Secret Sauce Interview with Miuros
Miuros is an AI-powered solution that helps companies reduce all types of churn by leveraging their customer service data. When we interviewed Benoit Gagnon, CEO and co-founder of Miuros, for our secret sauce series, he had a lot to share about data-driven decision making, involvement of other teams in support, AI in customer service and much more. Let’s dive in!
CEO and co-founder of Miuros
Tell us a bit about career journey before you started Miuros.
I studied econometrics and computer science but started my career as a support engineer for a French startup that was developing business intelligence solutions. I then became the director of its customer service function. Later, we got acquired by Oracle where I stayed for quite some time as its global head of technical support in charge of data integration products. I then moved to Germany and joined the executive board of directors of Ableton. There, I was in charge of customer service as well as general services, internal IT, and web development. And then I launched Miuros. I live in Berlin now where I’ve launched an association for customer service professionals which currently has 140 members.
What made you start Miuros?
In the customer service function, absolutely everything can be tracked. We had the data but I did not have the time, skills, or resources to get the most out of it, and I was frustrated because I’m a data-driven person who likes to support my decisions with data. But I felt that I was making sub-optimal decisions. That’s how the idea of Miuros was born.
How does Miuros helps brands improve customer service?
Miuros Assist automates tasks such as categorizing inbound requests for better routing and faster replies. It also suggests the most appropriate canned response to use and even automates replies when possible. Miuros Insights, our analytical solutions, reveals to the management team actionable insights to improve the efficiency of the customer service operations. This feature, coupled with Miuros Review, our quality assurance solution, helps detect coaching needs at individual and functional levels.
We’re also working on an AI-driven feature that does a semantic analysis of all the conversations between customer service and end customers, and clubs of similar topics together. The resulting groups are then enriched with customer service metrics such as customer satisfaction, sentiment analysis, volumes, or handling times. These qualitative and quantitative insights about problems reported by customers are then shared with the product and marketing teams so that they can address the root cause of each of the complaints.
How far do you think we can go in terms of automating agent tasks?
However simple their issue is, some customers won’t try to solve it on their own. They go straight to customer service agents when they could have found the solution themselves. Here’s where automated replies and canned responses are of great help. But, when we looked at our customer data, we observed that personalized canned responses yielded better customer satisfaction. This showed that while we can automate a lot of things, customers should still feel like there’s a human touch. Because at the end of the day, humans are the heart of everything.
How do you ensure that you put the customers first at Miuros?
Coming from a customer service background, it is really important for me that Miuros delivers great service to our customers. We are super reactive to their requests. We also organize meetings with them regularly and listen to their struggles very carefully. In fact, several of our new features are born from the needs they expressed. We also involve our customers early in the release process. We also have some alpha and beta testers who give us feedback on certain features before we go live with the feature for the rest of our customer base.
Many companies look at customer service as a cost center. What’s your opinion: is it a cost center or profit center?
Customer service is definitely a profit center. For many customers, it’s the only interface they have with a brand both before and after turning into customers. In short, it is a gold mine and the most underused corporate asset.
Why do you think many companies still have the mindset that customer service is a cost center?
Customer service exists because there are flaws in products, services, and communication. But many companies do not understand that what they perceive as a “cost center” actually helps them acquire new customers, retaining existing ones, and even building their brand and reputation. A lot of these companies are product-driven rather than customer-driven. While they may be in a position to force things on their customers, it’s the customer-centric companies that become successful in the long term, especially in the digital world.
What do you think can be done to change the perspective of product-driven companies?
Measuring the ROI of customer service is not an easy task. But data may help.
Also, the customer service team has to show its value. In many companies, service reps are the ones facing customers the most, and they need to invest time in being the voice of customers to their stakeholders such as product and marketing teams. Since these teams are thirsty for feedback from customers, actionable insights from customer service teams help change their perspective. But customer service teams often struggle with reporting valuable insights as they don’t have the proper tools, skills, or resources to support them.
What data points can the customer service teams use to show their value to the stakeholders?
It really depends from one business to another. In industries where customer service is interfaced with customers before and after their purchase, they can show the number of prospects who turned into customers assisted by customer service. They can also show how often customers repurchase when they get in touch with the customer service teams.
What do you think are the key differences between B2B customer service and B2C customer service?
The main difference is the volume of questions which is much lower in B2B than B2C because of the size of the customer base. In B2B, the questions tend to be more complex With B2B, the problems reported often have a direct impact on the customers’ ability to deliver their own product and services, thus impacting their business. Meanwhile, B2C customers are very demanding since they spend their own money on your products and services and hence have high expectations.
How big is your support team and what kind of questions do you get?
We don’t have a separate support team since we are a very small organization of 15 employees. We are still at the size where you get to be super close to your customers. While we provide support through the classic email channel, we also have a dedicated Slack channel for each of our customers. This way, they can contact us anytime and get immediate help. Everyone at Miuros has access to these channels. While not everyone has the knowledge to provide answers, especially to technical questions, I believe it’s valuable for everyone in the company to know what’s happening and understand the problems our customers are facing.
You provide all hands support now. Do you think this will change once you start growing?
Definitely. Every company goes through different phases. In the early stages, you can start without any helpdesk solution like Freshdesk. But, as your ticket volumes increase, you need to be equipped with proper omnichannel tools like Fresdesk, Freshchat, and others. You need to have a proper set up to ensure that there’s no deprecation in the support quality. Right now, given the lesser volumes, we don’t use any, but I’m sure we would need the tools in the near future.
As Miuros grows, would you still consider all hands support as an option?
