If you believed every AI advertisement, you’d think you’re only one subscription away from running an entire organisation with the click of a button.
- Your customer service? Automated.
- Your reports? Written.
- Your meetings? Summarised.
- Your marketing? Generated in seconds.
In Part 1 of this series, we discussed whether you should be using AI at all. Now let’s talk about what to actually expect if you decide to explore it.
It sounds impressive, and in some cases, it genuinely is. But this is also where many businesses and nonprofits begin their AI journey with expectations that no technology could realistically meet.
The disappointment isn’t usually because AI doesn’t work, it’s because the gap between the marketing and the reality is much wider than most people expect.
The truth is simpler. AI is a tool with remarkable strengths, equally real limitations, and outcomes that depend as much on the organisation using it as the technology itself.
What AI is genuinely good at today
The easiest way to understand AI is to stop thinking of it as a replacement for people and start thinking of it as an accelerator. When given clear instructions and good information, AI can complete certain tasks in seconds that might otherwise take hours.
Let’s examine some of them:
Generating and summarising text
This is perhaps AI’s strongest and most widely used capability. Whether it’s drafting emails, creating meeting summaries, producing first versions of reports, or simplifying complex documents, modern language models can dramatically reduce writing time.
The important word here is drafting. Good organisations don’t publish AI-generated content untouched. They use it to create a strong starting point, then rely on human judgment to refine, verify, and personalise the final result.
Finding patterns in data
AI is excellent at recognising trends that humans might overlook, especially when working with large volumes of information.
Businesses can use it to identify purchasing patterns. Nonprofits can analyse donor behaviour. Healthcare providers can detect anomalies in patient data.
The technology doesn’t magically understand your organisation, it simply processes far more information than a person can review manually.
Handling repetitive, high-volume tasks
One of AI’s biggest strengths is consistency. Tasks like categorising documents, responding to common customer enquiries, extracting information from forms, and routing requests to the right team can often be automated successfully.
These aren’t glamorous use cases. They’re simply the kind of repetitive work that consumes valuable staff time every day.
Working with language
Translation, transcription, and speech-to-text have improved dramatically. Meetings can be transcribed in minutes. Large document collections can be organised automatically.
For organisations dealing with high volumes of written or spoken content, this is a genuine time saver.
Speeding up research and idea generation
AI is particularly useful for overcoming the blank page problem. Need a first draft, ten campaign ideas, or alternative ways to explain something complex?
AI generates useful starting points quickly. The value is not the perfect answer. It is helping you begin.
Where AI consistently falls short
The same technology that performs brilliantly in one situation can produce disappointing results in another. Understanding these limitations is just as important as understanding its strengths.
Deep local knowledge
AI models are trained on enormous amounts of publicly available information. That doesn’t mean they understand every community, industry, or region equally well.
Ask a model about a major global company, and it will often perform well. Ask it about a small local organisation, regional regulations, or community-specific practices, and the quality can decline significantly.
Context matters. And local context is often the hardest thing for AI to understand.
Situations where accuracy is non-negotiable
AI can sound extremely confident while being completely wrong. This is one of its most misunderstood characteristics.
If you’re making legal decisions, financial recommendations, medical assessments, or regulatory submissions, AI should support human expertise, not replace it.
When the cost of being wrong is high, verification isn’t optional.
Genuine creativity
AI can remix. It can combine ideas. It can imitate styles. What it cannot do is experience the world.
The most memorable campaigns, products, and innovations usually emerge from human insight, lived experience, and understanding of people, not simply pattern prediction.
AI can accelerate creative work. It doesn’t replace imagination.
Poor data
If your organisation’s information is incomplete, inconsistent, or outdated, AI has very little to work with. It cannot invent reliable knowledge that doesn’t exist.
In many cases, improving data quality delivers greater business value than buying another AI tool.
Human relationships
Some problems are not information problems. They’re trust problems.
People don’t build long-term relationships because a chatbot answers quickly.
They stay because someone understood their situation, listened carefully, and responded with empathy.
For customer relationships, fundraising, partnerships, and community engagement, AI can assist. It cannot replace genuine human connection.
Why the same AI tool succeeds in one organisation and fails in another
People often assume success depends on choosing the “best” tool. In reality, success usually depends on everything around the tool.
An organisation with organised data, clear processes, realistic expectations, staff who understand the technology, and leadership willing to improve workflows will often achieve far better results than another organisation without these structures in place using exactly the same software.
Technology rarely fixes organisational problems. It usually reveals them.
A reality check for Nigeria and the wider African market
This challenge deserves far more attention than it receives. Many AI systems are trained predominantly on datasets originating from North America and Europe.
As a result, they often perform less effectively when working with African contexts.
This can show up in subtle but important ways. Nigerian names may be misinterpreted. Local businesses may not be recognised.
Regional expressions and cultural references may be misunderstood. Nigerian Pidgin and indigenous languages are often handled less accurately than English variants with greater representation in training data.
Local market dynamics, regulatory environments, and consumer behaviour can also be overlooked or misunderstood.
These are not minor inconveniences. For organisations operating in Nigeria and across Africa, they directly affect the reliability of AI outputs.
That doesn’t mean AI isn’t useful. It means outputs should always be reviewed by people who understand the local context. The closer your work is to local communities, languages, and culture, the more valuable human oversight becomes.
How to read past the hype
Every AI product promises to save time, increase productivity, and transform the way you work. Some genuinely will. Many won’t.
Before adopting any AI tool, ask yourself five simple questions:
- What specific problem does this solve?
- What evidence shows it works in organisations like ours?
- What information does it need to perform well?
- What happens when it makes a mistake?
- Who will be responsible for managing it six months from now?
If those questions don’t have clear answers, you’re probably looking at good marketing rather than a good solution.
Looking beyond the buzzwords
The organisations seeing the greatest value from AI today are rarely the ones chasing every new release or feature. They’re the ones making deliberate decisions.
They understand where AI adds speed. They recognise where human expertise remains essential. And they know the difference between genuine capability and ambitious marketing.
That balance is what turns AI from an expensive experiment into a practical business tool.
Closing thought
The hype surrounding AI isn’t entirely wrong. Modern AI can save time, improve productivity, and unlock opportunities that would have been difficult to imagine just a few years ago.
But the hype is incomplete. It often leaves out the limitations, the maintenance, the need for good data, and the importance of human judgment.
Understanding both sides of the picture is imperative for smart decision making.
The organisations that benefit most from AI won’t be the ones that adopted it first, they’ll be the ones that understood exactly what they were adopting, and what they weren’t.
___________________________
