AI that works for fundraisers — not against donors.
We use AI only when it makes giving easier, fundraising smarter, and our products safer. Donor data is protected, every use case is reviewed, and we take our footprint seriously.
AI should serve movements, campaigns and causes — not replace the people behind them.
Every AI feature we ship across GoodChange, SmartChange and AgentChange is grounded in one mission: helping campaigns, nonprofits, membership organizations and grassroots groups raise more — faster, more transparently, and with more respect for the people who give. Three commitments guide every decision.
01
Privacy & Security
Donor data is not currency. We protect it with encryption, strict access controls and infrastructure built for sensitive workloads. We never sell data. We never use customer or donor data to train AI models.
How we do it02
Governance & Ethics
We don't deploy AI for AI's sake. Every use case is reviewed before launch against our own values. AI on our products is mission-driven, designed for fairness, transparent in operation, and always under human oversight.
How we do it03
Environmental Sustainability
Power without waste. We deploy AI efficiently — using existing models instead of training new ones, and choosing right-sized models for the job. We review, improve and evolve our practices as the technology matures.
How we do itPrivacy & Security.
Protecting donor data isn't optional — it's the foundation of everything we build. These commitments hold on every Level Wins product.
We never use client or donor data to train AI models, on any Level Wins product.
Every contact, gift and record belongs to the organization that brought it. We never sell it, and you can export it.
An AI conversation is visible only to the person who started it. GoodChange and SmartChange let you choose to share one with teammates.
On GoodChange and SmartChange
- Governance aligned with PCI-DSS (payment card data) and SOC 2 principles.
- Encryption at rest and in transit, role-based access, full audit logging, and multi-factor authentication that organizations can require for their team.
- Payments processed through Stripe, with Stripe Radar fraud screening on every card payment.
- Every AI request is sent with zero-data-retention settings, so our AI providers don't keep your prompts or results. Conversations are stored in your account so you can reopen them, and deleted when you delete the session. Uploaded attachments expire after 30 days.
On AgentChange
- Every workspace is walled off from every other at the database level, and only the person who wrote a conversation with the Agent can read it.
- AI requests are sent with response storage turned off. Our AI provider may keep limited logs for abuse monitoring under its own policies.
- Connections to Google and other sources are optional, under your control, and stored encrypted. Google data is used only under Google's Limited Use requirements and never to train AI models.
- AgentChange never sends email to your contacts. It drafts; a person reviews, edits and sends from their own email.
Governance & Ethics.
Our use of AI comes from our mission and our values — nothing else. Every AI use case is reviewed by our team before launch, weighing the risks, the benefits and the alternatives. Governance is built into the foundation, not bolted on after the fact.
We build AI only when it advances the missions of the people we serve — campaigns, nonprofits, membership organizations, NIL collectives, advocacy groups. If a feature doesn't make fundraising better, it doesn't ship.
We work to identify and reduce bias in outputs. A tool that works well for some organizations and poorly for others isn't a tool we'll ship.
We tell you when and how AI is applied. On GoodChange and SmartChange, AI-generated suggestions are labeled and AI-assisted features are documented in the help desk.
Human oversight on every high-stakes decision. AI helps draft, suggest, summarize and analyze — real people approve, send, and own the outcome.
Environmental Sustainability.
AI has a real environmental cost, and we take it seriously — through design choices, infrastructure choices, and weighing whether each feature is worth its footprint.
We apply existing models rather than training new ones. Training consumes the most energy by far; using a model is dramatically lighter.
We don't reach for the largest model when a smaller one will do. Drafting a thank-you note doesn't need the horsepower of analyzing a year of giving.
GoodChange and SmartChange run on Amazon Web Services. AWS reports its data centers averaged a power usage effectiveness of 1.14 in 2025, and Amazon matches 100% of its annual electricity use with renewable energy, verified by an independent third party.
We'll keep adopting more efficient methods as they emerge. As research evolves, our practices evolve with it.
Where we use AI today.
Concrete examples, each reviewed and shipped against the principles above.
Ask about your data
Ask the AI assistant about donors, gifts, refunds and payouts in plain English. Answers come from the same data your dashboards use, limited to what your role allows.
Proposed changes you approve
The AI can propose updates to forms, events and donor records. Each shows exactly what will change, and nothing is saved until a person on your team approves it.
Donor email drafts
First drafts and one-click rewrites of donor updates and thank-you notes. You review every draft and send it yourself.
Fraud & risk detection
Every card payment is screened by Stripe Radar, Stripe's machine-learning fraud detection. Bot protection at checkout keeps automated attacks out.
Strategists you can ask
Ask the Agent what's on your mind and get a concrete next step, grounded in your own workspace. It advises; it doesn't change your records.
Segments and next moves
Ready-made segments surface the conversations worth having today — lapsed supporters, new relationships, high-capacity prospects.
Drafted outreach and thank-yous
Drafts in your voice for emails, posts and thank-you notes. A person reviews, edits and sends every one.
What we will and won't do.
What we do
- Use AI to assist people — drafts, summaries, suggestions, search.
- Keep humans in the loop for every high-stakes decision.
- Review every AI feature before it ships.
- Let organizations export their data.
What we won't do
- Train AI models on client or donor data.
- Sell donor data to advertisers or third parties.
- Let AI send messages to your donors or contacts on its own.
- Let AI auto-approve campaigns, refunds or escalations.
- Replace human support with a chatbot you can't escape.
- Use AI for AI's sake when a simpler tool would do.
Questions about AI at Level Wins.
Is my data used to train AI models?
No. We never use client or donor data to train AI models, on any Level Wins product.
Do your AI providers keep my prompts?
On GoodChange and SmartChange, every AI request is sent with zero-data-retention settings, so our AI providers don't keep your prompts or results. On AgentChange, requests are sent with response storage turned off; our provider may keep limited logs for abuse monitoring under its own policies.
Can other people see my AI conversations?
Not unless you share them. A conversation is visible only to the person who started it. On GoodChange and SmartChange you can choose to share one with teammates.
Can the AI send messages or make changes on its own?
No. AI drafts and suggests. On GoodChange and SmartChange, any change it proposes needs a person on your team to approve it. AgentChange never sends email to your contacts — a person reviews, edits and sends every message.
Does AI auto-approve campaigns or transactions?
No. Every high-stakes decision — campaign approval, refund processing, compliance escalation — is reviewed by a person. AI is a tool, not the decision-maker.
Do these rules apply to every Level Wins product?
Yes. The principles on this page — privacy, governance and sustainability — apply to GoodChange, SmartChange and AgentChange. Where the technical details differ between products, this page says so.
Where can I report a concern about an AI feature?
Email hi@levelwins.com, or the support team of the product you use. Every report is reviewed, and feedback on AI features goes directly to the people responsible for governance and product decisions.
AI must be ethical, secure and sustainable.
This work is never finished. Trusted AI is built in partnership — with our clients, our donors and the communities we serve. We refine through feedback and evolve as the missions we support evolve.