Calldesk is a Conversational AI company that helps large organizations to automate repetitive calls and focus on high-value interactions with their customers.
- The story behind the Calldesk’s idea.
- Calldesk’s focus on CX – ‘Caller Experience’.
- His company’s philosophy when it comes to hiring technical talent.
What is your background and what made you start Calldesk?
I am an engineer in Computer Science, graduated from the Ecole Polytechnique in 2007. I also have a Master’s Degree from Stanford University in Electrical and Electronics engineering. Since graduation, I have always been an entrepreneur, and calldesk is -already- my third company. With my previous company, Wozaik, I started working on an emerging technology: conversational artificial intelligence.
After selling Wozaik to the Solocal group, the French leader in digital services for local companies, we left for a 6-month boat trip with my wife, sailing around the world. During the trip, she had a big problem with her telephone operator, and as soon as we had network, I tried to call her customer service to find a solution. The experience was a real purgatory: with each call, the queue time exceeded 10 minutes, I had to explain my problem several times, I was sent back and forth from one caller to another… It finally took us 6 calls, spread over more than a month, to find a solution.
That’s when I realized the fundamental problem in customer service today: the lack of time to deal with all the requests. And that’s when I started wondering if conversational AI might be able to automate some exchanges, to free up customer service resources. Calldesk’s idea was born.
What is Calldesk’s mission?
Calldesk enables customer service of large organizations to automate repetitive calls, in order to focus on high-value interactions. With the right support, we believe that even a dissatisfied customer can be turned into a brand ambassador. We want to free call center agents of their repetitive tasks, so that they have more time to respond to complex customer requests and problems.
Who are Calldesk’s customers and how do you create value for them?
Calldesk already handles dozens of millions of customer service calls for some 30 top-tier companies. Our solutions deliver financial ROI (handling a call with a voice agent costs 5 to 10 times less than with a human), and improve caller satisfaction by removing irritants on their customer journey: queue time, annoying IVRs (“press 1, press 2…”).
What challenges do they face, or will they face in the future, that Calldesk can help them overcome?
Despite the development of digital channels, the number of phone calls received in customer relationship centers has never been so high. At the same time, customers now want an immediate and personalized response to all their requests. For customer relationship centers, the expectations are enormous!
That’s why the productivity gain associated with artificial intelligence offers such potential for customer services. By being more productive, companies can deliver a better experience and win the battle for customer retention.
Are most of your customers in a particular industry, and if yes, why?
We work with all companies that receive very large volumes of calls (from 2,000 per week), and that generally have significant B2C activities. The majority of our clients are therefore divided into four verticals:
- Insurance, banking and financial services (claims, account management…)
- Utilities (invoice follow-up, organisation of an intervention…)
- Logistics and delivery (parcel tracking…)
- Transport and tourism (booking a ticket, getting an invoice…)
Tell us about Calldesk’s team.
The Calldesk team now counts about 30 people. Half of them are dedicated to R&D, and are experts in artificial intelligence and machine learning. The rest is dedicated to supporting our customers, to help our partners make the best use of voice agents and build tomorrow’s contact center.
How many data scientists and/or machine learning engineers do you have in your team?
Apart from the product team, we have around 10 purely tech profiles in the team.
Which technology stack are you experts on and are you leveraging the most?
We extensively use Node.js in our back-end infrastructure in order to benefit from its asynchronous processes, event loop and incredible speed. Our platform is fully hosted on AWS that provides efficient services: Fargate, DynamoDb, RDS, SQS, Kinesis, Lambda… The whole architecture is serverless and can easily scale on-demand, whatever the number of calls we have to proceed. It’s as if we were building a contact center with one agent at 10:00 AM, and a hundred at 10:01 AM!
Our AI algorithms are developed in Python and are executed in Node.js processes in order to be easily integrated within our architecture.
We also use React.js to develop the calldesk studio, the interface that enables our users to create and deploy voicebots in production in a few minutes. It makes the AI usable by non-tech person!
What skills do you look for when hiring data scientists and/or machine learning engineers?
We are mainly focused on soft skills because we think that great talents can quickly learn new technologies, and we prefer to hire human and mindset rather than hard skills. We are looking for team players eager to learn, ready to make their teammates grow and progress together. Engineers at calldesk always help each other, and we strive for keeping this mindset in the team.
We are mainly focused on soft skills because we think that great talents can quickly learn new technologies
Our engineering team is divided into two sub teams. One team is focused on callers and has 2 core values : “no caller left behind” and “deliver WOW experience”. This team’s job is to ensure that every caller that has a conversation with a calldesk voice agent will have an amazing experience (in terms of latency, understanding, exchange…). The other team is focused on the calldesk studio, which is the tool used to create the voice agents, monitor phone calls… The core values are: “plug and play apps”, and “constantly improving bots”. Its goal is to provide pre-trained voice agents that are easily customizable and allows users to create new voice agents in minutes, as well as provide tools to improve existing agents: A/B testing, real-time analytics, intents classification analyzer…
What are the biggest challenges that you are currently facing at Calldesk?
The use of conversational artificial intelligence over the phone is a very important challenge, both from a technical and business point of view.
