follow-up meetings on your calendar - replacing 0.5-1.0 FTE of manual portfolio review. Enterprise only.

Weekly portfolio review across up to 50 donors
Outreach drafts in Content Studio - human approval required
Follow-up events scheduled automatically on your calendar

Platform Capabilities

Everything to manage AI agents

Templates & Recipes

26 pre-built templates plus guided recipes that deploy agents in 30 seconds. Pick a recipe, fill in a name and schedule, and your agent is live.

Cost Control

Per-agent spend limits, per-token cost tracking, and organization-wide usage dashboards. Circuit breakers prevent runaway costs automatically.

Health Monitoring

A 5-factor health score tracks completion rate, error rate, cost, latency, and recommendation quality for every agent. Automated alerts when an agent degrades or recovers.

5-Tier Integrations

Agents connect to any service through pre-built connectors, open protocol servers, ready-made actions, AI-generated connectors, or custom integrations. Credentials are stored encrypted and kept isolated to your organization.

Automatic Template Updates

System agent templates stay current automatically. When platform admins release template improvements, your deployed agents receive upgraded capabilities while preserving your custom name, description, schedule, and instructions.

Auto-Deploy

Feature-embedded agents deploy automatically when your organization upgrades its tier. No manual setup - the right agents appear when you activate the capability.

Audit Trail

Every agent action is logged with full context - what it read, what it wrote, what it recommended. SHA-256 hash-chained audit trail for compliance and accountability.

Full Activity History

See exactly what your agents did - every step, every decision, every result. Understand failures instantly instead of re-running agents by hand.

Self-Healing Agents

Agents recover from failures automatically. Expired credentials trigger reconnect prompts. Rate limits adjust schedules. Transient errors get retried. You fix the problem, not the symptoms.

Model Selection

Choose the right AI model for each agent. A fast model for routine monitoring, a balanced model for everyday analysis, and a powerful model for complex reasoning on any paid plan or with purchased credits. Templates recommend the best model automatically.

Cross-Agent Collaboration

Agents share discoveries through a shared workspace. When your compliance monitor finds an expiring credential, your grant writer can adjust timelines automatically. No complex workflows required.

External Agent Interoperability

AgentHQ is open. Any agent can work here.

Your AgentHQ agents don't live in a silo. External AI agents - Claude, ChatGPT, Cursor, and any compatible tool - operate TappHQ natively through open standards. Every platform tool is available to them.

Direct tool access

An open-standard connection lets Claude, Claude Code, and Cursor work with TappHQ directly. Pro plans unlock read access to all platform tools. Enterprise adds read and write.

Agent-to-agent delegation

External agents can hand tasks to your AgentHQ agents - and vice versa. Your compliance agent can pass research to an external funder-intelligence agent without any custom setup work.

Capabilities discovery

TappHQ publishes a public directory of everything agents can do - the available tools and their access rules. External agents discover what TappHQ can do with no setup required.

Explore the Agentic Platform

How it connects

This HQ is part of a bigger picture

Every HQ shares its data, its proactive alerts, and its agents with the rest of the platform. Here are three concrete ways this one works with the others.

Agent context from every HQ

Ask TappIQ tools resolve data from PeopleHQ, GrantsHQ, FinanceHQ, ComplianceHQ and more so agents answer questions without the user copy-pasting context.

Self-healing reliability

Agent run reliability events (agent.self_heal.action_taken, agent.run.retried) feed back into the scheduler so agents auto-adjust cadence when degradation is detected.

Shared memory across workflows

agent.shared_memory_written persists durable agent memory so downstream runs in other modules (Grants drafting, donor acknowledgments) reuse the same organizational context.

See how everything connects

Guardrails

Agents with brakes, not just a gas pedal.

Autonomy is powerful. Unconstrained autonomy is a liability. AgentHQ layers multiple independent safeguards so agents stay useful without becoming unsafe.

Autonomy levels

Set each agent to Strict, Standard, or Permissive (Permissive requires the Enterprise plan). Dial how much any specific agent is allowed to do on its own.

Approval before any change

Every action that changes something (send email, update record, create document) pauses for your confirmation. Nothing is saved until you approve it.

Untrusted-source handling

Content that arrives from outside - automated data feeds, emails, knowledge-base articles - is treated as untrusted. If an agent tries to act on it, the action is held for your approval, even at high autonomy.

Whole-conversation protection

Once a conversation is flagged as suspicious, that caution carries through the rest of it. Anything the agent saved during a compromised session is kept out of future answers.

Organization isolation

Agents can never reach another organization's data, period. Any attempt to cross that boundary is refused and logged.

Field-level output controls

Each tool can return only a specific, pre-approved set of fields. Anything outside that list is removed before the AI ever sees it - no accidental over-sharing.

Sensitive-data screening

Every generated response is scanned for personal data and secrets before you see it. Untrusted content is marked, and any sign of a data leak triggers a security alert.

Enterprise-grade AI safety

Read the posture: /trust/ai-safety

TAPPIQ Memory

The longer you use it, the smarter it gets

TAPPIQ Memory gives every agent persistent organizational knowledge across 6 dimensions. Preferences, decisions, behavioral patterns, voice profile, and contradiction detection - all applied automatically to everything the AI does.

Stated Preferences

Terminology, communication rules, priorities, and operational constraints persist across every session and surface automatically. Tell it once. It never forgets.

Decision History

Strategic decisions, funding priorities, and policy choices are captured and applied to grant writing, agent runs, and Ask TappIQ - so AI advice aligns with your actual direction.

Inferred Patterns

The platform observes how your team interacts with AI - acceptance rates, edit patterns, timing - and adjusts grant writing to match what actually works for your organization.

6 dimensions: Preferences, Patterns, Relationships, Decisions, Voice, Contradiction & Drift. Manage at /settings/memory. Starter+ only.

AgentHQ

Deploy your first AI agent today

Start with a free account and two agents. Pick a template, set a schedule, and let AI handle the work your team doesn't have time for. Upgrade when you're ready for more.

Start Free Browse Integrations