Bain in talks on generative AI with 120 South American firms

Consultancy firm Bain & Company, OpenAI’s preferential partner, has had talks with around 120 enterprises in the region since the beginning of the year for projects applying generative AI in business processes, Lucas Brossi (in photo), the consultancy's head of advanced analytics for South America, told BNamericas.

Currently, Bain has two projects underway in the region. 

Another 10 are at an advanced stage for contract closing, according to Brossi. For contractual reasons, he said the names of these companies cannot be disclosed.

Globally, Bain & Company reports over 20 enterprise use cases created to date over OpenAI, said Brossi. One of them is with Coca-Cola, which Bain claims to be the first company to combine GPT-4 and Dall-E for an AI-driven content creation platform.

The consultancy has been OpenAI's partner since April 2022, applying the company’s ChatGPT, Dall-E and Codex AI systems in its own internal processes. 

In February this year, the two formalized a global alliance, under which Bain supports other enterprises in developing use cases and applying the same technologies.

According to Brossi, interest in the region has mainly come from banks, insurance companies, retailers and companies in the education and health sectors. Talks are underway mostly in Brazil, followed by Chile, Colombia, Peru and Argentina. 

As reported by BNamericas, Bradesco, one of the largest retail banks in Brazil, started testing generative, natural language AI processes for customer service and open finance operations. 

USE CASES

Bain's alliance with OpenAI is not for the open version of the latter’s systems, those that are made available for end-users to test and feed.

"ChatGPT is eminently B2C. In this sense, part of OpenAI's interest was to open its use to individuals to, in exchange, improve the platform by storing data from conversations and from feedback, making the model better," said Grossi.

"When we talk about building applications using OpenAI with customers, it's something else. It's about creating APIs dedicated to supporting the company's processes and all the data transacted there is completely proprietary."

On its website, Bain claims to be able to "separate the hype from the real-world application, bringing experience across the value chain and a deep understanding of our clients’ industries."

There are five common use cases for generative AI in business, according to Brossi. 

The first is customer service and support – traditional chatbots but beefed up by the more powerful generative AI capabilities.

A second one involves sales consulting, with the tools helping the end-customer during their e-commerce journey, said Brossi.

A third interest of companies has been in marketing, specifically the use of text and audiovisual AI tools to create campaigns in different channels.

A fourth application is in IT itself, with generative AI supporting programmers to develop, enhance and restore systems and pages.

“The GPT came as a first layer of incident resolution. For example, if the company's website is down. Traditionally, this company's IT operations team would query logs, test hypotheses, etc,” said Brossi.

“GPT can come in and say, ‘I think the service that went down was this one. I've looked at the logs, consulted the systems, and it looks like you should do such and such a thing. Go there, change that parameter and your problem might be solved’.”

Finally, a fifth use case is as “knowledge assistants,” with generative AI helping employees to have quick access to all company policies, rules and compliance to perform different tasks.

Brossi cites as an example a technician who needs to carry out maintenance on a high-risk plant and is able to instantly consult safety requirements and rules related to that particular asset.

REGULATION

Brossi believes some regulation for the technology is necessary. This is no different than what the CEO of OpenAI, Sam Altman, has himself publicly voiced.  

For Brossi, however, such a regulation should center only on issues related to intellectual property, copyright and system transparency, keeping the tech unleashed for further developments.

He defended that companies employing AI must provide a disclaimer to their end customers about what is machine support and what is human. He also cites concern about biases and prejudices potentially reproduced by platforms.

“But I don't share this doom and gloom, sci-fi view that AI is a risk to humanity and that we are all close to the general AI system."

Out of 800 Bain employees in South America, 100 are currently dedicated to AI. The company has, at present, 20 open positions for the area, according to the executive.

Overall, Bain projects two-digit revenue growth in the region this year.