AI accelerates answers; who makes the decisions?

Column

Project Management

September 25, 2026

AI accelerates answers; who makes the decisions?

The speed of technology only becomes an advantage when it is accompanied by the capacity to understand, verify, and assume responsibility for the results

Imagine a team that starts producing in an afternoon the analyses that previously required a week. The change seems to justify the investment in artificial intelligence. But, before accounting for the gain, it is worth asking who examined the results, what information supports the conclusions, and how much time will be needed to use them safely. If production accelerates and evaluation remains limited, part of the work has merely changed places.

One of the most important issues in AI adoption is what happens when the capacity to produce answers grows faster than the capacity to verify them. For managers, this difference can appear in reports, projects, forecasts, and recommendations. The document is ready, but the decision still requires someone capable of vouching for it.

A recent debate on the Navier-Stokes equations helps to understand this situation. In September, OpenAI announced a proposed solution to a mathematical problem related to the equations that describe fluid motion. According to the company, approximately 10,000 AI agents participated in the effort that produced the demonstration in 88 hours. The announcement indicated a relevant computational research capability, but also raised questions about the origin of the contributions and their recognition (OpenAI, 2026).

For organizations, the most useful point of this episode lies in the distance between obtaining a result and transforming it into reliable knowledge. This distance needs to be factored into productivity evaluation. A team can produce more and still accumulate conclusions that no one has had the opportunity to examine adequately. The gain is only complete when the result can guide a decision.

Productivity needs to include verification

In management, it is usually easier to measure execution time than the quality of judgment. It is possible to record how many reports were delivered and how many hours were saved. Assessing whether the team identified an inadequate hypothesis, avoided a hasty interpretation, or realized that data was missing requires more careful monitoring. With AI, it is important to give more visibility to this second type of work.

Consider a demand analysis that recommends reducing the stock of a certain product. The system can organize the data and present a convincing justification. Before acting, however, someone needs to check if the history includes a stockout: low sales may reflect a lack of merchandise, not a lack of customers. The decision depends on knowing the business and examining the meaning of the numbers.

This example shows why the person responsible for validation should participate from the definition of the task. They need to know what will be evaluated, what data is available, and what evidence would justify accepting the conclusion. Leaving this responsibility until the end favors a rushed review, especially when the presentation already seems complete and the deadline has been consumed by the expectation of delivery.

The verification effort should also track the consequence of the decision. A draft internal communication might undergo a simple review. A recommendation that alters an engineering design or commits resources requires testing, documentation, and expert evaluation. Applying the same procedure to everything wastes time in some situations and leaves others insufficiently examined.

Terence Tao offers a useful contribution to this discussion by distinguishing the production of proofs from their verification, explanation, and incorporation into collective knowledge. His argument helps to realize that accelerating one step can increase the demand on subsequent steps (Tao, 2026). In companies, this reflection recommends evaluating the complete process: time to decision, rework, identified errors, and ability to explain the conclusion.

It is in this sense that a demanding adoption of AI is understood as necessary. The tool can take on extensive tasks, explore alternatives, and help confer results. It is up to the organization to create conditions to leverage this capability, including time and people to evaluate what has been produced. The verification must be foreseen in the project and budget.

Knowledge is also in the journey

The Navier-Stokes case adds another dimension. Mathematician Tristan Buckmaster, who worked on problems related to Levent Alpöge, reported using AI tools and inputting drafts into Codex. He questioned a possible influence of these interactions on OpenAI’s result, although he stated he did not know if his data had been used (Buckmaster, 2026).

The company responded that an internal investigation had ruled out the influence of Buckmaster’s prompts from the previous two months, including through training. This response should accompany the presentation of the suspicion, preserving the difference between an allegation and the conclusion disclosed by the organization involved (OpenAI, 2026).

Regardless of the outcome of this controversy, I see a practical issue for knowledge management. A research or a project holds value before the final delivery. Hypotheses, failed tests, choice criteria, and discarded solutions reveal what the team has learned. When this material goes through external tools, the organization needs to decide what information can be used, under what conditions, and how the path will be recorded.

Authorship must also follow this path. In an AI-assisted project, someone formulates the problem, selects information, establishes constraints, and interprets results. These contributions need to remain visible. Simply recording who requested the last response can erase an important part of the team’s work and make it difficult to identify responsibilities.

Consider a system developed with AI support that works during the presentation but fails after deployment. To fix it, the team needs to know the design decisions, the test conditions, and the known limits. A history of instructions can help, but it needs to be accompanied by documentation that allows another professional to understand and continue the work.

Therefore, a delivery must include the necessary explanation for its continuity. This practice protects the organization’s investment and reduces dependence on a person or tool. Time savings in execution lose part of their value if maintenance requires rebuilding everything that was done.

