Conversational Data Intelligence

Discover how Conversational Data Intelligence is enabling straight-through processing, continuous improvement and unprecedented business intelligence.

Conversational Data Intelligence

What is Conversational Data Intelligence?

Conversational Data Intelligence is a category of enterprise software that combines machine learning, natural language processing (NLP) and human employees to drive understanding and automation in business communications and conversations. 

Through machine-human collaboration, Conversational Data Intelligence creates structured data from masses of unstructured enterprise communications data, giving users the ability to analyse previously hidden business processes and automate low-skill manual processes that used to depend on human reading comprehension. 

See Conversational Data Intelligence in action as part of our platform tour.

What is conversational data?

Conversational data is data that's been extracted from a digital conversation. This can be done across a wide range of communications channels and message types, including emails, chats and even calls. Conversational data is valuable because it often contains information that isn't recorded anywhere else, and it can reveal things about your customers, employees, partners and suppliers you wouldn't have known otherwise. A customer may not provide comprehensive, honest feedback if asked directly, but their opinions and feelings will become clear once you have insight into all the interactions they've had with your agents or employees.

There are three main types of conversational data:

  • Intent data
    Intent data shows the intention of the sender or their reason for contact in the first place. Common reasons for contact include making a complaint, requesting a refund, and asking for an order update. However, beyond these broad categories intents can also be analysed in a very granular way on the level of an individual message.
  • Sentiment data
    Sentiment data tells you about the tone or emotional content of a message. When analysed as a whole, certain emotive and active phrases can tell you a lot about the emotional state of the sender. Once you can analyse the sentiment of a message, you gain fresh insight to inform metrics such as customer satisfaction (CSAT) and assess the quality of your service.
  • Entity data
    Entities are key words or individual phrases in a conversation that contain important information needed for responding to or completing a request. This could be anything, from a customer or order number to the date when a message was sent. This data is often extracted so it can be entered into other systems. When this process is automatic rather than manually performed, it's possible to automate entire conversational processes from end to end.

What can conversational data be used for?

Conversational data holds immense value for the enterprise, though is often neglected as an untapped resource. Conversational processes and workflows are unstructured, as is the data they produce. This unstructured conversational data normally resides inert in a corporate server, but there are numerous applications that can turn it into business value.

The most common use cases for conversational data are:

Conversational Analytics

Conversational analytics is the examination and interpretation of conversational data in order to extract useful insights from it. Companies use conversational analytics to understand the motivations and behaviours of their customers, and to gain a better understanding of their operational processes. It is a crucial method of extracting management information, operational and customer intelligence from business conversations and communications-based processes.

Learn more about conversational analytics.

Conversational Automation

Conversational processes - such as processing an account change or responding to an FAQ - are often highly transactional in nature. This means that they require only a small number of interactions to complete. Normally, these kinds of processes would be ideal fodder for automation tools. However, such conversational processes have been difficult to automate given their unstructured nature. Fortunately, conversational data is the missing link in automating these processes from end to end. Conversational data can be transformed into clean structured data through Conversational Data Intelligence. This data can then be sent downstream to automation tools like RPA to manage and complete the process.

Why is Conversational Data Intelligence important?

Conversational Data intelligence is important because it fills a major gap in business understanding and process efficiency. Business runs on communication. Almost every operational process and workflow involves human interaction - primarily digital - at some point. Customers reach out when they want something, employees collaborate to get work done. An ever-increasing amount of this business conversation takes place on digital channels, though primarily email.

The problem is that communications channels are notoriously difficult to analyse, and the linked manual comms-based processes are very hard to make more efficient. This is because unstructured digital conversations produce very little structured data, and this is what most traditional analytics and automation tools rely on to do their jobs. Human language and conversation is actually very complicated and difficult for machines to process accurately and understand. That’s why businesses have always relied on human employees, their contextual awareness and understanding of language, to process and action business communications correctly.

However, this status quo is rapidly becoming unsustainable. The scale of digital communications is growing rapidly - much faster than what human workers can process efficiently. Almost 320 billion emails are sent every day, and this is expected to grow to over 376 billion by 2025. Every day, employees are bombarded with messages from colleagues, partners, customers and suppliers. The average employee now sends or receives 126 emails a day, and up to 40% of their time is spent in Outlook alone. It’s a massive drain on productivity and staff morale - 30% of workers see their inbox as the biggest distraction from actual work, and 22% say they want to quit their current role due to exploding email volumes

Even then, the scale of communications is so great that this effort often isn’t enough. Important messages land in shared mailboxes and aren’t read or actioned for days or even weeks. Service levels are impacted, handling times increase, business process efficiency slumps, and customer dissatisfaction grows. The problem is getting worse and no business can afford to let it continue. However, it isn’t a problem that they can solve with humans alone.

