How PwC's GenAI Powered “TLDR” Tool Saves Workers From Information Chaos and What Businesses Can Learn From It

Some tech experts might have seen it coming long before but with the beginning of 2023 we saw a new era defined by the intersection of artificial intelligence and generative algorithms that has given rise to generative AI or GenAI. This synthesis of cutting-edge technologies holds the key to unlocking unprecedented benefits for both the external public and for the units in the background around us Chief Data Officers (CDOs).

As we delve into the intricate web of possibilities woven by generative AI, this article illuminates the impact it has on the broader user public, while exploring the transformative advantages it presents for the development of new data products, business models, and insights for training possibilities.

Looking first at the user perspective: For the users, various advantages are making their daily or work life easier. From personalized user experiences to enhanced services, GenAI creates a landscape where individuals can reap the rewards of tailor-made solutions that cater to their unique preferences and needs.

In a company such as PwC, this could be especially relevant in advisory and consulting mandates where parts of the project (management) can be taken over by an AI tool and contribute to large savings of time and effort making the project work more efficient.

At the same time, we as Chief Data Officers are at the forefront of going one step further and leveraging the wealth of information generated by GenAI to fuel strategic decision-making and enhance growth.

The abundance of data helps push forward innovation and create new data-driven solutions as well as utilizing the data insights and datasets for training purposes both for tool [improvement] and employee training.

At its core, generative AI harnesses the power of artificial intelligence and generative algorithms to generate data sets that transcend traditional boundaries. This synergy opens doors to new insights, enabling not only a deeper understanding of user behavior but also making it easier for Chief Data Officers to utilize user data.

That way, they can push initiatives in new product development and use the output data for training purposes. As we navigate through the multiple benefits, it becomes evident that generative AI is not merely a technological evolution; it also has the potential to redefine how we approach innovation and decision-making.

Better work with GenAI at PwC: TLDR for multi-purpose insights and data

How is it done in reality? First, we need to identify the most common problems and assess how far they can be tackled with an AI solution. Potential problems in large organizations are various. They range from:

• Having too many resources to navigate through

• The information is inconsistent and not properly maintained

• The content might be difficult to understand, and documents can be lengthy and time-consuming to read

• Intranet search function is not performing well-enough

It seems that oftentimes an overarching tool is missing that checks content, clears risks, and offers a digital solution. Let’s have a look at PwC. The sheer size of the global network structure is beneficial for a swift operation mode on the one hand but also favors the creation of silos and makes it difficult to keep the overview of our knowledge landscape.

AI can support such a huge structure to set up an internal knowledge management system and keep track of processes and information within the network. Such a system requires a solid data foundation to build on and reliable and regularly updated sources of information. An essential part of it is data quality.

Data quality is highly significant for generative AI performance and its reliability, as these systems leverage training data to enhance their outcomes. However, many organizations face a significant challenge: the lack of structured and formatted data.

Without them, companies cannot harness the full potential of GenAI and risk inaccurate and biased results due to inadequate training of data-based products. Thus, data management and the role of Chief Data Officers are crucial to build a solid underlying foundation.

An applied example of structured data usage and generative AI implication is PwC's TLDR tool. TLDR adapts to the busy lifestyle of the corporate world and stands for “Too Long Didn’t Read.” The ever-increasing amount of information makes it impossible for human capacity to keep up with and have an overview, especially when data is unstructured.

Unstructured data leads to a chaotic landscape of possibly valuable information that requires artificial intelligence to manage, process, and analyze it to extract insights and make them usable. TLDR supports employees in saving time and working more efficiently by summarizing the content of large document volumes.

The system leverages generative AI to generate summarized information from large documents for easy consumption by the user. It evolved from another system that utilized AI to extract entity information and content from unstructured documents. The benefit besides solving a user’s problem is that the data that was originally in an unstructured and disorganized state now becomes structured.

This structured data can form the basis for further analysis and training. From a data management perspective, the user act as providers of structured data to enlarge the corpus of the firm.

However, the advantages of this tool not only lie in user-friendliness and employee enablement, it also supports the empowerment of the entire organization. Moving forward, uploaded data might not only be useful for an individual worker - but all information compiled is highly significant for the entire company and can be used for further analytics activities.

This is primarily relevant for units of data experts, such as Chief Data Offices, as reliable data is the foundation for a broad set of work strings in the background such as product development, informed decision-making, operation enhancement, and strategy development. The reliability of data is hence an essential part.

Therefore, CDOs rely on high-quality and well-managed data to support the organization, drive positive business outcomes, and ensure compliance with essential regulatory requirements.

Wrap-up: Opportunities in GenAI for data-driven innovation

In conclusion, high-quality, structured, and formatted data is essential in the training of generative AI systems as well as for creating opportunities for data-driven innovation within companies. Creating such data assets can be facilitated by encouraging users to interact and use tools that solve specific products and the resulting data assets can be considered a by-product.

Accuracy, diversity, and relevance of the input highly influence generative AI’s outputs, leading to enhanced performance and reliability. Accordingly, organizations can leverage generative AI systems and their advanced capabilities and outcomes in various ways such as process automation, or personalized customer experiences.

Nonetheless, these results can also be used on a higher level - if companies collect all internal information, they can improve their data analytics, leading to informed decision-making and strategy formulation. As data plays a crucial role in the “era of GenAI,” CDOs become significantly important to support organizations in their data management.

About the Author:

Marcus Hartmann is a Partner and the Chief Data Officer for PwC Germany and Europe. PwC Germany. More than 13,000 dedicated people at 21 locations. €2.61 billion in turnover. The leading auditing and consulting firm in Germany.

As a proven data expert, Marcus has spent his entire career in the data & analytics industry, helping companies to move more easily and quickly in an increasingly data-driven world and to develop and implement new data-driven business models.

Marcus Hartmann joined PwC Germany in August 2019 and established the Chief Data Office and a corresponding internal Digital & Delivery Unit. He leads a team of data, software, and digital experts to create and develop the foundations for efficient and scalable data use within the company and the realization of highly scalable, market-oriented data products and new digital business models. In addition, he is teaching digital technologies at the Macromedia University of Applied Sciences.