Before You Automate: Why Purpose and People Must Come Before AI Adoption

I can still remember the first-time artificial intelligence moved from something I had heard about to something I wanted to understand.

I was sitting in a hotel room in the United States, half-listening to the news, when a story came on about the launch of ChatGPT. I was already familiar with Jasper and had been curious about generative AI, but this felt different. The attention was immediate. The language around it was bigger. Something was shifting.

So, I downloaded the app.

At first, I played with it. I asked questions, tested responses, challenged assumptions and quickly discovered both the promise and the flaws. The data was not always current. Bias was noticeable. Confidence sometimes exceeded accuracy.

Six months later, I found myself in another hotel room, this time in Toronto with my dad, who was recovering from hip surgery. We were there because the health system in his home province of Alberta had not supported him getting the surgery he needed in a timely way. In between appointments, recovery, logistics and family conversations, we researched the health system. We asked questions. We compared options. We used AI not as an oracle, but as a thought partner.

At one point, frustrated by a few of the answers, I told ChatGPT it reminded me of a politician. It did not take that especially well. The conversation, apparently, was over.

Looking back, that period shaped how I think about AI now. 

It was never only about technology. It was about access, systems, information, judgement, bias, trust and human care.

Three years later, I have now attended my third AI conference. I continue to learn an enormous amount, including from Julian Moore , whose ability to translate AI into practical business and leadership application has been valuable to me over several years.

At the Leading the Future AI Summit 2026, the conversation had clearly moved on from “what is AI?” to “how do we use it well?” Practical adoption, workflow automation, AI agents, cyber security, real-world business case studies and AI visibility were all part of the discussion.

And yet, the strongest message I took away was not about a platform, a tool or a technical trick. 

It was this: before we commit to AI solutions, we need to understand the problems we are trying to solve.

The AI Readiness Gap

Many organisations are still approaching AI in the wrong order. They start with the platform. They ask which tool to buy. They look for shortcuts. They want automation before diagnosis.

The better starting point is much simpler, and much harder: What problem are we trying to solve? What takes too much time? Where are staff stretched, systems fragmented or stakeholders poorly served? Where could AI improve quality, not just speed?

This matters acutely in the not-for-profit sector. Many purpose-driven organisations are under-resourced, under pressure and expected to do increasingly complex work with limited capacity. AI offers genuine opportunity. It can support research, drafting, analysis, workflow design, donor communication, segmentation, reporting, knowledge management and operational efficiency.

But AI cannot compensate for unclear strategy, weak governance, poor data, fragmented culture or a lack of trust. 

If anything, it will expose those weaknesses faster.

Purpose Before Platforms

For purpose-driven organisations, AI adoption should begin with purpose, not productivity.

Productivity matters. Efficiency matters. Reducing administrative burden matters. But the not-for-profit sector does not exist to produce more content, send more emails, automate more tasks or chase every new tool. It exists to advance mission, serve communities, steward trust and create impact.

That means the first AI question is not “what can we automate?” 

It is “what should we improve, and why?”

This distinction matters. AI can help fundraising teams draft donor communications, but it cannot replace the trust built through genuine relationship. It can help analyse stakeholder data, but it cannot decide what an organisation values. It can support a grant application, but it cannot create institutional credibility. It can help tell a story, but it cannot determine whether that story should be told, whose voice is centred, or whether consent has been properly considered.

One of the best comments I heard at the conference was that AI cannot collect eggs. In the context of a regenerative farm, that point was both practical and profound.

AI can help with many things. But not everything.

The discipline is knowing the difference.

Ethics Is Not an Add-On

The ethical questions around AI are not theoretical. They are already operational.

What data are we putting into AI systems? Who owns the output? How do we manage privacy? How do we check for hallucinations, bias and over reliance? How do we make sure people remain accountable for decisions?

This is why I was so interested in a recent The Economist article about why major AI labs are hiring philosophers. The article argues that AI now raises “thorny problems” around truthfulness, moral reasoning, safety, overconfidence and ethical decision-making, exactly the kinds of questions philosophy has wrestled with for centuries.

That took me straight back to conversations we had during the Vincent Fairfax Fellowship, where ethics, leadership, technology, and philosophy were not separate topics. They were deeply connected.

The Economist article also discusses concerns about “moral deskilling”: the risk that, if machines increasingly make ethical calls, humans may become less willing or less able to make their own judgements.

That concern should matter deeply to the not-for-profit sector.

Our work depends on judgement. It depends on values. It depends on context, empathy, accountability and trust. If AI is used to support those things, it can be powerful. If it is used to avoid them, it becomes dangerous.

