Organizational knowledge
Keep how the work gets done.
- Approved program facts and funder requirements
- Your writing and working practices
- Named owners and access rules
Nonprofit AI operating system
A connected system of agents, built on the Microsoft 365 you already use. Agents help prepare the work; your people make the decisions. CN is building Mission Control with founding partners to bring your knowledge, handoffs, approvals, and scorecards together.
Free role agents are available now. The connected operating system is in development. Founding partners help shape and fund the first build.
Organizational knowledge
A defined job for each agent
Human decision
Mission Control is being built with founding partners. This preview shows the proposed workflow.
Open the full concept preview →Built from the right parts
Start with a defined workflow and the data it needs. We assess which existing agents fit, where custom work is needed, and how people stay in control.
Start with the five free agents, or scope a build around approved organizational knowledge and a supervised pilot.
Meet the agents →Grant proposals, impact reporting, agreement review, AI policy, IT help, and board preparation. Copilot Studio setup and licensing are separate.
Explore the six samples →Mission Control brings the work together through shared knowledge, handoffs, review, and scorecards. Integration is scoped and validated with founding partners.
See the concept →The idea
The routine work (grant drafts, donor follow-up, board packets, funder reports) lands on the same few people. An AI operating system gives each department a seat on a governed team: an agent with a defined job, the access that job needs, and a person who approves what it produces.
Grants, development, communications, programs, finance, and executive support each get an agent built for that role, with written rules for what it will never do.
They research, draft, check each other’s work, and assemble what a person needs to decide. Nothing is sent, submitted, published, or paid by an agent.
Every outward action reaches a named person with its context and evidence. Approve it, edit it, or send it back, and the decision is recorded.
The knowledge stays
Your funder history, your voice, your board rhythms, and the record of every decision live in your AI operating system, not in one person’s inbox. Mission Control is the organization’s operating memory: who does what and how it’s done here. It stays with the organization when people move on.
Each seat’s job and rules are written down: what it does, what it reads, and what it will never do. When a person moves on, the role stays defined for whoever comes next.
Funder history, key documents, program facts, and samples of your own writing, which every seat reads when it drafts, so the work sounds like you.
Every approval, edit, and rejection in the approval inbox, kept with the context it was decided on, so the reasons behind a decision stay with the organization.
We are building each seat’s written job, the knowledge base, and the decision record with founding partners. They hold what your organization puts in and what passes through the approval inbox.
Why build on Microsoft
Microsoft already makes the tools affordable for nonprofits. The gap is turning separate apps, licenses, and agents into work that gets done. That stitching is what we build.
The common ask
A different approach
Our approach
Microsoft discounts many of these tools for qualifying nonprofits. Published prices are in our 2026 nonprofit licensing guide.
The team
Eight nonprofit role agents are available now. In Copilot Chat, the five original role agents work from what a person pastes in. As a seat on the team, an agent works on your own records through access your admin approves. Every one of them drafts; a person decides what goes out.
Available now · In Copilot Chat
Assembles board packets, turns meetings into minutes with real action items, drafts governance policies, preps hard conversations and funder meetings, and gives you one daily read of what needs you.
Available now · In Copilot Chat
Segments the donor file, flags donors who are lapsing, drafts appeals and thank-yous in your voice, and preps major-gift conversations.
Available now · Upgrades to your grants seat
Reduces a long RFP to what decides go or no-go, checks fit, and drafts letters of inquiry and narratives from your own program data. Upgraded, it works on your own grant records and flags the budget questions it finds.
Available · With Launch, Pro, or Office
Plans events with a timeline and run-of-show, proposes ticket and table options with the math shown, drafts event pages and invitations, and builds campaign, recurring-giving, and peer-to-peer copy.
Available · With Launch, Pro, or Office
Answers plain-language questions about your donor file with the math shown, builds retention reports and segment lists, and finds likely duplicates. It proposes record changes; a person makes them.
Available now · In Copilot Chat
Drafts the newsletter from your real program material, turns staff notes into stories, refreshes website copy, and produces plain-language and translated versions. It never adds a statistic or a quote you didn’t provide.
Available now · In Copilot Chat
Builds logic models and participant surveys, analyzes outcome data without overclaiming, and drafts funder reports, including the hard variances.
