AI at Work: How to Prepare Without Losing Critical Thinking

Artificial intelligence is no longer a distant trend.

It has entered the daily routine of anyone who writes emails, analyzes data, structures processes, researches information, creates reports, and makes decisions. In finance, this is even more evident: reconciliation, accounts payable, accounts receivable, month-end close, document analysis, and integration between systems are already being rethought with AI.

But there is an important difference between using AI occasionally and knowing how to apply it with maturity.

The professional who stands out in this new landscape is not necessarily the one who tests every tool of the moment. It is the one who understands their process, knows where AI helps, recognizes its limitations, and maintains critical thinking about what is being automated.

That was one of the main points of the conversation with Gio Mangone, founder of Móglia, on Papo Tech no Financeiro.

AI as a superpower, not a substitute for knowledge

For those who work in business areas such as finance, operations, marketing, sales, or customer service, AI can act as an accelerator.

It helps organize ideas, analyze information, review text, create drafts, structure processes, and explore hypotheses much faster. But the real gain appears when the person already understands at least minimally the context they are working in.

In finance, for example, this is essential.

An analyst can use AI to support document reading, review a spreadsheet, summarize a contract, or suggest an approval flow. But they still need to understand whether that answer makes sense, whether the rule is correct, and whether there is accounting, tax, operational, or governance impact.

AI can speed up the work. But it does not replace the technical foundation of those who need to make good decisions.

Productivity depends on a well-defined process

A common mistake is expecting AI to deliver instant productivity.

In practice, the first attempts may even feel slower. That happens because turning a manual task into an AI-supported process requires clarity about the step-by-step.

Before automating, you need to understand:

• what triggers the process;
• what data goes in;
• which validations must happen;
• what exceptions exist;
• who approves;
• which systems need to be updated;
• what final output should be generated.

When the process lives only in the professional's head, it is much harder to guide AI.

But when the flow is clear, AI becomes a layer of acceleration. It can help execute repetitive steps, interpret information, suggest paths, and reduce operational effort.

In finance, this is especially relevant because many processes still depend on parallel spreadsheets, manual checks, emails, ERPs, banks, and distributed approvals. That is the context in which a well-structured financial workflow makes a difference.

The role of critical thinking in using AI

One of the biggest risks in using AI is accepting any answer as truth.

AI tools can be wrong. They can misinterpret context. They can invent information. They can suggest logic that looks correct but does not hold up when applied to a real process.

That is why critical thinking has become a core skill.

It is not enough to ask better questions. You need to know how to evaluate the answer.

That "feeling" comes from experience, repertoire, and accumulated knowledge in the field. It is what makes a professional look at a response and think: "this doesn't seem right."

In finance, this point is even more sensitive.

A wrong answer can affect payment, reconciliation, accounting classification, compliance, audit, or relationships with suppliers and customers. That is why AI in financial processes must be used with control, validation, and traceability.

Using AI well is not using AI for everything. It is knowing where it truly improves the process without compromising control, quality, and governance.

Not everything needs AI

Another important point: using AI for everything is not maturity.

Many companies and professionals are under pressure to be "AI native," test tools, and show volume of use. But the number of prompts is not a productivity indicator.

The most important question is not "how do we put AI in this process?"

It is: "does it make sense to use AI here?"

Some tasks are better solved with simple rules, traditional integrations, deterministic automation, or solid data structures. In other cases, AI makes a lot of sense—especially when there is context reading, document interpretation, classification, text analysis, assisted decision-making, or interaction with systems without an API.

Maturity is choosing the right tool for the right problem.

What professionals need to learn now

To prepare for this new landscape, you do not need to start with complex courses or try to master every tool on the market.

The first step is to use it.

Test a chat tool. Understand how it responds. Notice where it fails. Learn to explain context better. Compare answers. Ask better questions. Understand its limits.

Then, it is worth going deeper into a few fundamentals:

• what AI models are;
• what LLMs are;
• why AI can be wrong;
• which types of tasks it performs best;
• where it is not reliable;
• how to protect sensitive data;
• when to use generative AI, automation, or traditional integration;
• how to validate answers before taking them into a real process.

