AI for Nonprofits (5 of 10): Map Your Nonprofit AI and Technology Landscape Before Buying Anything
Your nonprofit is probably already using more AI than anyone realizes.
It may be in Microsoft 365 or Google Workspace, the CRM, an email platform, a meeting transcription service, a fundraising product, a design tool, or a staff member’s personal account. The first technology task is not adding another product. It is understanding the environment you already have.
Build a practical inventory
When I conduct a technology stack audit, I start by asking for a complete list of the organization’s technology. Often, that list still needs to be created. Start with the tools people actually use, the work each supports, and the problems staff are trying to solve. Our earlier article, Evaluating Your Nonprofit Technology Stack: A Practical Framework, provides a broader starting point for that review. Then, for each platform, document:
- Product, plan, purpose, and business process supported
- Business owner and technical administrator
- Users and departments
- Information stored or accessible
- AI features and agents, with status marked as enabled, available but disabled, optional, or roadmap only
- License, add-on, usage, implementation, and ongoing support costs
- Integrations with other systems
- Sign-in methods, permissions, connected accounts, and whether AI can read information or take action
- Contract term, renewal date, and date the inventory entry was last verified
- Known limitations or adoption problems
Include unofficial tools discovered during staff interviews, browser extensions, personal AI accounts used for work, meeting bots, and automations built by individual teams. Ask without blame so staff are comfortable sharing what they use. Review invoices and administrator consoles to help confirm the list; do not record passwords or sensitive constituent information in the inventory.
Separate capability from availability
A vendor may advertise an AI feature that is not included in your plan, not available in your region, disabled by your administrator, or unable to work with the data your organization uses.
Confirm what is actually available in your account. Then determine whether staff know it exists and whether it solves a meaningful problem.
Record who verified each feature and when. An advertised capability, an active license, and a tool staff use successfully are three different things. Recheck the inventory at renewals and when vendors introduce major changes.
Follow the data
AI value often depends on access to useful information. Map where donor, beneficiary, volunteer, finance, program, communication, and operational data live. Identify how information moves between systems and where spreadsheets or manual exports serve as the integration layer.
Ask whether the data needed for a use case is complete, current, accessible, and appropriately permissioned. A sophisticated AI tool cannot compensate for records that are unreliable or trapped in disconnected systems.
For each important data category, identify the authoritative source and who is responsible for accuracy. A shared view of a constituent may require connected systems and consistent identifiers; it does not automatically require moving every department into one application.
Review the productivity environment
For many nonprofits, the Microsoft or Google environment will be the starting point for their AI discovery. It contains email, calendars, documents, spreadsheets, meetings, and shared files—the daily context for much of the organization’s work.
Review licensing, file organization, sharing practices, permissions, retention, identity management, and administrative ownership. AI can make well-managed information easier to use. It can also make disorganized or overexposed information easier to find.
Check the access granted to connected AI tools and agents separately. Record whether each can only retrieve information or can also update records, send messages, or trigger workflows. Identify required approvals, activity logs, and how an administrator can revoke access or stop an action.
Identify duplication and gaps
Compare overlapping capabilities. Are multiple products summarizing meetings? Are teams buying separate chat tools? Does the CRM provide a feature that another vendor is proposing? Could a straightforward automation solve the need without adding generative AI?
Also identify gaps: missing integration, inconsistent permissions, poor search, inaccessible data, or no one responsible for administration.
Evaluate vendors in context
Responsible vendor review should address privacy and security, data use, retention, model training, subcontractors, administrative controls, auditability, accessibility, bias testing, incident response, and contract exit.
This review supports the Fundraising.AI framework’s Privacy and Security, Data Ethics, Accountability, and Transparency and Explainability principles. Its Sustainability principle also asks organizations to consider long-term sustainability and environmental impact. Alongside those responsibilities, evaluate whether your nonprofit can afford and administer the capability over time. Read the Fundraising.AI framework
What the landscape should reveal
The completed landscape should show:
- AI capabilities already available
- Redundant products or licenses
- Important data and integration dependencies
- Permission or governance concerns
- Technology gaps that must be addressed first
- Tools that warrant deeper evaluation
Use the findings to decide what to keep, configure, connect, evaluate, or retire. Assign an owner and next review date to each priority. Resolve data, security, and governance gaps before allowing AI to use sensitive information or act on the organization’s behalf.
How nonprofits are putting this into practice
McKnight Foundation uses Microsoft 365 Copilot within its existing Microsoft environment to summarize information across Teams, SharePoint, and Outlook and help staff draft content. Microsoft’s customer story describes how the rollout builds on established security, permissions, and retention policies, supported by staff guidance and training. The lesson for a technology inventory is to examine the capabilities and controls in the environment your team already uses. Read the McKnight Foundation story
Centro de la Familia de Utah worked with JourneyTeam to build a reporting solution using Power Apps, Power BI, and Dataverse, with role-based access to data. It then added a Copilot Studio agent to help summarize notes and prepare reports. This illustrates why mapping the underlying data, reporting workflow, and permissions matters when evaluating an AI opportunity. Read the Centro de la Familia story
What you can do today
- Start an inventory of the platforms your team uses, including purpose, owner, renewal date, AI features, and current cost. A shared spreadsheet is enough to begin.
- Ask your major vendors to confirm which AI features are included, optional, enabled, or planned; their full costs; and what information they can access. Request written answers about data use, retention, model training, administrative controls, and contract exit.
- For any AI agent that can act, confirm its permitted actions, approval steps, activity logs, and how to stop it. Ask how outputs can be explained, corrected, or overridden, and how the vendor tests accessibility and bias.
- Identify one place where spreadsheets or manual exports connect otherwise separate systems.
- Review who owns administration and permissions for your Microsoft or Google environment.
- Assign one person to maintain the inventory and schedule a quarterly review with department owners. Start with the tools involved in your first AI pilot, then expand the map.
Responsible AI questions to consider
- Where is AI operating today, including features staff may not recognize as AI?
- Can each vendor explain recommendations, permissions, data use, limitations, and avenues for correction?
- Can the organization administer and sustain the capability without hidden risk or dependence?
Next ThursdAI: We will look at data quality, permissions, security, public transparency, and human oversight.what it can access and do, and the vendor questions that need answers.
Sign up to be notified when the next ThursdAI article goes live!
—
Cathexis Partners helps nonprofits assess their technology environment, improve business processes and select platforms based on documented organizational needs. Contact us to discuss your next step.