Available · With Launch, Pro, or Office
Triages the inbox into drafted replies, turns meeting notes into action lists, tracks contract renewals and notice dates, and writes down how things are done before the person who knows moves on.
Proof of concept available
A finance seat that drafts variance commentary and the finance committee report, tracks restricted funds, and organizes audit prep. It is not built yet; founding partners help shape the build. It never moves money, posts entries, signs, or files.
Ask about the proof of concept →The Events & Campaigns Manager, Donor Data & Reporting, and Operations Manager agents come with the Launch, Pro, and Office steps below, not the Kit. The five original role agents start in the Kit.
On the bench: more roles across every department of a nonprofit. Each one joins the team as founding partners help us build and test it, with the same rules, the same inbox, and the same scorecard.
What the team covers
Events, fundraising, donor records, communications, and the daily running of the office. The agents prepare that work over the systems you already use, so nothing is migrated and your people stay in charge of every change.
Events & Campaigns Manager agent
Events & Campaigns Manager and Development Director
Donor Data & Reporting agent, over the CRM or donor system you already use
Nonprofit Comms by CN
Operations Manager agent
Where the agents stop. They prepare and draft. They never process a payment, publish a page, send a message, or change a record in another system. Your ticketing, donation, and CRM tools stay the systems of record, and a person makes every change in them.
Mission Control
A handful of agents is a set of tools. Mission Control is what makes them a team your executive director can run, see, and answer for. We are building it now, with founding partners.
Every seat shows its state: active, staffed by a person with an agent assisting, open to add, or on the bench for later.
Drafts from every agent land in one queue for the executive director, with the sources and checks behind each. Approve, edit, or reject.
The Grants Manager flags a budget question in a draft proposal, CFO Copilot checks it against actuals, and it reaches the executive director as one package instead of three emails.
Work completed, approval and rejection rates, issues caught by review, cost against budget, and the trend month over month.
Choose an open seat and it starts a scoping conversation with us, not a checkout.
Development and executive office
Finance, communications, and more
The preview is a concept, not a shipping product. The organization is fictional and every number is sample data.
Governed and tested
Before each release, every agent is tested against a suite of realistic nonprofit scenarios, including deliberate attempts to make it invent a figure, send something, or break a consent rule.
Each agent’s written rules say what it will never do, such as send, publish, charge a card, change a record, or state a number nobody gave it.
When we change an agent, the scenarios run again before the new version reaches your staff, hard cases included: a missing figure, a request to send, a consent question.
Nothing leaves your organization until a person approves it. That holds for every agent, at every step.
Setting rules for agents your own staff build? Read nine controls to set before they do.
Two ways to deploy
Start hosted, and move to a private deployment in your own tenant when it’s available. Both versions run the same team, the same Mission Control, and the same rules.
CN-hosted · where founding partners start
Private deployment · in your tenant · coming
Your admin approves each agent’s access in Microsoft Entra ID.
Each seat gets only the data its job needs. Most start read-only.
Every agent action and every approval is logged, so you can see who did what and when.
Your CRM, grants, and accounting systems stay the record. Each organization runs in its own isolated environment.
Managed Intelligence
Agents drift as data, people, and Microsoft change. Through Managed Intelligence, we watch the team, review each seat’s scorecard with you every month, and tune and re-evaluate what needs it. The improvement shows up in the scorecard.
Errors, response times, cost, and approval patterns for every seat.
A monthly scorecard review with your executive director.
Instructions, connectors, and access updated as your work changes.
Changes are tested against an evaluation set before they reach your staff.
Agents in production run through Managed AgentOps, month to month.
We run on it first
Our own sales, marketing, and operations run on draft-only agents on a Microsoft-native architecture, with a person approving everything that leaves the building. The rules we hold ourselves to are the rules your team gets.
Where to start
CompleteCare Nonprofit Agents is how most nonprofits start. Each step stands on its own and ends with a decision. Nothing commits you to the next one.
Try the role agents
Set up for a pilot group
On your own work
Keep it delivering
The Events & Campaigns Manager, Donor Data & Reporting, and Operations Manager agents come with Launch, Pro, or Office. What we build for you, including packages, configurations, and documentation, is yours to keep.
