Please feel free to ask a full copy of the case study at office@iac.ai

• Context

• Credit card transaction fraud causes annual losses of some $28 billion* and is a source of frustration for customers retailers and financial institutions
• Even though losses tend to be insured, reimbursements take months and customers have to go through several administrative and inconvenient tasks.
• Banks proactively monitor transactions to flag potential fraud; a typical high-level process is shown in Figure 1

• Solution

• The company built a cognitive agent able to dialogue consistently across channels (e.g., emails, messaging, voice, and social media).
• It could also understand emotions by watching users through webcams. Its look and behavior were designed to be very similar to those of human agents. This increased acceptance of the cognitive agent by clients and improved their personal connection with it.

• Results

• Automation of more than 80% of the process
• Reduction by half of the end-to-end process time, from the identification of the issue to the reimbursement by the insurance

Intelligent Automation Solutions

• Machine Learning to improve itself over time from from observing conversations between people
• Computer Vision enabled through webcam to understand emotions and recognize users from their voice or face (using biometrics)
• Robotic Process Automation to process changes in systems

Number of fraudulent
transactions solved

0 %

Additional savings per year

>$ 1 M

• Context

• A hotel chain needed to manage more than 20,000 client interactions per day through emails, calls, messages, and social media.
• This workload was managed by a team of 240 customer care agents servicing clients 24/7, supporting sales, room reservations, inquiries, and after-sales.
• This team was facing low employee morale and high employee attrition, with an annual turnover of 40%.
• The low morale was mainly due to high and uneven workload, work pressure, and long shifts.
• These problems resulted in low service quality, rated less than 5 out of 10 by the chain’s customers. In order to solve this situation, the hotel chain decided to launch a digital transformation effort, including the implementation of a cognitive agent.

• Solution

• The company built a cognitive agent able to dialogue consistently across channels (e.g., emails, messaging, voice, and social media).
• It could also understand emotions by watching users through webcams. Its look and behavior were designed to be very similar to those of human agents. This increased acceptance of the cognitive agent by clients and improved their personal connection with it.

• Results

• The cognitive mainly covered the night shift and had taken a large part of the team’s workload
• The team size was reduced enabling the reminding members to focus on the more complex and exciting interactions that require higher-level relationship skills.

Intelligent Automation Solutions

• Machine Learning to improve itself over time from from observing conversations between people
• Computer Vision enabled through webcam to understand emotions and recognize users from their voice or face (using biometrics)
• Robotic Process Automation to process changes in systems

Increase in
customer satisfaction

0 %

Decrease in
employee turnover

0 %

• Context

• The main objective of this banking process is to collect and assess future client data to provide better service and manage risk.
• In the past, this process was mostly manual, and hence, it was long and painful for clients and employees alike.
• Complex process combining a wide range of capabilities, including executing, reading, dialoguing, understanding, analyzing, and learning..

• Results

This end-to-end process automation resulted in

• Reduction of process time by over 90%
• Cost reduction of over 80%
• Enhanced compliance, and a significantly improved customer experience.

Intelligent Automation Solutions

• Chatbots to welcome prospective clients 24/7 and answer their questions
• Application programming interfaces (APIs) to share data between government regulators and banks as well as for collecting client information from sources such Google and Facebook
• Analytics and Machine Learning to estimate and continuously update the risk levels and segmented the client pool to customize services.

• Computer vision and natural language processing to intelligently process unstructured documents such as bills, contracts, and identity documents.
• Robotic process automation to drive the execution of the overall process, performing all rule-based activities such as reconciling data, checking data, and sending emails.

After implementing 12 similar
end-to-end processes over an
18-month period

Cost reduction

0 %

Revenue increase

0 %