I always get this question.

So I decided to create a mini-app to explain my thoughts. You can access it here:

And also an article ↓

Why GenAI Doesn’t Give You the Same Answer Twice

And Why That’s Exactly the Point

Imagine this scenario. You have two interns. Let’s say they both went to the same university in this case let’s say Oxford University.

They’ve completed the same Bachelor’s degree. The same Master’s programme. They’ve had identical training.

You sit them down and hand each of them the same brief: perform a variance analysis on this dataset.

You give them the same instructions, the same data, and the same deadline — two days.

“Would you expect both interns to produce 100% identical analysis and output?”

Of course not. And no reasonable leader would.

Intern A might start with a revenue breakdown, use waterfall charts, and focus on cost-centre gaps across twelve slides. Intern B might lead with margin analysis, use heat maps, and zero in on seasonal patterns across eight detailed slides. Both are valid. Both are insightful. Neither is wrong.

Same training. Same data. Same instructions. Different — yet both valuable — outputs.

Now Replace Interns with AI

Generative AI works the same way. You can give the same model the same prompt and the same data, and the output may vary each time. This is not a bug.

This is how intelligence — artificial or human — works.

Deterministic vs. Probabilistic

Traditional software is deterministic. One plus one always equals two. Every time. No variation. That’s what calculators and spreadsheets are for.

Generative AI is probabilistic. It reasons. It explores. It discovers. Like a brilliant analyst, it weighs multiple paths and may choose a different one each time — because it’s thinking, not just computing.

DETERMINISTIC

Traditional Software

Input → Fixed Output

PROBABILISTIC

Generative AI

Input → Rich Possibilities

GenAI is probabilistic, not deterministic — and that is its power.

Why This Is a Feature, Not a Flaw

If you wanted a calculator, you’d use a spreadsheet. You chose AI because you wanted intelligence — and intelligence is inherently creative.

“Why doesn’t it give the same answer twice?”

Because neither would two Oxford graduates. Variability in intelligent analysis is a sign of depth, not a defect. Each run can surface different valid perspectives on the same data.

“Doesn’t that make it unreliable?”

A calculator is reliable. An analyst is valuable. GenAI is the analyst — it brings judgement, pattern recognition, and creative problem-solving to your data. Reliability and value are not the same thing.

“How should we think about it instead?”

Think of each AI run as a fresh expert perspective. Run it three times and you don’t get errors — you get a richer, multi-angle understanding of your data.

The Strategic Advantage of GenAI for Finance and FP&A

The probabilistic nature of GenAI is not something to fight against. It’s something to leverage.

Here’s what it unlocks for finance teams:

• Uncover hidden trends. AI explores analytical paths you wouldn’t think to take — surfacing patterns invisible to conventional analysis.

• Multi-angle perspective. Each run is a fresh lens on your data. Multiple perspectives mean more robust decision-making.

• Speed of exploration. What takes an analyst days, AI does in seconds — and then does it again from a completely different angle.

• Financial storytelling. With visual outputs generated in seconds, you can focus your expertise on interpreting the numbers, telling the story, and driving strategy — rather than waiting for dashboards to be built.

The question isn’t “why isn’t it the same?” It’s “what did it find this time?”

GenAI can go to places in your data that you didn’t even think to explore. It can uncover trends, run analyses from angles you hadn’t considered, and come back with insights that change the conversation. That’s not unreliability. That’s intelligence.

The next time someone asks you why the AI gave a different output, ask them this: would you expect two Oxford-trained analysts to deliver the exact same work?

Embrace the variability. That’s where the value lives.

Continue Learning

🚀 Join the AI Finance Accelerator — Accelerate your AI + Finance journey

🤝 Join the AI Finance Club — Connect with finance professionals using AI

▶ Watch AI for Finance guides on YouTube​

About Christian Martinez

Christian Martinez is a Finance Senior Manager and AI for Finance Professor.

He has trained over 20,000 CFOs, Fractional CFOs and other senior finance leaders in AI for Finance through his online courses across AI Finance Club, AI Finance Accelerator , LinkedIn Learning and YouTube.

He is also the author of Smart Finance: Leveraging AI for Enhanced Financial Planning and Analysis and an international conference speaker, having presented at the World Finance Forum, World Summit AI, and the EMEA FP&A Summit.

He has over 10 years of experience in FP&A, Finance Transformation, and Finance Analytics at multinational companies across The Netherlands, Australia, and Mexico. He leads a team of 40+ developers and data engineers building advanced analytics tools for finance professionals across 20+ markets worldwide.

Christian holds a Master’s Degree in Finance and Data Science, Lean Six Sigma Green Belt, and Product Owner certifications.