Anthropology and artificial intelligence used to transform corporate listening

Tool allows issues such as harassment to be identified before they become critical, enabling companies to take preventive action.

By Roseli Andrion  |  Agência FAPESP – Human relations are essential in any corporate environment. They are what make companies successful. That’s why listening to people can be an organization’s main differentiator.

That’s what the startup Tech Viz does: it allows companies to listen to the professionals and customers who engage with them. Founded by anthropologist Adla Viana in São Carlos, in the state of São Paulo, Brazil, the company has developed Antropos.ai, an artificial intelligence agent that uses anthropological knowledge to capture and analyze narratives from people on the front lines of organizations – both employees and customers – to identify problems before they become critical. “Anthropology can be used by companies for a number of solutions,” she points out.

Adla’s initial goal was to develop a data platform focused on diversity and inclusion. With support from FAPESP’s Innovative Research in Small Businesses program (PIPE), the researcher participated in PIPE Entrepreneur, as the High-Tech Entrepreneurship Training Program is known. “As researchers, we have little experience with business tools. This opportunity showed us how to go to the market and evaluate our research hypotheses.”

During PIPE Entrepreneur, the startup was able to adjust its goals. “The program takes the academic researcher out of the lab and into the marketplace,” Adla summarizes. “This transition is especially important at a time when many startups are struggling to survive, and investments in research and innovation need to produce concrete results.”

After listening to some 120 stakeholders and realizing that the market was looking for something more comprehensive, Adla decided to create a tool that would enable effective listening to people on the front lines. “When it comes to employees in a company, it’s on the front lines of operations, where minorities are often concentrated, that people have the information that the manager and leadership need to understand what the problems and solutions are,” says Adla. “When a problem is solved in a large company, those on the front line say: ‘Wow, finally.’ Often, customers feel the same way: they suffer for a long time with some difficulty until it’s solved.”

Adla says that the team realized that managers wanted a tool that was informed by anthropological knowledge to know what people are really experiencing and to really listen to what they have to say. Tech Viz then created an artificial intelligence agent that uses scientific research to meet market needs. “We control the information to train the models and ensure that the qualitative information is useful in solving problems according to the customer’s requirements.”

Anthropological basis

Tech Viz’s solution differs from organizational climate surveys and market research in two key ways: its anthropological basis and its use of narratives rather than conventional forms. “Many front-line employees don’t necessarily read and write as part of their daily work. Forms with ‘strongly agree’ or ‘strongly disagree’ may not be appropriate for this audience. Also, it’s common for these forms to be very individualized, and the idea behind Antropos.ai is to evaluate the information from an anthropological point of view to approach relationships within companies.”

According to the researcher, the tool can identify difficulties before they become critical. “Before harassment, for example, there may be discomfort. And what precedes that discomfort is a woman who can still talk about it,” she describes. “Our tool captures information that precedes very serious situations,” Adla explains.

She says that the solution increases the feedback rate. “It’s very easy to use: all the participant needs to do is send two or three audios, so we have a good amount of information to take to managers in real time, allowing them to make decisions,” she explains. “The solution reduces informational stress while guaranteeing, by capturing information, the personalization of the narratives and the extraction of micro-realities.”

Early detection allows companies to take preventive action, reducing costs and avoiding other consequences. “You don’t have to intervene across the entire company. You find the focus that precedes harassment, for example, and act to mediate the conflict and seek solutions without embarrassing anyone.”

The startup has already tested the platform with companies in different industries. At an energy company, for example, the tool helped identify stories from electricians about the importance of communication in preventing workplace accidents. In another case, it identified situations where women felt uncomfortable in male-dominated environments, allowing preventive action to be taken before harassment occurred.

The system is also effective in situations where traditional reporting channels don’t work well. “People don’t report things to complaint boxes. When the ombudsman gets a call, the conflict already exists,” Adla explains. “The artificial intelligence agent points out possible complaints to anticipate the problem because it looks at what’s happening, not something that’s already happened.”

Artificial intelligence

Adla emphasizes that the use of artificial intelligence requires the right specialists. “Antropos.ai shows that artificial intelligence needs to be built, used and applied by people with experience in scientific knowledge related to the humanities, communication and operational information,” she describes. “This means involving anthropologists, social scientists and psychologists.”

In addition, the quality of the input data is critical to the tool’s success. “If you have good inputs, the performance of the artificial intelligence is excellent,” explains Adla. “Because the solution is anthropologically based, it’s possible to get high-quality inputs and then process them with artificial intelligence algorithms.”

The tool uses natural language processing technologies and artificial intelligence models to analyze narratives and identify patterns and trends that may indicate potential challenges. During its development, Antropos.ai received support from the IPT, the Institute for Technological Research associated with the São Paulo State Department for Economic Development, in the AI Factory program.

Human sciences

According to Adla, investment in scientific research in the humanities is crucial in the context of artificial intelligence. “Engineers alone won’t be able to handle this technology,” she points out. “We need investment in the humanities so that we can really take advantage of all that it can offer,” she argues. “The current technological revolution requires a combination of knowledge from all fields.”

She points out that artificial intelligence is changing the way we access knowledge, transmit information and build large data conglomerates. That’s why it needs to be used ethically and carefully, even as the technology reaches its maximum capacity. “Experiments around the world have already shown that artificial intelligence won’t – and can’t – do anything on its own. Antropos.ai is an example of how to build a responsible solution that guarantees ethics by choosing the most appropriate algorithms.”

The researcher points out that investment in scientific research leads to innovations that benefit the market and society. “At a time when there’s a tendency to devalue scientific knowledge and an exaggerated belief in the capabilities of artificial intelligence, we need to remember the importance of uniting technology and the humanities. Combining different fields of knowledge to create more ethical and efficient solutions is a way to create horizons for other possible futures.”

Born in the innovation ecosystem of São Carlos, Tech Viz received support from FAPESP from the beginning. “Without this support, the company wouldn’t exist. I’m a woman, black and from the northeast of Bahia: the Foundation has opened doors that are usually closed to entrepreneurs with this profile.” Antropos.ai’s methodology is unprecedented. “We’re pioneers in combining anthropological knowledge with technology tools for this exponential active listening. In competing solutions, the underlying science is organizational or behavioral psychology with a focus on the individual.”

In the future, Adla plans to return to academia to pursue a Ph.D. and study the anthropology of algorithms and artificial intelligence. This will allow her to combine her practical experience in the market with academic research. “This is a task that we already do very well, but we need to look at the world and the technological revolution from an anthropological point of view.”

* Image by geralt on Pixabay