4 Enterprise IT goals & their AI solutions

A look at four real-world examples of AI implementations.

For years technology professionals have been bombarded with messages proclaiming that AI will revolutionise their business. Experts have highlighted AI's ability to improve agility and productivity, reduce costs and mitigate risks - and yes, this is all true. However, with so many options available to them, IT managers really need to think about how AI could meaningfully benefit their business.

Like any emerging technology, AI projects run the risk of falling into the "science project trap" says Sudheesh Nair, CEO of business intelligence (BI) firm ThoughtSpot. He points out that organisations that have successfully deployed AI within the business are those who thought about specific use cases and how AI could advance them. 

"Whether you're leveraging AI to personalise customer interactions, improve processes or reduce operational costs, start with the outcome, then work backwards," he advises.

To help you see where your business might best benefit from AI, we share four real-world examples of AI implementations, illustrating the goal they were looking to achieve, and the AI technologies that enabled them to accomplish it.

 

Goal: Expand customer support services

AI solution: Virtual assistant

As part of its digital transformation programme, banking firm BBVA was looking for a way to develop digital customer support services, maximising the user experience by offering a customised service across all channels.

BBVA decided to trial a virtual assistant from IPsoft in its Mexico contact centre, which fields over 100m calls a year.

"Carrying out the implementation was complex. First we had to integrate it with multiple banking systems, identify and establish customer service processes it could support, and then adapt our processes - most of which were designed ten years ago for a traditional way of doing things," says Sergio Torres, BBVA Mexico's director of strategy and innovation.

The solution, which BBVA named Blue, has now been up and running for two years and the bank has seen tangible benefits including lower operating costs, higher customer satisfaction and increased employee productivity.

"Today, 32% of the total calls are resolved by Blue, which also delivers an 80% customer resolution success rate," says Torres. "Following the success of this project, we expanded the use of digital assistants into other areas and markets with the goal of developing digital sales, advisory and support services, and providing customers with a fully digital experience."

 

Goal: Grow revenue faster

AI solution: Predictive marketing

Gamma sells unified communications systems to small businesses. Joe Leverson, its head of digital marketing, was keen to explore ways to drive revenue and scale direct and partner customer acquisition, and the first step was to make direct acquisition more efficient.

"We needed to drive productivity from our existing sales and marketing channels and make life easier for our field sales teams - if we didn't do that, they would've been left scrabbling to create leads themselves, and it's very hard to scale that kind of activity up."

Gamma worked with Growth Intelligence to refine its lead generation process so that its intelligent revenue generation platform would add maximum value.

"The second step was for us to provide them with our marketing outcomes so they could understand what our best prospects looked like and prospects to avoid. They then analysed our data and built the AI system, which took a couple of weeks."  

Gamma has seen a significant improvement in lead conversion rates and a substantial uplift in lifetime value generated per customer.

"We're generating opportunities worth 2.4x more revenue and now have a comprehensive view on our total addressable market. We know how likely a company is to buy, which means all our marketing is aligned. Before we risked having multiple initiatives and lead sources," Leverson says.

 

Goal: Empower employees to use analytics and drive business value

AI solution: Search and AI-driven analytics

Paul French, Nationwide Building Society's director of business intelligence, visualisation and reporting, data and analytics, wanted the capability to challenge some of the society's institutional knowledge, and find new patterns and trends in its data.

He was seeing more and more requirements for colleagues to get right into the details quickly, and wanted to reduce the costs and inefficiencies associated with large numbers of analysts preparing and readying data. His idea was to develop an Amazon-like keyword search and shopping experience that would dramatically simplify data engagement.

"We wanted an AI tool that was easy to use and could interpret natural language, so any employee, regardless of technical or data expertise, could easily learn and use it," he says.

Having made the case for an additional BI tool to be added to Nationwide's existing technology estate, the society went live with an AI solution from ThoughtSpot, and has now implemented four use cases including mortgage lending, savings and customer products, with more in progress. In just one-use case, the system replaced 9,000 static emails being sent monthly, turning off clunky BI reports. 

"Before implementation, our team received raw mortgage data in huge Excel files attached to emails - we were sending out about 9,000 per month. Going through these manually required more time and effort than most people were willing to put in; in fact, only 90 people were looking at the data. With our new self-service system, we have more than 200 people sharing the same view. Even better, they can explore and drill down into live data helping them answer questions as they arise."

 

Goal: Drive revenue generation and optimise productivity

AI solution: Natural language processing (NLP) and deep learning

Finance broker Redburn was looking for a way to improve client service, create productivity improvements and drive smarter enterprise decisions.

It decided to implement FeedStock's Synapse deep learning technology, which turns existing data within the business into actionable insights.

"We started with a proof-of-concept phase, which involved the integration of FeedStock's automatic data capture modules into our existing IT infrastructure and the collection of the first data points in the data warehouse," says Redburn's head of technology, Matthew Norman.

"The second stage of the project was the delivery of a refined deep learning model which classifies Redburn's outgoing client interactions into five distinct categories and further enriches the dataset we use to maximise enterprise opportunities and streamline workflows. It also involved a new front-end dashboard with custom-built user-interface. This UI solves additional business use cases and provides instantaneous client engagement insights for the front office to use to help refine their outreach and drive better sales performance."

This project has enabled sales professionals and managers to assess and improve the human-to-human relationships taking place within the organisation and across the client base automatically, and revealed three times the number of chargeable client events from within Redburn's structured and unstructured communications data.

"It provides tangible, quantifiable commercial benefit to Redburn. In today's increasingly digital and remote working environment, gaining access to this structured dataset enables enterprises to streamline operations, manage costs effectively and refine and improve revenue-generating opportunities."