Conversational Data Intelligence provides a solution. It gives organisations easy access to advanced NLP and machine learning capabilities that anyone in the organisation can use. These platforms analyse the masses of unstructured communications at speed and at scale, pulling out the most valuable information and creating structured data that can then be used for automation. 

This is game-changing for process insight and efficiency. For the first time, businesses have total visibility and the ability to understand everything that is happening in their comms channels. They also have the raw data needed to create intelligent automated workflows from manual comms-based processes. This frees employees from the chore of manual email processing, letting them focus on only the most valuable work.   

What are the benefits of Conversational Data Intelligence?

Conversational Data Intelligence is a crucial part of an organisation’s operations infrastructure, and the only way to process communications quickly and effectively en masse. 

Having a complete overview of communications channels and the data to create powerful automations helps a business drive key improvements:

  • Increased cost savings
    Conversational Data Intelligence speeds up many analytics activities and facilitates the automation of manual comms-based tasks. This cuts handling times and costs across the board.

  • Improved productivity
    Employees spend less time on low-value email tasks, giving them more time to focus on value-add.

  • More value from digital automation
    Through Communications Mining, Conversational Data Intelligence reveals valuable automation opportunities that were previously hidden within unstructured data.

  • Data-driven decisions
    Gain visibility into opaque channels and processes to identify problems, inefficiencies and quantify change opportunities.

  • Operational scalability
    Conversational Data Intelligence uncovers the costly processes and blockers slowing down the business, while providing automation and workflow tools with the structured data they need.

  • Improved service & customer experience
    Streamlined communications means clients are serviced more quickly, leading to a better customer experience.
  • Improved customer intelligence, experience and personalisation
    The factors impacting customer loyalty and fuelling churn are more visible when you can process what they are saying en masse and in real time.
  • Happier employees and improved retention
    Conversational Data Intelligence reduces the need for employees to manually process emails and other messages, allowing them to focus on more interesting work.

Book a platform demo to see what Conversational Data Intelligence can do for your organisation.

Operational scalability

Communications Mining uncovers the costly processes and blockers slowing down the business, while providing automation and workflow tools with the structured data they need.

More value from automation

Communications Mining reveals valuable automation opportunities that were previously hidden behind a screen of unstructured data.

Increased cost savings

Communications Mining speeds up many of the processes related to customer service, product intelligence and service monitoring. This cuts handling times and costs across the board.

Better compliance

Communications Mining gives compliance teams real-time insight into the business’s communication channels. Operational risks are surfaced and can be addressed more quickly.

Happier employees and better business intelligence

Communications Mining reduces the need for employees to manually process emails and other messages, allowing them to focus on more interesting work.

Improved productivity

Employees spend less time investigating issues and validating opportunities, giving them more time to focus on value-adding tasks.

Improved customer intelligence, experience and personalisation

The factors impacting customer loyalty and fuelling churn are more visible when you can process what they are saying en masse and in real time.

Data-driven decisions

Gain visibility into opaque channels and processes to identify problems, inefficiencies and quantify change opportunities.

Where can Conversational Data Intelligence be used?

A Conversational Data Intelligence platform can be used by anyone in a business - whether they are a C-suite executive, a manager, or a front-line worker - and it’s not limited to a single function or department.

Conversational Data Intelligence can be implemented anywhere that email and communications-based processes - including tickets, chats, CRM and ERP system notes - are prevalent. It works best when there is a high volume of messages passing between colleagues, customers, suppliers and partners, and where shared mailboxes and systems are commonly used to process those messages. All that a Conversational Data Intelligence platform needs to start creating value from communications is a large amount of unstructured data, and people ready to take it through a fast and easy model training process.

Conversational Data Intelligence is most commonly used for analysis or automation, but it achieves its true potential in an organisation when both applications are actively used. It strives to help users perform in-depth analysis of their communications data, discover and scale new automation opportunities.     

Conversational Data Intelligence platforms have already been deployed in sectors where high-volume, high-complexity transactions are integral. They can be found across the world’s largest banking, asset management, insurance, ecommerce, telecommunications, consumer and business travel companies. Conversational Data Intelligence is also present in the large-scale back offices of companies with extensive shared service functions.

