GenAI is rapidly transforming business operations worldwide, from the SMB level to the largest enterprises. Conversational GenAI applications—tools that augment existing large language models (LLM) with your organization’s data—can significantly boost the practical value of the data you already possess. In-house LLMs can provide detailed summaries of meetings, personalize marketing content and automate customer service experiences, auto-generate process methodologies, streamline specialized research and document analysis, answer technical questions, enhance content creation, and much more—all within the boundaries of organizational standards and data security concerns. 

The potential benefits of a successful in-house LLM could be huge, but any such project starts with the underlying hardware, whether it is on-prem or in the cloud. GenAI is resource-intensive, and the solution you choose to power your LLM must be able to handle the number of simultaneous users you need to support with the kind of response time they require. To choose the right platform, you’ll have to make decisions about CPUs, GPUs, networking, and memory. Objective real-world performance data is mandatory for IT teams seeking the right platform at the right price. 

To meet the need for objective in-house LLM sizing data, we built PTChatterly—an AI benchmarking and sizing framework that can quantify the performance and experience you can expect while running an in-house-data-augmented LLM on a given underlying solution. PTChatterly includes a full-stack AI implementation of an LLM—augmented with in-house data—and a testing harness that lets you determine how many people the system can support with acceptable response time. The tool reports test results in the form of meaningful metrics that are simple and easy to understand. For example, it might say that the server under test supports 28 people having simultaneous conversations with a complete response time—how long it takes for a full answer to appear—of 10 seconds or less. 

Of course, if you want to dig into the more technical side of things, PTChatterly also provides a wealth of more in-depth results. 

If you're considering building an in-house-augmented LLM and are deciding among multiple possible solutions, PTChatterly can give you the objective, repeatable, real-world data you need to confidently make a smart buying decision. 

To find out more about how PTChatterly can help your organization, check out the PTChatterly home page for detailed answers to FAQs and more or view the introductory video below.