CEOs Are Confusing Two Types of AI
Executives keep hearing about “AI”, but most conversations mix two very different technologies: predictive AI and generative AI.
Most discussions about “AI” combine two different technologies: predictive AI that analyses data and generative AI that creates content.
Predictive AI has operated quietly inside organisations for years improving forecasts, fraud detection and operational decisions.
Generative AI changes how employees interact with information by turning complex systems into conversational interfaces.
The AI Conversation Is Confusing for a Reason
Many executives feel slightly disoriented by the current AI discussion.
Every week brings another announcement:
AI will transform your organisation.
AI will automate jobs.
AI will reshape industries.
But the conversation often becomes confusing because people use the term “AI” to describe two very different types of technology.
Understanding that distinction is the first step to making sensible decisions about where AI will actually matter inside your organisation.
The Two Types of AI
When most people say “AI”, they are referring to one of two broad categories.
1. Predictive AI
Predictive AI has existed inside organisations for years.
It analyses historical data to predict, classify or optimise outcomes.
Typical examples include:
fraud detection in banks
product recommendation engines in retail
demand forecasting in supply chains
predictive maintenance in manufacturing
These systems rely on machine learning, where models analyse large volumes of historical data to detect patterns.
But predictive AI does not create anything new.
It simply produces probabilities.
In practical terms, predictive AI answers questions like:
Will this customer churn?
Is this transaction fraudulent?
What demand should we expect next week?
Predictive AI improves operational efficiency.
But it largely operates in the background.
Most employees never interact with it directly.
2. Generative AI
Generative AI is different.
Instead of analysing data to predict outcomes, it creates new outputs.
Those outputs can include:
written text
software code
images
structured analysis
This is the technology behind tools such as ChatGPT, Claude and Microsoft Copilot.
Employees interact with generative AI directly.
They ask questions.
The system generates answers.
That simple shift is why generative AI has moved from the technology department to the boardroom so quickly.
A Retail Example
Retail companies provide a useful illustration of the difference.
For years, retailers have used predictive AI to forecast demand.
Machine learning models analyse historical sales, weather patterns and promotions to estimate how much stock stores will need.
Those models operate quietly inside supply-chain systems.
Merchants rarely interact with them directly.
Now generative AI is appearing on top of those same systems.
Shopify recently introduced Sidekick, a generative AI assistant for merchants running stores on its platform.
Instead of reviewing dashboards and reports, merchants can ask questions such as:
Why did sales drop this week?
Which products should I reorder?
What inventory is moving fastest?
The system analyses store data and generates an explanation or recommendation.
The predictive models underneath still exist.
But generative AI becomes the interface through which people interact with those systems.
A Banking Example
Banks show the same pattern.
Financial institutions have used predictive AI for years.
Machine learning models analyse transaction patterns to detect fraud, assess credit risk and flag suspicious activity.
These systems run constantly inside banking infrastructure.
Customers and employees rarely see them.
Generative AI introduces a different capability.
Instead of simply detecting patterns, it can summarise complex financial information, explain risks and assist staff in responding to customer queries.
Banks including Commonwealth Bank and JPMorgan have begun experimenting with generative AI tools to support employees analysing documents, preparing responses and interpreting operational data.
The underlying predictive models remain in place.
But generative AI becomes the interpretation layer sitting on top of those systems.
Why This Difference Matters
Predictive AI improves decisions.
Generative AI changes how work happens.
Predictive AI answers questions like:
What will happen?
Generative AI answers questions like:
Help me do this.
That distinction explains why generative AI is spreading so quickly inside organisations.
It operates at the interface between employees and information.
Instead of navigating dashboards, databases and reporting systems, employees can simply ask questions.
The Technology Behind Generative AI (In Plain English)
Modern generative AI systems are built using very large neural networks known as large language models.
These models are trained on enormous datasets and learn patterns in language, images and code.
When a user asks a question, the system generates a response based on those patterns.
The technical details matter for engineers.
For executives, the key point is simpler:
Generative AI makes complex information systems conversational.
The Strategic Reality
For most organisations, the distinction is straightforward.
Predictive AI improves operational decisions.
Generative AI changes how employees interact with knowledge.
Both matter.
But they affect organisations in different ways.
Predictive AI tends to sit inside operational systems.
Generative AI sits at the interface between employees and those systems.
The Leadership Question
When leaders say they are “investing in AI”, they should be clear about which type they mean.
Are they improving operational models that predict outcomes?
Or are they deploying systems that change how employees work with information?
Those are very different investments.
And they lead to very different organisational changes.