A Guide to Nonprofit AI Implementation for Nonprofit Leaders and Staff

In a previous blog on the Nonprofit Standard, we discussed how nonprofit boards can help their organizations implement AI securely, effectively, and equitably. 

But nonprofit AI implementation does not depend on the board alone. Nonprofit leaders, including executive directors and staff members, have a critical role in successful AI adoption.

Leaders and staff have direct insight into how AI can streamline the day-to-day work of their organizations. They can use their on-the-ground knowledge to show the board where AI can make the greatest impact and why. 

For AI to achieve its potential, there are implementation should follow several steps. Nonprofit leaders and staff need to identify where AI can add value, develop a plan for implementation, address data governance and security, and look to continuously test and iterate on AI projects. Each of these steps requires input from the board so that AI initiatives can advance the organization’s objectives.  

Identify Where AI Can Be Valuable

Board members may be able to provide insight into how other organizations leverage AI or bring relevant experience from how they are using AI elsewhere in their professional and business capacity. 

Top Uses for AI at Nonprofits

Leaders should identify two to three nonprofit AI use cases that will unlock the most value, whether that is streamlining their grant writing process, automating volunteer management, or enhancing financial management. Once nonprofit leaders can demonstrate that AI initiatives have achieved project goals, they can expand AI initiatives to additional areas of the organization. 

Develop a Plan for AI Implementation

After identifying where AI can provide tangible value to the organization, the next step is to develop a plan for implementation.

An effective plan should address: 

• What success means for the organization: AI can help optimize resources, encourage innovation, or both, so leaders and staff should determine their main objective for AI implementation. Then, after rolling out specific AI pilots, define the metrics that will determine their success. Initial projects should be measured on whether they can validate a hypothesis. For example, is this pilot actually making the grant writing process more efficient, or are staff spending too much time reviewing and correcting the AI’s outputs? As projects mature, leaders will likely need to adjust success metrics.
• Which stakeholders will need to provide buy-in: For AI projects to be successful, various stakeholders will need to be engaged and understand how to use the selected AI tools. In addition to working closely with the board on the AI rollout, nonprofit leaders will need to provide training on how staff, and potentially volunteers, can use new AI technology programs. In addition, beneficiaries and donors will need to consent if their data is being used in new ways. 
• The impact on programs and service delivery: How might AI implementation impact programs and services, if at all? For example, if an organization is using AI to analyze program effectiveness, the organization may need to update the data being collected from program beneficiaries. This would require surveying or interviewing beneficiaries. 
• How AI supports the organization’s mission: Above all, state how AI projects will further the organization’s mission. For example, reduced costs on the financial, operational, and administrative side means more dollars can be dedicated to program delivery. Also, greater insight into program effectiveness can help leadership understand where to make adjustments to increase impact and effectiveness.

As leaders and staff build out an implementation plan, they should work closely with the board to make sure it aligns with their AI vision and adheres to governance, ethics, and security requirements. 

Address Data Governance and Security Concerns

AI should be used in a manner that is safe, effective, ethical, and aligned with data governance policy. While the board’s primary role in AI implementation is to provide oversight, this will also require close collaboration with nonprofit leadership. Additionally, nonprofit staff can provide critical feedback on how new data governance structures impact their day-to-day activities. Boards and leadership will want to find the right balance of security and allowing access to data that can unlock valuable insights. 

Consider BDO’s eight key elements of Good AI Governance: 

1. Principles and Guidelines: Identify oversight roles and responsibilities. Develop the framework and create user guidelines to facilitate the initiation of AI projects. 
2. Operating Model and Standards: Establish and apply technical standards and adopt common AI standards for the organization. 
3. Human-centric Design: Consider privacy and data protection from the onset of design and implementation, while mandating human review of AI inputs and outputs. 
4. Transparency and Explainability: Track and analyze the development and implementation of AI. Stay educated on AI development, opportunities, and challenges, as well as its economic, social, and environmental impacts. 
5. Accountability and Responsibility of AI: Protect individual rights by monitoring and reporting on the compliance of AI systems with data privacy standards.  
6. Systems and Applications: Develop impact assessments and audits that scrutinize AI applications and their compliance with legal, ethical, and technical requirements. 
7. Compliance: Test and monitor systems throughout their life cycle, ensuring that users maintain records and report incidents. 
8. Cybersecurity: Establish the processes, procedures, and controls to identify and respond to cybersecurity risks. Develop incident response plans and templates. 
   
   

Learn more in our Practical Guide to AI. 

Test, Grow, and Change

Some nonprofits interested in exploring AI may get stuck in the planning phase. Once the initial frameworks are established, focus on introducing a minimum viable product (MVP) use case. An MVP should use the least amount of time, budget, and resources, and may require a private sandbox environment to test the product and protect intellectual property. Test the MVP with a small group to determine if it provides value. 

If the MVP does provide value, continue to test and iterate. Encourage staff to test programs and provide feedback. Some AI models allow users to experiment with new use cases, which allows innovation to emerge from any corner of the organization. Throughout this testing journey, provide feedback to the board on how projects are progressing, any initial success metrics, and challenges or areas for improvement. 

Collaboration Precipitates Success

The key to any successful nonprofit AI implementation project is for nonprofit leaders, staff, and board members to collaborate closely. While there will naturally be some overlap in the duties of each role, the board’s role is to provide oversight and governance policies. Staff and leaders, on the other hand, have more hands-on knowledge of where AI can provide the most value, foresee any obstacles to implementation, and become AI advocates within the organization. 

By working together, nonprofit leaders, staff, and boards can transform nonprofit work using AI and enable their organizations to increase their impact. 

For a more comprehensive overview of the structure and guardrails necessary for using artificial intelligence responsibly, AI implementation, download our AI Governance Guide.