How to Boost Workforce Productivity with ChatGPT while Managing Risks

Insights | 30.06.2023 | By: Jakub Lutter

Generative AI, has gained widespread recognition for its ability to engage in hilarious and entertaining conversations, earning it a reputation for being the life of the digital party. However, what often goes unnoticed is its remarkable potential as a powerful productivity accelerator in professional settings, allowing users who leverage Generative AI to achieve a performance of 6 percentage points higher than their peer group not using such tools or replacing almost 50 hours of work weekly with 30 minutes of ChatGPT usage within a whole team.

Team leaders and managers should not underestimate the value of Generative AI and large language models (LLMs) as a productivity enabler, as they can streamline workflows, provide instant information, and assist with complex problem-solving. By harnessing the dual nature of LLMs, organizations can unlock a world of efficiency and enhanced employee experience, propelling their teams to new heights of success.

LLMs as a productivity booster: The evidence

By deploying at workstations of users within teams dealing with transactional and recurring activities, we were able to derive insights into how people work and how ChatGPT is leveraged, providing evidence of its impact on enhancing efficiency and streamlining workflows, thanks to’s insights and measurements of productivity gains,.

Among the users of LLMs in a Pharmaceuticals company in the APAC region, an impressive average productivity rate of 96% was consistently achieved, surpassing the performance of the rest of the team by 6 percentage points. These findings highlight the tangible benefits that LLMs brings to the table:

  • Two teams working on industrial projects composed of 13 people each perform complicated pricing calculations. One of the teams utilises LLM for 30 minutes a week, allowing them to save almost 50 hours of work each week on filling in pricing calculation templates in comparison to the team not using LLM. Meanwhile, the second team of 13 people does not utilise an LLM during their work and the productivity gap is noticable, mainly in the additional 50 hours of work spent on the pricing calculation templates.
  • Another user leverages LLM’s capabilities to generate quick Excel formulas, empowering him to manage invoicing process efficiently and enabling him to be 10% more productive than the rest of his team.
  • Another example is another user who relies on LLM to generate concise and structured responses to client queries regarding shipment processing, enhancing customer satisfaction and gaining 5% of productivity over his colleagues.

These diverse use cases demonstrate the wide-ranging applicability of LLM in driving productivity across various functions and industries.

LLMs as an enabler of automation: Finding the opportunities

While almost everyone has already tried ChatGPT or a similar LLM model, not many people find opportunities to use it in a work setting and team leaders struggle with proper identification of opportunities for usage of such tools. provides a complete insights into activities executed by employees and the tools that are used while executing them. Not only can serve as a source of an automation or improvement pipeline for LLM use cases because of the capability to identify opportunities for LLM usage, but also tangible benefits in terms of productivity increase and cost savings can be calculated prior to any LLM deployment.

Finding opportunities for LLM usage, documenting them and implementing them into action can be as easy as simply asking an LLM a question.

LLMs as a source of problems: Managing risk of SaaS LLM models

While LLMs increase productivity, they also pose some risks that are associated with their usage. LLM models in general and ChatGPT in particular are closed-source applications, usually deployed in cloud infrastructure of a certain provider. Therefore, it might not always be clear what happens to data that is input into an LLM and whether it is used for potentially unwanted activities. Furthermore, confidential data should stay confidential and should not leave the company.

For this reason, many organisations have outright banned usage of ChatGPT and similar applications for the fear of data breaches or GDPR non-compliance. However, by banning ChatGPT, a productivity and automation goldmine stays locked.

So how can organisation steer clear of risk and still enjoy benefits of LLMs?

As part of a strategy for adoption of LLMs that addresses their risks, organisations can understand LLM usage patterns from insights and steer clear of any associated risks. At the same time, opportunities for automation and productivity enhancement can be identified with, leaving the path open to embrace LLM usage with open arms and enjoy considerable financial and productivity benefits, also contributing to a great employee experience.

Note: A big thank you to our colleague Arpit Gupte for providing valuable examples and data for this article.