PwC simplifies data retrieval with intelligent bots built on Power Virtual Agents

A wealth of content spread across multiple repositories

The global PwC network of business consultancy and accountancy firms serves a diverse range of industries, including aerospace, utilities, hospitality, pharmaceuticals, and many others. PwC continues to drive innovation, learning, and development across its network, working with Microsoft to stay at the forefront of every industry that it serves. But operating at a global scale means PwC faces an overwhelming volume of data and content spread across multiple repositories in its network. This can make it difficult for teams to find the content and learning resources that they need.

PwC wanted to give its people a faster, more seamless way to search through and extract information from its content repositories. By doing so, it would speed up learning and development and save valuable time for teams across the PwC network.

The need for quick and easy content searching

PwC already used Microsoft Power Platform globally to develop a low-code culture and help teams create quick-win automations across its operations. After exploring the platform’s capabilities, PwC added Microsoft Power Virtual Agents, part of Microsoft Power Platform, to its tool set.

“We knew that Microsoft Power Platform was greater than the sum of its parts—we have so many different use cases for its components,” says Dushyanth Chandramouli, Director at PwC. “We saw huge potential in Power Virtual Agents to help us analyze our data and enable content retrieval at scale in a more contextual way.”

Instead of a traditional search engine that relies on keywords and phrases, PwC wanted to create a content search functionality personalized to the roles of individual team members. “Every team member across our network—whether they’re a project manager, software engineer, or consultant—has different areas of interest, and not all the content in our repositories is valuable to any one individual,” says Chandramouli. “We wanted to offer individuals a way to filter the content they’re interested in, whether they’re looking for a strategy presentation, training resource, or client-related document.”

A helping hand from intelligent chatbots

To solve its large-scale data challenge, PwC used Power Virtual Agents and Microsoft Bot Framework to create smart chatbots that automate searching and information sharing across repositories. The bots can support team members across multiple content repositories including, for example: SharePoint, Confluence, and Google Drive. Instead of using the search bar in a content repository, team members interact with the repository’s chatbot by submitting queries, links, and keywords, and the chatbot retrieves the content most relevant to their request. 

“Our people have the ability to create their own chatbot experience in whichever repository they’re using very quickly,” says Chandramouli. “Using Power Virtual Agents, team members can seamlessly translate the repository’s hierarchy into a more meaningful, decision tree-like structure that acts as an input for the bot, all with minimal coding.”

Importantly, PwC can embed its bots in any website or repository—such as Confluence—and they can also be used as a Microsoft Teams app. Behind the scenes, the solution uses Microsoft Power Automate to trigger flows from the hierarchy file in SharePoint that create a bot based on where each person is trying search. 

PwC’s Power Virtual Agent chatbots apply in myriad use cases, including targeted content retrieval from large, multi-section audit user guides to streamline the audit experience (figure 1). PwC uses the bots built with Power Virtual Agents to support the use of Microsoft Power Platform at scale across its network of professionals (figure 2). It is also being used to advise users with business use case solution recommendations leveraging appropriate Power Platform components (figure 3).

“Our teams can use the Power Virtual Agents bots to learn more about Microsoft Power Platform,” says Chandramouli. “The bots answer a wide range of frequently asked questions, along with questions about the platform’s more nuanced capabilities.”

Using Microsoft Bot Framework Composer, built on the Bot Framework SDK, PwC’s developers can add advanced capabilities to the low-code Power Virtual Agents bots—such as automatic greeting initiation and multiple choice capabilities around trigger phrases. The organization’s developers can author, test, provision, and manage conversational experiences using Bot Framework Composer, allowing them to create more dynamic bots than those spun up using Power Virtual Agents. 

Efficient and effortless content navigation

PwC started by rolling out the bots to a target group of 3,000 team members and is rapidly extending broadly after measuring the project’s success. These initial business users reported finding the specific content they need at a much faster pace using the bots. 

“The bots we’ve built with Power Virtual Agents have proved especially useful for repositories like Microsoft Teams,” says Chandramouli. “It’s possible to lose track of files in the best repositories, but using a bot, we can retrieve whatever content we need, no matter where it’s stored.” 

PwC predicts time savings for both developers and non-technical team members throughout its firms. “For our developers, using Power Virtual Agents prevents them from having to create new solutions from scratch. And for other team members, interacting with the bots significantly speeds up content searching,” says Chandramouli. 

Faster content searching will also help PwC respond to RFPs, strengthen its customer relationships with quicker responses to queries, reduce the risk of duplicate asset creation, and offer an enhanced learning and development experience.

 “Power Virtual Agents is going to help us when we roll out new technology, because there are always a lot of different governance processes involved,” says Chandramouli. “Using the bots, we’ll be able to easily scale training across the entire organization.” 

Big bot ambitions on the horizon

PwC can now use the bots to drill down into content repositories in a more meaningful way. “The bots created with Power Virtual Agents really shine when the content has multiple degrees of nesting,” says Chandramouli. “Our teams can interact with the bots to keep navigating further into the decision tree and quickly find all the content that’s relevant to their searches.”

Next, PwC wants to take its bots further by adding AI capabilities with Microsoft AI Builder . Team members with little technical experience can then add prebuilt AI models using an easy, intuitive interface. “We’re looking at how we can use Microsoft AI Builder to include capabilities like continuous learning and natural language text recognition in our bots,” says Chandramouli.

Find out more about PwC on Twitter, Facebook, and LinkedIn.