| Predictive | Generative |
|---|---|
| What it does | Assigns a number or a label to something that exists |
| Examples | Fraud, credit risk, demand, machine failure |
| Ground truth | Arrives eventually, so accuracy is measurable |
| Hardware | Modest, often an organization's own machines |
Ground truth is the whole difference
A fraud model flags a transaction. Weeks later the record shows whether it was fraud, and the model has been scored. That loop is what makes a predictive system testable: you can measure its accuracy, watch it drift, and compare it with whatever it replaced.
A generative model drafts a paragraph. No record will later say what the right paragraph was. "Accuracy" is frequently not even well defined, so evaluation becomes judgment: did a reviewer accept it, did it state anything false, did it break a rule.
Why the infrastructure differs
Predictive systems run on modest hardware and have been in production in banking and insurance for years, under existing regulation. Frontier generative models are what require hyperscale data centers, and most public argument about AI's energy and cost concerns them, usually without saying so. More in Does all AI need a hyperscale data center?
Where the measurable return sits
Predictive systems already do bounded work such as fraud detection, forecasting, underwriting, and maintenance, where later outcomes can measure performance. Generative deployment is also growing, but its returns are often measured through review, acceptance, or time saved without a single ground-truth outcome.
Where returns are measurable, they concentrate in predictive systems doing bounded, repetitive work with a ground truth to check against. Generative deployment is real and growing, and its returns are harder to attribute. That is a measurement problem, and it is no evidence of failure.
What this means for an organization
- Ask which kind a proposal is. A predictive project should arrive with a baseline and an accuracy target, because the answer will come in to check it. A generative one needs a different measure, such as time saved or the share of drafts accepted.
- Expect different evidence from vendors. For a predictive system, ask for accuracy measured against outcomes. For a generative one, ask how outputs are reviewed and what gets recorded.
- Give generative work a fair clock. Gains spread thinly across many people are real and slow to show up in a six-month payback.
For the generative prompts a team sends to supported AI tools, Sentinel, an AI gateway, measures each request and shows it in the account, which gives usage to review. It records metadata only and does not judge whether an output was right.
Frequently asked questions
- What do predictive and generative AI each do?
- Predictive AI assigns a number or label to something that already exists, such as fraud, credit risk, demand, or machine failure. Generative AI produces new text, images, code, audio, or video. The real answer eventually arrives for a predictive system, so its accuracy can be measured, while a generative output usually has no answer to check.
- Why is generative AI harder to test?
- A generative model drafts a paragraph, and no record will later say what the right paragraph was. Accuracy is frequently not even well defined, so evaluation becomes judgment: did a reviewer accept it, did it state anything false, did it break a rule. Predictive systems get scored when the outcome arrives.
- Do both kinds of AI need the same hardware?
- No. Predictive systems run on modest hardware, often an organization's own machines, and have been in production in banking and insurance for years. Frontier generative models are what require hyperscale data centers, and most public argument about AI's energy and cost concerns them, usually without saying so.
- Where is the measurable return from AI?
- Where returns are measurable, they concentrate in predictive systems doing bounded, repetitive work with a ground truth to check against, such as fraud detection, forecasting, underwriting, and maintenance. Generative deployment is real and growing, and its returns are harder to attribute. That is a measurement problem, and it is no evidence of failure.
- What should you ask when someone proposes an AI project?
- Ask which kind it is. A predictive project should arrive with a baseline and an accuracy target, because the answer will come in to check it. A generative one needs a different measure, such as time saved or the share of drafts accepted. Ask vendors how outputs are reviewed and what gets recorded.