People Before Technology

Another theme that stood out to me is that implementation is largely a people challenge, not a technology challenge.

AI adoption is often described as a project. I think that is the wrong frame.

AI is not a project. It is a practice.

It requires experimentation, training, testing, policy, review and cultural change. It requires people to feel safe enough to learn but disciplined enough not to be reckless. It requires leaders to create sandboxes where staff can explore use cases without exposing sensitive data or making unsupported claims.

This is particularly important in organisations where people are already stretched. Staff should not simply be told to “use AI” on top of everything else. They need guidance on where it is appropriate, where it is not, and how to use it well.

The best approach is to identify, pilot, select and scale.

Start small. Play in the sandbox. Test with low-risk tasks. Build confidence. Check quality. Understand the risks. Then decide what should be scaled.

Do not scale what you have not tested.

What This Means for Fundraising and Advancement

I had a small but useful wake-up call recently.

During an interview with a Mercor avatar for an expert advisory project, I was asked about the different uses of AI in fundraising. I answered, but I stumbled. That bothered me, because it should not happen. Not anymore.

As a consultant in the not-for-profit sector, I need to be able to provide practical leadership and guidance. That means moving beyond general enthusiasm into disciplined fluency.

In fundraising and advancement, AI can support donor and prospect research, segmentation, stewardship, grant preparation, campaign planning, board reporting, data cleaning, event follow-up, content repurposing, scenario planning, policy drafting, training resources, and impact reporting.

But the answer is not always “invest in AI instead of hiring staff.”

The answer depends on the role, the problem, the risk and the relationship at stake.

If the work is repetitive, rules-based, time-consuming and low-risk, AI may create significant efficiency. If the work requires trust, judgement, lived experience, relationship-building or ethical sensitivity, AI may support the person doing the work, but it should not replace them.

The opportunity is not to remove humans from the work.

It is to give humans more time for the work only humans can do.

AI Visibility and the New Search Reality

One of the more practical ideas from the conference was the question of whether a business or organisation is visible to AI.

This matters for not-for-profit organisations too.

We are used to thinking about search in terms of Google, websites and keywords. But AI does not present information in the same way. Increasingly, people are not asking for ten links. They are asking for an answer.

That means organisations need to be clear, consistent and findable.

Your website, LinkedIn profile, about page, FAQs, testimonials, media mentions, and external citations all matter. If AI tools are forming a shortlist, they can only recommend what they can confidently describe.

For not-for-profit organisations, this is not about chasing the latest version of SEO. It is about clarity.

Can your organisation be easily understood? Is your purpose clear? Are your services, impact, and credibility visible? Are your facts up to date? Do you answer the questions your stakeholders are already asking?

AI slop is real. So is invisibility.

Both are risks.

Practical Reflections for Boards and Executive Leaders

For boards and executive teams beginning to consider AI adoption, the following questions are worth asking before committing to a tool or platform:

• What problem are we trying to solve?
• Where could AI improve efficiency without weakening relationships?
• What data should never be entered into public AI tools?
• Have we considered cyber security before experimentation becomes widespread?
• Where is human judgement non-negotiable?
• Are staff being trained, or merely encouraged to experiment?
• Who is accountable for AI-assisted work?

These questions are not designed to slow progress. They are designed to make progress safer, smarter and more useful.

The Courage to Learn

I have been an early adopter of AI, but I have not yet benefited from it to the extent I could have.

That is about to change. I have a lot of work ahead of me.

Not because AI is a trend I need to follow, but because it is now part of the leadership landscape. As a consultant, educator and adviser to purpose-driven organisations, I need to understand not only what AI can do, but what it should do, where it should be used, where it should not be used, and how organisations can adopt it without compromising trust.

For the not-for-profit sector, this is only the beginning.

The organisations that benefit most will not be the ones that automate everything. They will be the ones that understand their purpose, identify the right problems, build staff capability, protect sensitive information, strengthen governance, and keep human judgement at the centre.

The question is not whether AI will affect the not-for-profit sector.

It already has.

The question is whether we will adopt it with enough purpose, discipline, and ethical clarity to ensure it strengthens the human work we are here to do.

And did AI help me with this article? Of course it did.

It was a collaboration.

The thinking, judgement, experience, and final responsibility remain mine.

References

• The Economist (2026) ‘Why big AI labs are hiring so many philosophers’, The Economist, 24 June 2026.
• Leading the Future AI Summit 2026, Peel Chamber of Commerce & Industry.
• Moore, J. Leading the Future AI Summit 2026 keynote: Making AI Work in the Real World.
• OpenAI (2022–2026) ChatGPT.