At the business level, customer satisfaction remains the priority of the companies we work with. We need to make sure that automating certain calls will not degrade their callers’ experience. We have therefore developed a real expertise in CX – caller experience, in order to create a fluid and natural experience in discussions with voice agents. For this, we have developed several features: A/B testing, automatic measurement of post-call satisfaction, collection of user feedback…
At the technical level, we need to build a technology that can successfully meet these business objectives. To create a bot capable to conduct a natural conversation with a human is to act as a conductor: we must ensure the best comprehension rate, reduce latency to provide a satisfactory answer in one to two seconds, be able to handle requests out of context (e.g. the caller asks to repeat the question or wishes to be transferred to a human)
What are Calldesk’s biggest achievements in the last 12 months?
This year, we are proud to have successfully launched our partnership program, to integrate ourselves into the vast ecosystem of customer relationship players. AWS, Microsoft, Nice InContact, Softel Communications… are some of the companies that are increasingly interested in our technology, and train their teams to create and deploy voice agents to their customers.
This required a real technical investment, in order to create an ergonomic platform usable by non-technical users: this is the role of calldesk studio.
Walk us through Calldesk’s products, and what problems they solve.
All our products are voice agents, which will be able to partially or end-to-end automate incoming calls. The use cases are very numerous: a voice agent can answer the phone call and direct the caller to the right competence center, it can ask him questions to identify him before transferring him to an agent, it can fully solve his problem and trigger the sending of an email or a sms with the information sought by the customer…
Which product are you currently focusing the most on?
Many companies have understood the potential of natural language for telephone interactions, and are tackling a technology that hasn’t evolved for 40 years: IVRs (interactive voice responders). Their objective is to deploy a routing voice agent, capable of greeting callers without queue time, 24 hours a day, and understanding their need to guide them in the right direction.
Which product do you think is creating most value for your clients, and why?
The voice agents that create the most value are able to automate the most tasks, and thus handle incoming calls from end to end. Even with today’s technologies, the challenge is great; but this allows companies to offer a true voice customer space, where customer autonomy becomes the norm. The customer is then able to solve his own problem.
Do you customize your product based on each customer’s needs? How?
Yes, each voice agent is tailor-made to the customer’s needs, thanks to our team of Customer Success Managers. The set-up can be done in a few days thanks to two unique innovations: our studio, which easily allows non-technical people to create bots thanks to its ergonomic interface; and a system of modules that can be duplicated from one voice agent to another (for example a “first name identification” block).
What channels do you provide chatbots for?
We have found that for our customers, the telephone still represents 10 to 20x more volumes of requests than chat. We therefore focus solely on the telephone channel.
What are channel-specific challenges and/or opportunities for enterprises to leverage bots?
On the telephone channel, automation is for now reserved to large companies. Our approach to identifying sources of value consists of analysing all incoming call flows to determine the solicitations with the highest volume and the lowest added value. On these channels, the company generally wants to remain available by phone, but reduce the resources dedicated to call handling. This is where voice agents are positioned.
Which internal processes are you helping enterprises automate with bots & Conversational AI?
We use voice agents to automate end-to-end calls, and also to reduce the amount of data entry tasks dedicated to humans. Typically after a call, the agent needs 30% more time to type the information related to the interaction into the CRM. These tasks can be performed immediately by the bot.
How do you measure the performance of your bots?
We monitor the comprehension rate and response latency of our voice agents on a daily basis. The most accurate indicator comes directly from the customers themselves: at the end of each call, the bot asks the caller for his satisfaction with the interaction that has just taken place, and can also retrieve a verbatim. This allows us to benefit from a clear KPI and concrete recommendations to improve the experience of our callers.
What opportunities do voice-enabled interfaces bring for enterprises?
Voice is the most natural channel of communication. It allows companies to be more present for their customers and to process their simple requests more quickly. But we must keep in mind that customer relations is a complex business, where agents are not destined to disappear: on the contrary, AI must allow them to increase their skills, in order to offer an ever more relevant response to the customers who need it most.
How can enterprises get started with their conversational ai strategy?
Conversational AI is not a gadget: it must meet a clear need expressed by the business. In general, awareness begins when the company realizes that it receives more calls than it can handle. In this case, after having identified the call flows to be automated, a POC can quickly be deployed on a limited call flow, in order to estimate caller appetence and estimate the ROI of the solution.
How can enterprises prepare for the changes in the industry brought by Conversational AI technology?
Companies already need to think about how they want to be present to their customers in the future, and their preferred communication channels. It is this analysis that will determine which channels automation is most relevant.
What are the biggest opportunities for enterprises to leverage Conversational AI?
Consider the customer experience from end to end, and know how to mobilize AI only when it is relevant within that journey.
How can enterprises leverage Conversational AI to make their customers happier and more satisfied?
Artificial intelligence’s ability to be available immediately and regardless of the number of solicitations is already improving the customer experience. Moreover, it is especially at the level of personalization that companies can invest: thanks to its ability to immediately identify the customer and retrieve the history of the relationship, the voice agent can create a true conversational experience and help each customer feel valued by the company.
Which Conversational AI-related technology trend do you think will have the biggest impact in your industry in the coming years?
The main challenge over the next few years is to develop collaboration between humans and AI. In companies, employees are the only ones who have the emotional intelligence that AI will not have for a long time. Mobilizing this emotional intelligence at critical moments in customer relations, while mobilizing the productivity gains allowed by AI, is an important challenge for the future!