What is Steward Agents?
Steward Agents gives staff at Canadian non-profits their own AI agent, shaped to their role: executive, finance and administration, programs, development, communications, grants, or policy and advocacy, plus an advisor on funding and planning AI adoption. Each agent drafts and analyzes the kind of work that role does, from board updates to grant narratives. People review and decide.
Who are the agents for?
Executive directors and CEOs, and staff who lead finance and administration, programs, fundraising and development, communications, grants, and policy or community engagement. Most agents are designed through a short onboarding interview with the person who will use them; ready-made agents are available to start on day one.
Which Microsoft apps does it work in?
A Steward Agents panel opens on the right in Outlook, Word, Excel and PowerPoint, on the web [and in the desktop apps: confirm after testing]. There's also a web app at app.stewardagents.com, and an optional connection for Microsoft 365 Copilot. You sign in with your Microsoft work account.
Does it send email?
No. Steward Agents doesn't have permission to send email from your mailbox. In Outlook it opens a reply for you to edit, or saves a draft in your Drafts folder. You send it yourself.
Will it overwrite my Excel numbers?
No. In Excel it reads what you select, and when you choose to add its summary, it writes below your data or on a new sheet. It doesn't change existing cells.
Does it replace staff?
No. Agents take on first drafts, summaries and checking so your people spend more time on relationships, judgment and decisions. A person reviews everything, and your organization stays responsible for what goes out. See AI Use & Human Review.
Can we use it with client or family information?
No. Steward Agents is for staff work, not case work. Don't put client or family personal health or case information into it. It checks for signs of that information and refuses to send it to the AI, but that's a safeguard, not permission. See the Acceptable Use Policy.
Where is our data stored? Is it in Canada?
Stored data stays in Canada: the application runs in Toronto and the database in Montréal. To answer each request, the request and the content needed for it are sent to our AI provider, Anthropic, and by default that processing happens mainly in the United States. A Canada-only processing option, with some features limited, is available on request. Under the provider's commercial API terms, that content isn't used to train its models. Details are in the Privacy Policy.
Who can see our conversations?
Each person's agents and conversations are private to them. Steward Agents staff see usage and cost information to run the service, and conversation content only if your organization turns that access on. See the Security Overview.
How long is a pilot, and what happens after?
A pilot gives one to three staff members their own agents for a set period, with a review against the success criteria you set at the start. [Confirm standard pilot length, for example 30, 60 or 90 days.] At the end, you decide whether to continue on an annual subscription, change the setup, or stop. If you stop, we delete your data as described in the Privacy Policy. Pricing is shared on request.
What does setup involve?
Each person does a guided onboarding interview about their role, which becomes their agent's design. Your Microsoft 365 administrator approves the Steward Agents app once and installs the Office add-in for the people using it. We walk you through both.
How does support work?
Email us at [support email]. We reply within [response time] on business days, Pacific time. [Describe pilot support, for example a named contact and scheduled check-ins.] See Contact & Support.
Can we cancel?
Yes. Cancellation and refunds are covered in the Terms of Service.