Yes and no. While everyone in the company should know what customers are struggling with, they need not be responsible for solving these problems. After all, we can’t expect every team to resolve support issues.
But, I believe it’s very valuable for the product and development teams to be involved in support to an extent, or to work very closely with customer service even if the support person is the one communicating with customers.
What metrics do you use to measure the performance of customer service?
Traditionally, there is customer satisfaction. But I believe less and less in CSAT metric because of the advent of self-service, chatbots, and automated replies.
The simple questions, the ones that usually generate the highest CSAT, don’t reach the support team anymore. They are either addressed within the product directly or through chatbots. Only the complex problems that take more time to be solved reach the customer service team. So naturally, the CSAT scores would be lower, portraying poor performance, which is not true. There’s another problem with metrics that are based on customer surveys. The problem lies in the fact that very often customers do not rate what you expect them to. Often, when asked to rate the quality of the service they received, they give you a rating for their overall experience with your products. Does it then help assess the performance of customer service properly? I don’t think so and that’s why I believe CES (customer effort score) is a better metric to use to assess the performance of customer support teams.
What kind of insights do you draw from the CES?
CES reflects the amount of effort the customer needs to put in to get their issue solved. So, a high score indicates that we need to make the resolution process easier and better for customers, including self-service.
What are some of the best practices to improve CES?
Make sure your customers don’t have to repeat themselves when their issue is passed from one agent to another. Similarly, when a customer switches from one channel to another, the agent handling the request should know about the past interactions. In terms of self-service, maintain a proper knowledge base where both agents and customers have quick access to the right solution. If you implement chatbots, make sure that the bots would quickly detect when it can not address a problem and hand the case over to a human.
How important is data to the success of a product and its support functions?
Data is very important, but we have to be extremely careful. Data is not everything.
When you’re driving your car, you have the counter to make sure that you’re not speeding. But if you only look at the counter while driving, you’re going to hit a wall. It’s the same with data. If you look only at the data, your product is not going to succeed. You need to put everything in context and data is there just to support your decision.
What does the future of AI look like according to you?
AI will be everywhere and you won’t even notice it. To me, there is a big hype around AI which will soon be gone because people will realize that AI is a tool to automate simple things so that people can focus on things that are more sophisticated. While AI will be everywhere, it will not be so strong that it will replace humans and eliminate jobs as some people foresee. I believe we are far from that stage of AI.
Many companies are investing a lot in AI and the kind of technology that powers AI. Do you think it is smarter to invest more in AI or in improving our teams’ skills?
I think you have to do both because AI can never replace humans. AI can take care of simple things, and maybe over time, things that are complex too. But humans will remain at the heart of everything because AI mimics what humans are doing at a much larger scale. And of course, you have to invest in your people to develop their skills because the most complex problems will still have to be solved by humans. You need to develop their soft skills and technical skills, so they can handle these problems well.
At the end of the day, it’s all about empathy and AI is not that great at it yet.
Support teams usually struggle to handle negative customer feedback. How do you go about responding to negative feedback, and what is your advice to agents?
First of all, support reps should always remember that they are not responsible for the issues that customers report. And customers often understand that the support person is just the messenger. So while facing negative feedback, the support team needs to focus more on acknowledging the problem with empathy rather than apologizing for it. Customers don’t get frustrated when they see that the company understands their problems.
It is also very important to understand the reason for the negative feedback. We need to check how the issue came up in the first place, who needs to be engaged to resolve it, and how we can address it. Yet sometimes, we can’t do much about certain negative feedback and that’s totally fine.
At Miuros, how do you nudge your customers to adopt new features?
Two kinds of new features are released. One is what we create believing that it will be useful to customers. The other is something that has been requested by multiple customers.
For features that fall under the first bucket, we make sure that they understand the pain point it solves, what they can get out of it, and how they could use it. If the adoption rate doesn’t improve, we try to find why they’re not using it. If the feature solves a real problem and is not used, it is often an indication of a UX/design issue.
What do you do to make your customers understand about the feature?
Since we are a small company, we are very close to our customers. We talk to them on a regular basis where we try to understand their concerns and the new problems they’re trying to solve. We use this is an opportunity to promote new features.
Promoting features within the product using videos that cover concrete use cases often helps. Showing how they can benefit from the feature using their own data in their own context is even better.
What do you do when customers tell you that a feature isn’t very useful? Do you roll back the feature?
Actually, customers don’t always voice their opinions even when we talk to them. This is where you can use tools to view how customers are using the product, what features they use, and how they use them.
But when they express that a feature is not very useful, we first need to understand what “not very useful means”. Does it mean that it’s not solving a problem or does it mean that usability needs to be improved? Sometimes the timing is just wrong where people are not ready to use specific features or products.
What’s your take on culture and its impact on the organization?
I’m a strong defender of a multicultural environment. When a business starts getting end customers from around the globe, it is very important for them to be familiar with the different customer cultures. In Miuros, we have people who come from Spain, Italy, and Latin countries whose cultures are very different from the ones you observe in countries like the Netherlands or Germany. We will be able to interpret the customer’s communication style only if we are aware of these cultural differences. While uniform culture may work in specific market segments, you need to bring diversity within the group for further growth.
Do you like the customer support of any other organization? If yes, why?
I was once delighted by the customer support I received from Airbnb. When I was facing an issue, the support conversation started with email where they said they would call me. But, due to a technical glitch with the phone line, I didn’t receive any call. I was really upset because it was already 9.55 pm and their support hours would end at 10 pm. Just when I thought my problem wouldn’t be fixed, the person called me back at 10.15 pm once the issue was fixed on their end. I often recount this experience since the company went the extra mile to solve my problem.