CustomerServiceAI: Fully language-independent AI platform for automated and augmented customer service

Projektinformationen

CustomerServiceAI

ID Finanzhilfevereinbarung: 880954

DOI

10.3030/880954

Projekt abgeschlossen

EK-Unterschriftsdatum 16 September 2019

Startdatum 1 Oktober 2019

Enddatum 30 September 2021

Finanziert unter

• INDUSTRIAL LEADERSHIP - Innovation In SMEs

Gesamtkosten

€ 3 055 677,50

EU-Beitrag

€ 2 138 974,25

Koordiniert durch

CULINAR OY
Finland

Periodic Reporting for period 2 - CustomerServiceAI (CustomerServiceAI: Fully language-independent AI platform for automated and augmented customer service)

Berichtszeitraum: 2020-10-01 bis 2021-09-30

Zusammenfassung vom Kontext und den Gesamtzielen des Projekts

Businesses today provide customer service via one-to-one conversations between support reps and consumers. We believe that this is, secretly, the single source of almost all of the industry’s problems.

Consider costs, the primary pain point of every customer service operation. People are expensive to hire, retain, train and manage. Globally, we are spending over $900 BN on agent salaries alone.

If a business wants to reduce costs, they must accept a drop in customer satisfaction. Long wait times; inconsistent service quality; and stressed and poorly trained staff are all symptoms of this compromise.

Damage to customer satisfaction means lower revenues. Customers are walking away from brands following a bad experience - and they are telling their family, friends and audiences about it. In the USA alone, businesses are losing $1.6 TN every year due to poor customer service.

Software vendors reducing the need for customer service contacts (FAQs, knowledge bases) or making agents more efficient (better agent tools) are focused on the wrong thing. These efforts provide only a marginal impact and do not address the root cause; one-to-one conversations do not scale.

Virtual agents that can outperform humans in customer satisfaction will completely disrupt the customer service industry. They will change everything, from the jobs and software in customer service, to the way consumers and brands communicate.

That’s what CustomerServiceAI is, and that’s what we do.

Arbeit, die ab Beginn des Projekts bis zum Ende des durch den Bericht erfassten Berichtszeitraums geleistet wurde, und die wichtigsten bis dahin erzielten Ergebnisse

We can summarise the achievements of the project thus far by grouping them into 3 categories: artificial intelligence; software development; process improvements. To keep this concise, we will not describe the work performed to manage the project (largely internal meetings and administration) or to disseminate the results (strong focus on digital marketing and content creation, as COVID19 limited our plans around Events & Trade Shows).

Artificial Intelligence Overview:
CustomerServiceAI uses deep neural networks to provide the first fully language-agnostic artificial intelligence platform for customer service. Our goal is to become the world’s leading customer service automation platform. This mission puts artificial intelligence at the core of our value proposition. Our objective was to improve our core deep learning clustering technology by experimenting with different deep learning techniques and state-of-the-art methods. Our goal is to progress towards achieving fully autonomous clustering which will allow rapid onboarding of new customers, reducing time-to-value and increasing ROI.

Artificial Intelligence Results:
- Hired an interdisciplinary team of AI Researchers and AI Engineers
- Designed and implemented a prototyping and user testing platform for training dataset creation
- Ran and concluded experiments around the exploration of new clustering algorithms
- Launched AI Model Impact tool
- Implemented transfer learning to increase classification accuracy and reduce data requirements
- Developed iterative clustering methodology for unsupervised clustering at high performance

Software Development Overview:
Software development focused on two primary objectives.
1) To build an intuitive, easy-to-use web dashboard which customers can use to manage and track the performance of their AI deployments. Focus here is on developing a clear information architecture that regular non-technical users can use with ease.
2) To advance our integration framework, with a focus on developing integrations to three selected partner contact centers: Salesforce, Zendesk and LiveChat.

Software Development Results:
- Built one-click integration to Salesforce & listed on Salesforce AppExchange
- Built one-click integration to Zendesk & listed on Zendesk Marketplace
- Built one-click integration to LiveChat & listed on LiveChat Marketplace
- Launched Knowledge Base with articles on how to connect CustomerServiceAI to your contact centre
- Develop flexible integration architecture
- Launch Web Dashboard with 7+ core product areas, providing customers and partners an intuitive, self-service UI to manage the AI model and design automated conversation and process workflows
- Extend integrations with partner CRMs
- Begin development of Analytics product

Process Improvement Overview:
Our goal is to create training processes and materials to support the onboarding of customers to our Platform. Success in meeting this goal is guided by two primary objectives: Delivering the highest value for our clients in the shortest possible amount of time and ensuring scalability of our operations to optimally support our growth. As an important initial step, both goals have informed the development of the onboarding process for new clients over the past 24 months.

Process Improvement Results:
- Launched new customer onboarding process, with onboarding timeline and key milestones
- Launched new Service Plans
- Roles and responsibilities within our customer success team
- Customer training processes & materials (e.g. published e-book Guide To Great Conversational AI)
- Awarded Best Support on G2 as reviewed by our customers
- Reduced onboarding time from 6 weeks to 2 weeks
- Achieved 34% average automation rate across all customers, 3 weeks from kickoff

Fortschritte, die über den aktuellen Stand der Technik hinausgehen und voraussichtliche potenzielle Auswirkungen (einschließlich der bis dato erzielten sozioökonomischen Auswirkungen und weiter gefassten gesellschaftlichen Auswirkungen des Projekts)

Modern consumers are mobile-first, always online and used to instant gratification. They expect to receive customer service on their own terms, wherever and whenever they want it.

Digital consumers are also less loyal. 47% of Gen Z will walk away from a brand they love after just one bad experience, compared to 32% in the general population.

Higher expectations and rising stakes have increased pressure on support teams. In response, brands have ramped up hiring - the number of support reps in the USA grew 9.2% from 2018-19.

But, to control costs, they have also increased the demands that they put on their support staff. Agents have to respond quickly and resolve cases non-stop and are penalised harshly for any slow down or drop in CSAT. Managers have limited control over what agents do and so create large, pyramid structures to monitor and QA. Year on year, customer service is rated one of the most stressful jobs in the world.

CustomerServiceAI virtual agents meet the needs of the modern consumer. They respond instantly, round the clock, in every language. They give a consistent, on-brand experience across every device and channel. They are plugged into every business system and have the latest product information, without the need for training. They give managers full control over the experiences they deliver and an x-ray view into what is going on in customer support.

CustomerServiceAI has the potential to improve customer service for hundreds of millions of consumers throughout Europe. The platform gives brands control of their customer experience, driving better retention and thus revenues. And finally, CustomerServiceAI improves the working lives of the 14 million Europeans workings as customer service agents, who under resource scarcity and rising customer service volumes currently experience one of the most stressful jobs in the world.

ultimate-ai-logo-positive-rgb-72dpi.png