Educate people to assume responsibility

As a Computer Engineering professor, I consider the relationship between outcome and learning especially important. An activity may be completed without the person who submitted it being able to explain the choices made. Education needs to create opportunities to identify this difference, because the professional will be called upon to make decisions in situations that were not foreseen in the exercise.

This requires reviewing what is asked and what is evaluated. In addition to presenting a solution, the student may be asked to justify their hypotheses, compare alternatives, and explain under what conditions they would change their decision. The use of AI can integrate this process, provided that its participation is explicit and the work allows for the evaluation of the reasoning developed.

In companies, the same concern should guide the training of beginner professionals. If all analysis tasks are delegated to the tool, it will be necessary to preserve other opportunities to learn to recognize problems. The review capability depends on technical knowledge and contact with the consequences of choices. It is not enough to appoint a reviewer without providing the conditions for them to develop this judgment.

For Brazilian universities and companies, there is room for cooperation in projects that combine the use of AI and validation in real-world situations. Working with a defined problem, pre-defining success criteria, and comparing results with observed data can yield more useful gains than adopting tools solely based on the hype of their announcements. It also allows for discussions on confidentiality, credit, and responsibility before conflicts arise.

The potential of technology is significant, and harnessing it requires changes in work organization. Professionals capable of formulating good questions, recognizing limitations, and explaining decisions will play a central role in this process. Their contribution needs to be valued even when it manifests as a less optimistic conclusion, a deadline revision, or the decision to perform another test.

The Navier-Stokes case helps put that choice in perspective. The ability to produce results is advancing, and institutions need to develop ways to incorporate them with confidence. For a manager, a question should accompany every promise of acceleration: who is able to explain and assume the decision that will come next? The answer indicates how much of the contracted technology is being transformed into the organization’s capability.

To access the references of this text click here

Who wrote this column

Maurício Acconcia Dias

Possui graduação em Ciência da Computação pela Universidade Federal de Lavras, mestrado e doutorado em Ciências da Computação e Matemática Computacional pela Universidade de São Paulo e MBA em Data Science Analytics pela USP/Esalq. Atua com desenvolvimento de hardware para sistemas inteligentes aplicados à robótica. É consultor em Data Science & Analytics e Desenvolvimento de sistemas embarcados e orientador do MBA USP/Esalq.

You may also like

November 16, 2026

Exploratory modeling of the effect of preemptive cytomegalovirus treatment in AIDS patients

Cytomegalovirus (CMV) is highly prevalent and causes severe disease in immunocompromised individuals, such as HIV patients. Given the scarcity of robust data on preemptive treatment for this infection, an exploratory model was developed. The objective was to identify potential subgroups that would benefit from this approach. A retrospective observational study was conducted, analyzing 176 medical records of patients admitted to a tertiary hospital between 2017 and 2022. After applying inclusion and exclusion criteria, 116 patients were divided into a preemptive treatment group (19) and a control group (97). Machine learning techniques, such as LinearDML and CausalForestDML, and variable analysis by Feature Importance and SHAP were used. Variable analysis indicated that high CMV viremia was associated with death, with an Odds Ratio of 1.85 (p = 0.004). The model estimated an increase in overall mortality associated with preemptive treatment (ATE of 48%, p = 0.0180), possibly reflecting the greater baseline severity of the treated patients. However, heterogeneity analysis identified that older patients showed a relatively more favorable estimated effect. Limitations included the small sample size and the retrospective and observational nature. Nevertheless, with an expanded evaluation group, the model could serve as a basis for building clinical decision support tools.

Keywords: Machine learning; Double machine learning; Causal inference; Mortality.

November 16, 2026

Regression models for personal hygiene and personal care product expenses in the Southeast region

Regression models were evaluated to correlate expenses with personal hygiene and care products of the population in the Southeast Region of Brazil. Data from the 2017-2018 Family Budgets Survey (POF), conducted by the Brazilian Institute of Geography and Statistics, were used. The models were estimated by ordinary least squares and verified for the assumptions of normality of error terms, multicollinearity, and heteroscedasticity. Linear regression, nonlinear regression with Box-Cox transformation, and linear regression with dummy variables were evaluated. Coefficients statistically different from zero were obtained for all three models, confirmed by the p-values of the F and t tests, and the adherence of residuals to normality was graphically assessed. However, heteroscedasticity problems were observed in all cases, indicating the omission of relevant variables from the model. The low R² values obtained, between 2.19% and 5.56%, suggested that the dependent variable was poorly explained by the models. Despite this, a growing trend in expenses for personal hygiene and care as a function of disposable income per capita was observed. For future studies, it is recommended to evaluate other explanatory variables to improve the model’s adjustment coefficient and eliminate heteroscedasticity.

Keywords: Family expenses; Ordinary least squares; Per capita income; Box-Cox transformation; Dummy variables.