It is also important to understand that there are different categories of tools.

A chat tool is not the same as a financial automation platform. An automation platform is not just an AI tool. And a solution with AI agents does not work the same way as an isolated prompt.

The clearer professionals are about these differences, the better their ability to apply AI with real impact.

Checklist: before using AI in a process

• Is the process clear?
• Is the input data reliable?
• Is there a validation criterion?
• Does the answer need to be auditable?
• Is there financial, tax, or operational risk?
• Is AI really necessary?
• Who reviews exceptions?

Human skills remain the differentiator

Despite all technological evolution, human skills remain decisive.

Communication, curiosity, creativity, critical thinking, and problem-solving ability are even more important in a world with AI.

Many people ask how to write better prompts. But in practice, much of the challenge is knowing how to explain what you want—with context, objective, and quality criteria.

That requires communication.

It also requires curiosity to test, fail, adjust, and learn continuously.

And it requires business vision to identify where AI truly creates value.

In finance, that means moving beyond merely executing tasks and starting to think in processes: where there is rework, where there is risk, where there is lack of visibility, where repetitive decisions exist, where the team still depends on parallel controls, and where well-structured automation can free time for analysis.

AI is not going away, but the hype should fade

AI is not a passing fad.

Like the internet, the cloud, and other major technological shifts, it should remain part of work. What is likely to change is how we talk about it.

Today there is still a lot of noise: new tools every day, exaggerated promises, fear of replacement, and pressure to adopt AI quickly.

Over time, the conversation should mature.

Companies will realize that AI is not the answer to everything. Professionals will better understand where it helps. And the most successful processes will be those that combine technology with governance, context, and human validation.

The professional who will stand out

The professional who will stand out in the coming years is not the one who only uses ChatGPT to improve text.

Nor is it the one who tries to automate everything without criteria.

It is the one who can combine three things:

1. solid knowledge of their own field;
2. practical understanding of AI and automation;
3. critical thinking to apply technology responsibly.

In finance, that means using AI to gain productivity without giving up control. It means automating processes while maintaining traceability. It means reducing manual work while preserving governance. It means using technology to free the team from repetitive tasks and increase capacity for analysis.

AI can be a superpower.

But only for those who know where, how, and why to use it.

See how companies are applying this in practice in our customer stories and in the content on AI applied to accounts payable.

FAQ — AI at work and in finance

How can I use AI at work productively?

Start with processes you know well. Define objective, context, and validation criteria before asking AI for an answer. Use the tool to accelerate drafts, analyses, and structuring—and review the result with business knowledge before applying it to any real workflow.

What skills are important for using AI?

Clear communication, critical thinking, curiosity, process vision, and domain knowledge. Knowing how to write a good prompt helps, but the differentiator is evaluating whether the answer makes sense and whether using AI creates value without increasing risk.

Does AI replace finance professionals?

Not broadly. AI replaces repetitive tasks and accelerates analyses, but decisions with financial, tax, and operational impact still require human validation, governance, and contextual understanding.

Why is critical thinking important when using AI?

Because models can be wrong, invent information, or suggest logic that looks correct but fails in practice. In finance, an error can affect payments, reconciliations, compliance, and audits.

When does it make sense to use AI in financial processes?

When there is document interpretation, context-based classification, text reading, assisted decision-making, or interaction with systems without a structured API. If a simple rule or integration solves it, AI may be unnecessary.

What is the difference between using AI and automating processes?

Automating means structuring a flow with rules, integrations, approvals, and traceability. Using AI means applying models to interpret context, handle variation, and support decisions within that flow—or in specific steps of it.

How do you prepare to work with AI?

Use tools in daily work, test limits, learn LLM fundamentals, understand solution categories (chat, automation, agents), and practice validation before scaling any use to critical processes.

---

Want to understand where AI can create real impact in your finance function?

See how Abstra helps finance teams automate processes with AI, integrations, and governance—connecting ERP, banks, spreadsheets, documents, and internal systems in traceable workflows.

Talk to a specialist