Client:
Various internal teams across PwC Poland
Our role:
We delivered CDP, an intelligent, AI-powered, secure, and tenant-based document platform built for scale. We set up data extraction rules to search for and extract data, classify, and translate documents automatically.
Country:
Poland
Setting the scene
At PwC Poland, our back-office human resources, finance, and legal teams faced the constant challenge of processing a high volume of diverse documents. Manually reviewing employment contracts for specific clauses, extracting data from hundreds of financial statements, or verifying details across invoices and reports was time consuming and susceptible to human error.
CDP was selected to address these challenges as it provides auto-classification, semantic and exact data extraction, document matching, translation maintaining original formatting, and anonymisation.
With accuracy exceeding 95% in locating requested data, and over 90% in correctly identifying, matching, or processing cases, CDP provides a dependable solution for optimising back-office operations and enhancing data quality. The solution also provides enterprise-grade security.
How we helped
The CDP platform works across internal back-office functions. The primary goal is to automate and accelerate data extraction from a wide array of document types and formats. These include Word documents, PDFs, Excel files (including .xlsxm files), and scanned documents—and translating and anonymising content as required.
Our team configured CDP’s data extraction capabilities to meet the specific needs of each department:
- Human resources: The platform is designed to automatically extract key data—such as start dates, non-compete clauses, and entitlements—from employment contracts and translate documents across over 40 languages.
- Finance: CDP facilitates the extraction of figures, customer details, data from invoices, and reports and financial statements. It also automatically performs calculations to speed up validation.
- Legal: The tool swiftly examines lengthy legal agreements to find specific clauses or defined terms—a task that previously took hours—and anonymises content to comply with specific legal requirements.
The CDP platform’s data extraction tools use smart search features, such as semantic searches, to extract information both by meaning and by exact match. To make validation more efficient, every data point extracted was clearly marked within the source documents. By building a tool that our employees could easily tailor to their unique data requirements, it equipped our people with a powerful and practical tool for real-world use.
"By using CDP, we made it much easier for our teams to handle all types of documents quickly and accurately, no matter the language or format. This change means less manual work and fewer mistakes, and now 6,500 of our colleagues can save significant time every day. It's a big step for our business, making us stronger and ready to grow in the future. This platform is a powerful testament to our capabilities—not only do we propose this platform to our clients, we run our business on it."
Michał Targiel, Partner, PwC PolandImpact and potential
CDP has significantly reduced the manual effort required to process large volumes of diverse documents. It has delivered accuracy above 95% in identifying key data and over 90% in end‑to‑end workflow processing. This has allowed HR, finance, and legal teams to work faster, with fewer errors, and with clearer validation steps.
By standardising extraction, translation, and anonymisation across functions, the platform strengthened data quality and operational efficiency. Long term, CDP positions PwC Poland to scale document‑heavy processes securely and consistently, while enabling teams to focus on higher‑value tasks. The project also demonstrated how adaptable, AI‑driven tools can effectively support complex, real‑world use cases across an entire organisation.
Contact us