November 16, 2026

Cognitive Accessibility and ASD: Design Guidelines for Augmentative and Alternative Communication Applications

Autism Spectrum Disorder (ASD) imposes specific challenges on the design of digital interfaces, affecting cognitive processing, sensory regulation, and functional communication. In this context, Augmentative and Alternative Communication (AAC) is an essential resource to support individuals with oral language limitations. This study aimed to analyze the usability and accessibility of AAC applications for the neurodivergent public. The research was characterized as descriptive, with a quantitative approach, and was based on the systematic observation of the interfaces of the Brazilian applications Livox, Matraquinha, and Expressia. A structured checklist, developed from the convergence of Human-Computer Interaction (HCI) guidelines, Applied Behavior Analysis (ABA), and the GAIA Guide, was used as a data collection instrument. The results revealed that Livox and Expressia showed greater technical compliance and learning support, while Matraquinha stood out for its semantic literalness, but with limitations in customization and architecture. A generalized critical gap was identified in the proactive detection of frustration, highlighting opportunities for strategic innovation for the digital business market and for the advancement of neuroinclusive design.

Keywords: Neuroinclusive Design; Digital Inclusion; Strategic Innovation; Assistive Technology; Usability.

November 16, 2026

Multidimensional Poverty and Bolsa Família Program: challenges and perspectives of inclusion through public policies

The concept of poverty has evolved over the 20th and 21st centuries, incorporating a more inclusive perspective that transcends monetary income. The Multidimensional Poverty Index (MPI) has emerged as a complementary tool to characterize poverty and guide programs like Bolsa Família, aiming to improve their effectiveness and targeting. The study aimed to create a national MPI using data from the Continuous National Household Sample Survey (PNADC) of 2024, seeking to identify the correlation between receiving the benefit and multidimensional poverty, as well as to detect underserved populations. The Alkire-Foster methodology was used to calculate the MPI, decomposing the results into groups of beneficiaries and non-beneficiaries of PBF. Subsequently, a Generalized Linear Model with a logit function was applied to analyze the correlations. The results revealed that 4.84% of the Brazilian population lived in multidimensional poverty, with an incidence three times higher among Bolsa Família beneficiaries. Two profiles of multidimensional poverty and two potential blind spots in PBF coverage were identified, including 9.0% of non-beneficiary multidimensional poor in extreme monetary poverty and 25.2% in low-income vulnerability. Being a PBF beneficiary was associated with a two-fold higher chance of being classified as multidimensionally poor. It was concluded that the MPI is a valuable tool for designing public policies, increasing targeting capacity and demanding actions to address structural problems that perpetuate poverty.

Keywords: Alkire-Foster; MPI; Public policies.

November 16, 2026

The influence of People Management on the Health and Well-being of collaborators

In a dynamic organizational scenario, understanding the influence of People Management practices on employee health and well-being has become essential. The study aimed to analyze how People Management practices influence employees’ perception of health, well-being, and quality of life in various organizations. The methodology adopted was descriptive, with a quantitative approach, and employed the survey technique. Data were collected through a structured questionnaire with 15 sociodemographic questions and Likert scale items, disseminated for 40 days on social media. The sample consisted of employees from different companies, selected for convenience, with voluntary and anonymous participation. The results revealed that people management practices impacted motivation, engagement, and the perception of quality of work life, but were not consistently perceived by all employees. It was observed that organizational actions focused on health and well-being were recognized in a moderate and non-equivalent manner, with the presence of neutral and dissenting perceptions. It was concluded that, although People Management practices have strategic potential, their effectiveness depends on greater structuring, continuity, and communication. The study contributed to highlighting gaps between the adopted practices and employee perception, providing subsidies for the improvement of organizational strategies and for the production of academic knowledge in the area.

Keywords: Organizational well-being; Organizational climate; Engagement; Motivation; Quality of work life.

November 16, 2026

Optimization of Delivery Logistics Network through Mathematical Modeling and Clustering

Efficiency in distribution logistics constitutes an essential competitive advantage for the reduction of operational costs and the financial sustainability of contemporary organizations. The general objective of this work was to optimize the delivery network of a company, aiming at cost reduction through the decrease of mileage traveled. The methodology employed was based on the principles of Operational Research and was structured in three sequential phases: geographical allocation of clients to depots respecting Federal Unit restrictions, clustering of stores using the K-Means algorithm, and routing of the heterogeneous fleet using the mathematical formulation of the Vehicle Routing Problem. The main results obtained demonstrated a reduction in the total distance traveled, which represented an absolute saving of 5,116 km and a percentage gain of 3.24%. The model validated itself as a robust tool, capable of optimizing vehicle occupancy and eliminating subjectivity in logistical planning. It was concluded that the use of advanced mathematical modeling allowed operational restrictions to be converted into a measurable competitive advantage. The solution validated the viability of data-based decision support systems for the automation and optimization of large-scale logistical flows.

Keywords: Clustering; Logistics optimization; Operational Research; Cost Reduction; Vehicle routing.