Large Language Models (LLMs) are quietly weaving their way into your organization, empowering employees with capabilities like automation, content creation, and data analysis. Yet, for many leaders, their actual use remains unclear. What tasks are LLMs performing? What impact are they having? Tackling these questions can be challenging, given the vast amount of data generated by numerous users, prompts, and content. But worry not, there’s a method to the madness!

Imagine facing a mountain of text data from LLM interactions – a daunting task for humans alone. This is where topic modeling enters the picture, acting as a powerful tool. Techniques like BERTopic analyze this vast data, automatically identifying and summarizing key themes. The result? A clear picture of the various ways LLMs are being utilized across your organization.

But that’s not all! We can translate this mountain of words into meaningful insights. By feeding the structured list of topics and their associated words to an LLM like OpenAI GPT, we unlock the hidden story within the data.

Here’s what you gain:

  • 1. Clear Visibility: Gain a comprehensive understanding of how LLMs are streamlining workflows, boosting productivity, and driving innovation, even in unexpected departments.
  • 2. Bottleneck Identification: Pinpoint areas where LLM adoption is low or ineffective, allowing you to address challenges and maximize their value.
  • 3. Data-Driven Decisions: Use factual insights to guide future LLM investments, ensuring they align with your strategic goals.
  • 4. Employee Empowerment: Equip employees with the knowledge to leverage LLMs effectively, unlocking their full potential and transforming daily tasks.

Don’t let the data mountain obscure the value LLMs bring to your organization. By leveraging the power of topic modeling and LLMs, you can gain clear insights and unlock the true potential of these transformative tools.

Figure 1

Ready to embark on this journey of discovery? Let’s work together to understand how LLMs are shaping your organisations future.

Prompt

“Write an email to a colleague to reschedule a meeting thursday because I have an unexpected conflict involving a phone call. Apologise for any inconvenience and offer two alternative times for the rescheduled meeting.”


    "1": [
        [ "meeting", 0.015928226435436704 ],
        ["tomorrow", 0.007408962865214033 ],
        [ "call", 0.007053428896275531 ],
        [ "thursday", 0.005805450289975135 ],
        [ "availability", 0.005333975677416746 ],
    ],

“I’ve categorised these into general thematic areas:

Topics related to professional interaction and discussions, such as meeting schedules.”