“”

AI Solutions in Customer Service: Q&A with Simplr Data Scientist Damien Thioulouse

TwitterFacebook

Simplr was recently named “Best AI Solution for Customer Service” in the AI Breakthrough Awards. The award recognizes innovation, hard work, and success in Artificial Intelligence (AI) and Machine Learning (ML). 

But what does being “the best” actually look like? With so many automation and chatbot vendors in the market, what should CXOs and CX leaders be looking for when incorporating AI into their strategy?

We sat down with Damien Thioulouse, Simpr’s Head of AI and ML, to discuss the current state of AI in customer service. 

What does it mean to be named “Best AI Solution for Customer Service”?

For me, this award is validation that we’ve made the right design decisions over the years and have established a strong AI culture within the organization. I’m extremely proud of our AI team at Simplr and grateful for the support from our operations and engineering teams. 

What should CX leaders consider when applying AI to their customer service strategy? 

There’s a lot to consider when thinking about using AI in customer service. Here are my five recommendations for CXOs and CX leaders:

  1. Ironically, I would tell them to forget about AI and focus on what specifically would make their customer service better and faster. Depending on the context, this might not even need any advanced AI!
  2. When building a solution, don’t get obsessed about pushing the so-called “accuracy,” but try to adapt the product or operational processes to maximize model impact in its current state and iterate from there. You might realize your “accuracy” might not reflect what you had anticipated.
  3. Know how to pivot quickly. In the customer service space, there are too many problems to solve to afford to drag projects that “might” work at all or that “might” get better. I recommend using the Pareto rule: 20% of the work delivers 80% of the value. Better to spend extra time on a new project than on the same one.
  4. Overall, AI shouldn’t necessarily be seen as an automation tool, but as a way to empower your in-house agents to deliver world-class customer service. Simplr recently uncovered that only half of large and enterprise businesses are using AI to help agents, while nearly 75% are using AI for customer-facing interactions. At Simplr, we help our agents connect with the customers by adding personal connection prompts and thoughtful touches to their messages or guiding them through the resolution process.
  5. Customer service is dynamic and customer inquiries constantly evolve over time. AI can help you keep your knowledge base optimized or up to date. For example, by analyzing its usage from agents, some types of inquiries might display unusual patterns signaling agent confusion, or some types of inquiries being resolved in multiple ways indicating duplication or even inconsistencies related to dynamic aspects of customer service. AI is great at learning from the past, it is important to consider that traditional approaches will have issues adapting and will be lagging behind. At Simplr, we are ensuring our solutions can autonomously learn from constant & current flows of data, satisfying our performance requirements and ensuring AI remains current at any point in time as the knowledge base changes. 

Why do you think Simplr partners love the Simplr platform? What makes it so special?

I believe we’re building the best possible Machine Learning solutions on the market. Our internal matching system always ensures that a given inquiry gets routed to the most appropriate agent in the shortest amount of time. Given our expertise in the space and data assets we have built, we have been making our platform so easy to use that it makes it hard for our agent to resolve a customer inquiry incorrectly.

On top of it all, our solution has been able to increase conversions and create higher-quality customer experiences. Simplr prides itself on getting “better and faster as we get bigger ” – everything we’ve built is built to scale so we are confident we will be able to maintain our excellent service quality and speed as we grow.

Some might not know this, but Data Science is very much a team sport. Can you walk us through a typical day with Simplr’s data science team?

There is no typical day in Data Science 🙂 Joke aside, every problem has its own challenges which might be related to data, product, infrastructure, and internal processes. Therefore, data scientists at Simplr will adapt to what works best for them and their current projects. Data Science is definitely a team sport and keeping a good understanding of each other’s projects is critical as peer feedback and exchanges of ideas are invaluable. 

Data science relies heavily on creativity, collaboration with diverse internal stakeholders, and adjusting/suggesting project requirements. Therefore, we do not rely on a “Data Science Routine” BUT we always keep in mind our foundational questions in daily activities, including:

  1. Why am I working on this?
  2. What would be the business impact if my solution was 100% accurate?
  3. How soon can I test the most simple prototype?
  4. Would this solution design still work if we grow 10x?

What excites you the most about where AI is headed? Anything you’re watching closely? 

I’m excited about how good AI is becoming, particularly that it’s learning from less data. Generative models have become more and more trendy and made significant progress over the past few years. 

I have been very intrigued by their application to customer service. While the generative model works astonishingly well on general “happy path” scenarios, they are hard to interpret and troubleshoot. Additionally, there is currently no guarantee that they will gracefully handle cases deviating slightly from the norm. I can’t wait to see how this is going to evolve.

In my opinion, one of the main bottlenecks of AI in customer service (especially for generative model approaches) is about feeding a set of human-generated internal CX policies (dynamic – as mentioned above) and making sure the model can safely understand them, incorporate them into its “learning,” and output them in a grammatically-correct customer-facing response. I am excited about seeing how this evolves.

Simplr is the conversational experience platform for the NOW CX era. The company’s AI-enabled platform unites chatbots and human assistance to deliver instantly scalable premium pre-sale shopping assistance and customer support. The result for Simplr customers is best-in-class experiences throughout the consumer journey, increasing loyalty, satisfaction, and revenue. Simplr is funded by Asurion, which continues to support its growth.