Collaboration of humans and AI in the enterprise

Nobody loves cold calls and bad telesales pitches, even if the product or service is something you might want. Researchers have developed a smart system that helps sales teams improve their interactions with customers over the phone using artificial intelligence (AI).

The botconnect system is a cloud-based software application designed to help service and sales teams create effective conversations with every customer interaction. The innovative system uses methods such as a deep learning neural network to interact with a human more effectively.

The system is designed around natural interactions to make it easier to work with. “For example, account managers or customer service agents can interact with the botconnect system via chat and voice in natural speech at their desk or mobile devices,” says project coordinator Alex Franke.

The EU-funded botconnect project resulted in a software application that aims to increase the number of leads and sales during calls to customers from call centres. The team successfully made botconnect the first real-time sales advisor for enterprises. In effect, botconnect partners with salespeople and makes selling over the phone a more effective approach.

Working efficiently to create botconnect

“Botconnect has been built as an agile company with a flat hierarchy in order to ensure fast adoption by changing markets and requirements,” Franke says. The researchers adopted established processes such as Design Thinking and Scrum to successfully drive the project.

The team applied cutting-edge deep learning algorithms that rely on large chains of neural networks running within a cloud infrastructure. The botconnect team trained the neural networks using large business datasets to optimise dialogue and increase sales made over the phone.

The researchers found that service and sales representatives easily accepted the botconnect sales assistant. The team also demonstrated that using botconnect led to increased revenue for sales teams.

Overcoming barriers to acceptance

The project was not without its challenges, as the researchers found it difficult to manage the user’s expectations on what an AI-driven sales assistant can do. They overcame this issue with regular customer visits and by working closely with key users in a test group.

The researchers discovered that generating sales can be difficult in short interactions, especially when the products sold are complex, multiple systems are used, or there is a large amount of customer data. To solve this, the team analysed their clients’ customer bases to optimise how the sales dialogue should flow. “While most service and call centre software tries to reduce costs, botconnect works towards increasing the customer's revenue by applying state-of-the-art AI to design the best possible sales conversation dialogue,” Franke says.

The reduced service interaction times also resulted in increased sales revenue and more satisfied customers.

Next-best action

Botconnect’s primary offering is what researchers call the ‘next-best action’: essentially the AI help salespeople offer something to clients at every stage of interaction, resulting in a happier customer and increased revenue or customer retention.

Looking forward, the team hopes to optimise dialogue by adding objection handling and other proven AI methods within the sales process. “We will also optimise our reporting module and work towards better ways to automate importing and exporting data.”

Reference source: Collaboration of humans and AI in the enterprise