See what impact Conversational Data Intelligence is having across numerous industries:

What features and capabilities are important for Conversational Data Intelligence?

Conversational Data Intelligence is one of various tools a business can adopt for digital transformation and intelligent automation. In analysing and digitising unstructured communications, it forms a complementary role with other tools like Process Mining, RPA, optical character recognition and intelligent document processing. 

It’s very important therefore, that a Conversational Data Intelligence platform can easily integrate with the rest of your technology stack. When evaluating a solution, consider the number and quality of partnerships the vendor has with system integrators, leading consultancies, and technology vendors. This provides a good indication of how easily the solution will integrate with your existing technology infrastructure.

The user experience is also crucial, as this indicates how quickly a Conversational Data Intelligence platform will be adopted by users and start generating value. Subject matter experts (SMEs) - usually employees - play a critical role in the model training process known as Active Learning. This is where SMEs train a model by consistently correcting and ‘grading’ its predictions until the model is accurate, balanced and reliable enough to be deployed. 

This process can be daunting for employees without technical training. Yet to scale any machine learning tool across a business mandates a large pool of users, so the underlying models can be exposed to and train from as many different messages and scenarios as possible. This is especially important given a persistent shortage of technical - and especially AI - skills across industry. A good Conversational Data Intelligence solution will provide prompts, training materials and added tools to make the process easy and seamless for users. The best solutions can help even non-technical SMEs train up powerful machine learning and language models in a matter of hours. 

Is Conversational Data Intelligence the same as Conversation Intelligence?

Conversational Data Intelligence and Conversation Intelligence are different tools, though they often leverage similar technologies to understand human language and interactions.

Conversation Intelligence uses a form of NLP to record and understand the spoken words of humans, picking out the important parts of the conversation for later analysis. It is most commonly used by sales and customer service professionals to look back at previous interactions with prospects and customers to pick out the most important information or to improve their techniques for future interactions.

By contrast, Conversational Data Intelligence uses the power of NLP and machine learning to understand and action the unstructured data contained in digital communications. It is used to understand and automate communications and email-based business processes at speed and at scale, structuring the data and filtering the most useful information for the business user.

Most Conversation Intelligence solutions come off-the-shelf and are fully pre-trained. They are black box solutions which don’t inform users how the underlying language models work, and which rarely let them be customised to suit the needs of the user. Conversational Data Intelligence platforms, meanwhile, fully involve users in the model training process through Active Learning. These solutions usually provide pre-trained options and concepts to shorten training times, but the focus is on allowing users to actively train the language models to pick out the most useful insights from communications accurately and reliably. 

How is Conversational Data Intelligence different to Communications Mining?

While closely related concepts, Conversational Data Intelligence and Communications Mining are quite different. Communications Mining is only one popular application of Conversational Data Intelligence technology. It describes the use of Conversational Data Intelligence to convert unstructured communications into structured data for detailed insight and analysis. 

However, Communications Mining does not cover the exploitation of that new data for automation and downstream processing. While the creation of intelligent, automated workflows requires an additional automation tool, such as RPA, to execute the process from end-to-end, the creation of that data is a vital part of the process and a key application of Conversational Data Intelligence.  

Learn more about Communications Mining.

How do I get started with Conversational Data Intelligence?

Developing your own Conversational Data Intelligence platform is a difficult undertaking with no guarantee of success. Building a platform from scratch that combines NLP with machine learning requires a wealth of data science and AI talent that very few organisations have.

Even large organisations with plentiful resources can struggle to see the development process through - the average tenure of a technology or transformation leader is significantly less than the time it takes for such a solution to reach production and start delivering value. This increases the chances that the project will be scrapped once the original creator and project team have left the company.

Even if a project survives development, the finished product is unlikely to perform as expected. This is compounded by the fact that the solution is a product of multiple different dev teams and project leaders. Post-deployment fixes and support will further inflate project costs and lengthen the path to profit.

It’s recommended, therefore, that you adopt an existing Conversational Data Intelligence platform. The most difficult development decisions will already have been taken by the technology vendor, and the only time to value will be the time taken to tweak the solution to the customer’s needs.

The Conversational Data Intelligence Platform by Re:infer, a UiPath company, is an ideal solution that delivers full Conversational Data Intelligence capabilities, and which starts delivering value after a short training period. Re:infer has worked with many companies to extract the value from conversational data and build intelligent products, services and workflows.

Request a demo or get started with a free trial.