How A Generative AI Experiment Led to Improved Employee Training for Liberate Labs

Liberate Labs

Austin Texas, USA

Challenge: How Can We Prove Our Generative AI Capabilities?

Liberate Labs was a newly assembled team in January of 2023. As an experiment within the company, the whole team decided to work on a Generative AI solution. This would require a lot of unity as a new team, and the challenge was exciting.

Specifically, CIO Jawad Iqbal wanted to do this experiment on a real-world solution. Not some useless tool that they’d never use. He asked, “How can our team build software that uses Generative AI and that has practical applications in our business?”

The solution was LiberateChat.

Generative AI: from Idea to Validated Solution

LiberateChat is a modular, customizable software. It helped Liberate Labs directly in hiring, firing, training, and promoting employees. The software helped enhance all employee functions.

It started as “an experiment into implementing generative AI by and for Liberate Labs itself.” Here are some of the unique features of LiberateChat:

Knowledge Graph Generation

The knowledge graph feature shows relationships between concepts. This helps users and serves the ultimate purpose of using AI to understand data and generate new insights. In the example above, the knowledge graph was able to understand and graph the relationship of concepts from a scientific paper.

For Liberate Labs, this feature matters because it is a novel way to use Generative AI and LLMs in real-world scenarios. In fact, CIO Jawad Iqbal said he has “never seen this capability before with AI.”

Sales Agent

The sales agent feature is a conversational chatbot made to have prospect conversations. Using internal documents like SOPs, the employee handbook, marketing materials, and others, it acts like a live representative. This feature was a success in that it works intelligibly to interact with prospects using Generative AI and LLM. It gathers prospect info, then sends them an email while CC’ing a Liberate Labs employee for further communications.


The multi-tool combines an LLM with other tools to answer questions and solve problems. For example, a user can ask:

  • “Who is Leonardo DiCaprio?” and the multi-tool will use Wikipedia to help answer
  • a math question and the multi-tool uses a calculator to solve the problem
  • Present-moment questions, such as “What is the weather today in Boston?”, and the multi-tool will use a weather app to answer.

The multi-tool matters to Liberate Labs because the experiment in integrating multiple tools with a LLM proved successful. This allows Liberate Labs to be at the cutting edge of Generative AI use cases. It also has delighted prospects on sales calls and led indirectly to client acquisition.

LLM Employee Trainer

This feature takes internal documents like the employee handbook, safety data sheets, and others. Then, it quizzes and grades employees on their understanding. Using this trainer feature, Liberate Labs was able to train new employees without using a supervisor or manager to do so.

Data Visualizer

This feature uses Python to generate data visuals. Users can plug in any data set and tell the program to take specific sections of that data and generate line graphs, pie charts, and other data visuals. This matters to Liberate Labs because it was a successful experiment in using Generative AI and LLMs to quickly and accurately display data in a visual format.

Tech Stack Used:

Front End: React
Backend: Flask
Vector Database: Mongo Atlas
LLM API: Open AI GPT-4 Turbo
Deployment: Digital Ocean

Results: Enhanced Demoing, Employee Management, and Client Services

Liberate Labs now uses LiberateChat in live sales calls for demoing purposes. Though it had bugs at first, LiberateChat has quickly become a key player in winning over clients. Seeing Generative AI used for their employee training in real-time always produces a “wow” response.

Additionally, building this Generative AI solution has enhanced the regular workflows and communications of the Liberate Labs dev team. The team is more cohesive and experienced in using Generative AI for real-world applications. This experiment, of which Liberate Labs has many, was a massive success.

Lastly, the benefits to employee management have saved Liberate Labs dozens of hours. This has allowed Liberate Labs to be streamlined and focus on higher ROI activities – such as building Generative AI solutions for clients.

Does your B2B SaaS company need Generative AI? Don’t let your competitors implement Generative AI before you do and steal your market share. Hire Liberate Labs today and get Generative AI solutions that work (without having to wait months to get them).

We are a high-velocity product team. We drive PLG motions for your SaaS through product management, design, and full-stack development

Clients feedback

Experimenting with Generative AI in a real-world, internal-use scenario helped our team go from scripters to product-oriented developers.

Create an account to access this functionality.
Discover the advantages

Create an account to access this functionality