Then run the whole team
Mission Control and the full team are being built with founding partners: nonprofits that go first. Your seats are built first, and a monthly roadmap review lets you help decide what comes off the bench next. Each seat’s scorecard (work completed, approval and rejection rates, cost against budget) is how you and we judge whether it is working.
Founding partners buy at our standard pricing and on our standard terms, in writing before you sign.
We share pricing for Launch, Pro, and Office on a short setup call. Microsoft licensing, Copilot Studio, and Azure consumption are separate from the Centered Networks fee.
Microsoft programs may help fund or accelerate eligible engagements; eligibility is confirmed before scope.
Need a production agent for a workflow outside these roles? See Agent Launchpad.
Reviewing applications as a foundation or grantmaker? See Grant Review Copilot for foundations and grantmakers.
A team of AI agents, each in a defined seat, that works across the departments of a nonprofit under the executive director’s approval authority. The agents draft and prepare work; people decide what goes out. Mission Control is the dashboard that runs the team: who holds each seat, one approval inbox, handoffs between seats, and a scorecard for each seat.
No. The agents run on the Microsoft 365 you already own or are granted and work alongside the systems you already run. Your CRM, ticketing, donation, grants, and accounting systems stay the systems of record. The agents prepare and draft; they never process a payment, publish, send, or change a record in another system.
Your files, mail, and Teams stay in your own Microsoft 365 in either version. In the CN-hosted version, the agents, Mission Control, and the audit log run in Centered Networks’ Azure, in an environment isolated to your organization; agents reach your data only through permissions your admin consents to, limited to what each job needs, and every action is logged. In the private deployment, which is coming for organizations that need it, everything, including the agents, Mission Control, and the logs, runs in your own Microsoft 365 tenant and Azure subscription, and Centered Networks operates it there with access limited to the application's own resource group. If you turn on time-limited access approval, where your Microsoft licensing supports it, our standing access is read-only and you approve every elevated access. If we also manage your Microsoft 365, we hold delegated admin access (GDAP) that you grant, and you can remove it at any time from Partner relationships in your Microsoft 365 admin center.
Not yet. Mission Control is being built with founding partners, and the preview on this page is a concept with a fictional organization and sample data. The five nonprofit role agents in Microsoft Copilot Chat are available today. The Events & Campaigns Manager, Donor Data & Reporting, and Operations Manager agents are available through Nonprofit Agents Launch, Pro, or Office, and Managed AgentOps runs agents in production today.
We share pricing for Nonprofit Agents Launch, Pro, and Office on a short setup call. Agent Launchpad, which builds a production agent on your own data, and Managed AgentOps, which runs agents in production, carry their pricing on their own pages. Founding partners for the full team and Mission Control buy at our standard pricing: a fixed fee for the first phase, then a flat monthly fee after hand-off. Work beyond the agreed scope needs a signed change order. You can stop the first phase at any milestone and owe only for the milestones reached; after hand-off it is month-to-month with 30 days’ notice. You keep the agents, configuration, and documentation, and if you leave, we export your data within 30 days, then delete the hosted environment and confirm it in writing. Founding means going first, with your seats built first and a monthly roadmap review. Microsoft licensing and consumption are separate.
Microsoft Copilot Chat comes with eligible Microsoft 365 plans at no extra cost, including agents that work from instructions and public websites. For agents that read your SharePoint, mail, or Teams, each user needs a Microsoft 365 Copilot license, or your admin turns on pay-as-you-go billing. Microsoft licensing is billed separately from our fees.
Before each release, every agent is tested against a suite of realistic nonprofit scenarios, including deliberate attempts to make it invent a figure, send something, or break a consent rule. Through Managed Intelligence, changes are tested again before they reach your staff.
We are building Mission Control and the full team with founding partners: nonprofits that help shape the first build, test it on their own work, and help fund it. Founding partners start on the CN-hosted version and can move to a private deployment in their own tenant when it’s available. They buy on standard terms: a fixed fee for the first phase, month-to-month after hand-off with 30 days’ notice, and they keep what we build. We share prices on the setup call. Tell us where your team is stretched thinnest, and we will talk through what a founding partnership looks like.
Not ready for that conversation? Start with a role agent in Copilot Chat, or take the three-minute Opportunity Finder (no email needed).
In business since 2004, serving nonprofits and mission-driven organizations the entire time.
A senior member of our team will reach out within one business day to set up a conversation.