LIEV accelerates the use of AI at Inmetrics using AWS

Inmetrics developed LIEV, an internal initiative to accelerate its processes and increase team productivity. The objective was to reduce manual workload and allow teams to focus on their tasks with greater added value. The initiative was so successful that the company decided to make the solution available to its clients, allowing them to leverage the same efficiency and productivity benefits.

Daniel Messias, solution director at Inmetrics, recalls that the company was already developing algorithm models for operating data and that the arrival of large language model (LLM) networks encouraged new development possibilities. One of these possibilities resulted in the creation of a generative AI solution that works as an accelerator in the development of AI solutions, boosting the efficiency and productivity of various operations within a business—with intensive use of open-source LLM networks, strict security standards, and in full compliance with the Brazilian Data Protection Law (LGPD) requirements.

Referred to as LIEV, the solution represents a leap into generative AI, combining efficiency, versatility, and safety. With LIEV, Inmetrics is positioned to lead and define the future of generative AI, offering innovation and performance as real and long-term competitive advantages to its clients.

LIEV works as a set of codes and architectures to offer business solutions or the generation of technical artifacts, facilitating the use of LLMs in private environments and running on the client's own cloud—this is why the data is never shared. In order to guarantee that the solution was successfully implemented, Inmetrics needed to rely on an adequate infrastructure. That’s where AWS comes into play, providing an infrastructure that enables the use and adoption of generative AI at scale.

“AWS has always been our cloud provider and we chose it for being the most mature cloud in the market. That's why LIEV was born within AWS, as a result of this trajectory between the two companies,” Messias says, recalling that AWS infrastructure helped the company to train models quickly and efficiently, creating the ideal solutions for Inmetrics’ clients with the lowest investment possible.

Today the entire LIEV infrastructure runs on AWS, using services such as Amazon Elastic Compute Cloud (Amazon EC2); Amazon SageMaker; and Amazon Bedrock, a managed service that that offers a choice of high-performing foundation models (FMs) from leading AI companies and Amazon with a single API, along with a broad set of capabilities to build generative AI applications, simplifying development while maintaining privacy and security. Amazon Bedrock agents are fully managed and facilitate the building of generative AI applications that can provide updated answers based on proprietary knowledge sources and complete tasks for a wide variety of use cases.

In addition to Inmetrics' internal use, AWS services are also the basis for offering LIEV to the market. As an AWS Partner, Inmetrics also implements LIEV on AWS for other clients. Therefore, since its creation, the solution was integrated with all Inmetrics’ services, making it the first Brazilian technological consultancy to incorporate AI across its portfolio, including development, software testing, modernization and migration of applications to the cloud, and the construction of production belts.

LIEV is already being operated in pilot version by some of Inmetrics' central clients. They have been using it to build personalized solutions oriented to solving their key business challenges and impacting different success indicators. "From the moment we released this engine in the client's environment, we started leveraging AWS services in order to assimilate and promote this stream. In that sense, this is a great value that we have,” Messias says.

Sidney Rocha, another managing director at Inmetrics, recalls that, in the first months using the solution, the company already worked on 20 generative AI projects. “Today we don't talk exclusively about digital transformation, we discuss digital efficiency. That’s what the use of generative AI is all about. Making this connection marks a new moment in the technology market,” he explains. Rocha also describes the steps companies should take to be able to capture all this value. Created based on the experience gained, the company relies on a roadmap based upon six steps: identify a specific domain and select a persona; define OKRs and assemble the team; determine data sources; define the technology stack; experiment with users and measure results; create a business case.