USING ARTIFICIAL INTELLIGENCE TO MAKE REAL CUSTOMER INTERACTIONS MORE REWARDING
Modern consumers interact with their service providers in many more ways than in the past – including email, social media, instant messaging – on average, almost six different channels for each customer relationship. Today, the sheer volume of these contacts means that only a fraction of them can be reviewed and analysed.
Our Nespresso Canada Customer Relationship Centre, for example, accumulates approximately 61,000 hours of telephone recordings per year. Due to the prohibitive costs of manually listening to and scoring these calls, only around 2% of these phone calls are analysed for quality, compliance, and agent performance.
This process is not efficient for data collection or improving consumer experience. So even though our consumers are giving us feedback on a daily basis, most of that feedback is never used.
Which is why we want to use Artificial Intelligence (AI) - powered speech analytics to provide more reliable data to improve agent performance, enhance business intelligence and consumer experience.
AI speech analytics will allow us to process a full spectrum of our customer interactions and focus our analysis on the most insightful ones, rather than randomly selecting a small percentage of calls to identify patterns and problems areas.
We wish to implement this in Nespresso Canada & Nestlé Canada Inc. call centres. To achieve this, we require a Minimum Viable Product (MVP) to test and learn, and also to train our own algorithms, which should be capable of being scaled up.
Where you come in
We are looking for a partner who can provide an AI-powered Platform to run Speech Analytics (Speech-to-text, Natural Language Processing (NLP) and Business Analytics) on call centre phone calls to obtain operational insights, enhanced consumer experience and increased sales.
We’d like you to build, implement and test an AI-powered MVP with:
- Speech-to-Text to transcribe recorded conversations that can scale to serve any speech recognition needs
- Natural Language Processing to extract information from transcripts with NLP to surface and summarise the important parts of call recordings
- Sentiment analysis to determine negative, neutral or positive scores for call segments or complete calls
- The ability to train our own Machine Learning algorithms (e.g. AutoML) to identify call segments and increase accuracy
- Customised Business Intelligence reports based on quality management needs, with insights accessible from analytical dashboards
We will be judging submissions to this challenge on the following criteria:
- Easily available insights, once transcriptions and sentiment analysis from recorded conversation are extracted and analysed
- Automated call scoring and agent monitoring to provide KPIs from call interactions and track call drivers and business trends to make more informed decisions
- Business Intelligence based upon call analysis and AI predictions
Currently, Nespresso Canada analyses approximately 3000 calls (around 1% of total calls) for quality assurance and compliance, taking 2800 men hours.
Our goal is to be able to auto-analyse and transcribe 100% of the calls, increasing by 90% the call monitoring (assuming a 10% of calls don´t have enough quality) and reducing by 50% the time a Quality Manager spends with each call.
Why it matters
In a consumer-centric organisation, AI technology offers big opportunities to enhance consumer experience, from more accurate evaluation to personalised actions. It can also help develop the quality of our customer relationship teams and help reduce costs associated with agent turnover. With this new process, one bad call will not be make or break for the agent’s monthly scorecard. It will now reflect their true performance throughout the month, and enable effective coaching plans to be developed on the basis of all the calls.
If you have any questions about this particular challenge, you can get in touch with Oscar here.
To make sure you don’t lose your application while you are working on it, please make sure to save your answers into a separate document as backup. If you don’t receive a confirmation that your application has been submitted, please get in touch with us via the Contact Us page and we will help you with your application.
Please note that your solution must be able to support Canadian English and French.