Home > Columns > Executive Interviews

Loris Executive Interview

Please click the image below to view the ebook

Sheri Greenhaus, Managing Partner, CrmXchange and Etie Hertz, CEO, Loris, discuss the creation of Loris and its benefits for companies with customer facing agents. You can find Loris in booth 723 at CCW 2022.

How did Loris start?

Loris was founded on the idea of bringing more empathy to the world, and we started in customer service because it’s just about the hardest job out there and is really in need of that empathetic approach.   The initial idea was to build empathetic training videos and test the concept as fast growing companies needed more guidance on how to navigate the rise of real-time digital channels and trend towards messaging, SMS, and chat communications.

After a year and a half of testing, we noticed that while the content was valuable, there was an issue: like most training based content and platforms, while it is helpful on Monday - by Wednesday nobody remembers what they learned. 

It was at this point, in 2019, Loris pivoted to being a tech company, with an aim to have interaction level impact.  We built a natural language processing based software solution that can integrate with existing customer service platforms in order to guide agents in real time with our best practice language suggestions.   Loris exited stealth mode in 2021 as a powerful AI solution in the form of a chrome extension.

Is this with an agent, or without an agent?

It’s only with an agent.

Loris sits on top of most existing platforms agents use day to day such as Zendesk, Salesforce and Live Chat.  During a chat conversation, within a second, we absorb all text messages, suggest techniques and language for customer facing live agents. We increase the efficiency of the agent, while ensuring high quality conversations. 

Does it guide them?

Yes exactly.   100% of incoming messages get scored based on customer sentiment in real time.  Based on what we see, we tell the agent the best way to approach that customer. We don't tell them what to say, we give them suggestions. If they agree, they click to send the text. If the agent does not agree with what we sent, they can edit or X out.  Agents benefit from suggested responses and proven techniques to ensure brand tone and policies are upheld - while maintaining a distinctly human touch.

Is your differential the fact that you're picking up the sentiment?

It’s a couple of things. What we do now was not technologically possible a year and a half ago.  To ingest that amount of data, analyze customer sentiment and intent, and instantly suggest a response was not an option before Loris.

We have some of the best data scientists in the world and that is continuing to expand solutions we’re able to bring to companies of all sizes - solutions that have until now been reserved for large enterprise companies with in-house data science teams. .

We started with language suggestions and now we’re taking on policy suggestions. As an example, one of our clients spends a few million dollars per year on coupons.  They did not know if couponing was working.  If the customer was upset, the agent may give the customer a $20 coupon.

One of our data scientists said ‘if we predict the language, why can’t we predict dollars’?  We know the sentiment, we know the intent, we can suggest a range of dollars, or we can give them some guidance. Maybe the customer doesn't want money, maybe they just want their feedback to be heard.

We’ve started to further roll this out to our clients with churn issues. It looks like we are going to triple the number of customers that they're able to save with their agents.

 

 

How does the system know what I want? It is profiling people like me and my demographic? Is it looking at what I've done previously?

We don't store any personal identifiable information. Everything is scrubbed. I can only see what you show me in your text. I might be able to pull something like – you’re customer 123, you've been here 3 times in 24 hours, you're probably upset about the same thing.

The agent would know who I am.

Exactly. The agent in a loop helps. The agent can bring together any next best action data (such as past purchase, demographic, LTV, etc.) and apply it to Loris suggested responses that cover the rest of the equation - response best practice language, tone and policy recommendations.   The agent has flexibility and control to see the native customer profile data available and augment our suggestions, which all in all is what maintains a personalized and human touch.

How is the agent trained?

We have built a UI that looks the same regardless of the platform. It is basically a Chrome extension on the front end for the agent, and it's not forcing them into another system. Within 3 clicks, they get it. We tell them ‘here’s how to navigate the suggested language options and here's why our suggested language will benefit your day to day workflow’.  Agents pick it up quickly. It doesn't require a lot of training.

Can the agent can make a decision on what to give to me as a customer?  Since it’s identifying me, it must be based on something.  Does the system know psychologically this is what people want based on these words?

Ultimately, companies define policies and agents uphold them. However, typically those policies are laid out in complex situation based matrixes that make compliance a challenge - especially when support demand is high or customer sentiment is negative. 

As an example, you’re connecting riders and drivers.  There is a problem if the ride doesn't show up or vice versa. Someone's yelling. How do you know what that is worth to someone?

Should I coupon you? 

Should I give you free rides? 

Should I move you up a tier in support?

There’s a lot of things the agent needs to assess for every response within the framework of what this customer is worth. Now, think how hard it is for agents who are typing three conversations at once and English may not be their first language. It gets exhausting. They have a hard job to begin with and now the job is getting more and more difficult.

We make all of this easier for them by suggesting appropriate language, tone, and policy while giving agents the flexibility they need to personalize a response. 

 

 

Do you think this concept might not have worked as well before the pandemic? People were calling in for a very simple thing. Now, the simple things they can self-serve. As an agent, I can make a decision to give this person 20 dollars as opposed to free rides because I understand that is what this person would want. The agent is given more flexibility to decide what is right for the client.  The days of calling to change a password are over.

It's a great point. There were a few societal shifts. More transactions moving online as a result of the pandemic, people expecting more from brands, and an influx of companies that look the same. Take Uber and Lyft... it's the same car with the same driver, but it's a different company. What differentiates those two companies?  With Fintech there's like, 50 companies. They all do the same thing. What’s unique about you aside from your logo? It's basically the service. This level of expectation of service from agents is just going to continuously increase. The agents have to step up to meet the demand.

In the future, do you think customers won’t need human agents because AI is going to be smart enough that you won't know if you're talking to a human or you're talking to a machine?

No, because people in general don't want to talk to a machine and brands who turn heavily towards automation and deflection will not be able to compete. I think the next 15 years you’re going to vacillate between a human and an AI assisted human and you won't know the difference. You're going to basically be talking to the human the whole time, and they're going to get AI assistance. Maybe at the beginning AI automation will  ask you context based questions before routing to an agent, but for at least the next two decades, AI assistance technology, like Loris, will enable brands to deliver quality, human based service at scale..

Are there certain industries you're finding that would use your solution more than others?

In the last year and a half or so, fast growing companies, marketplace, Fintech, ecommerce, delivery have been the first to adopt Loris but it’s going to get much broader, especially as a result of the current state of the economy. Companies have to find ways to be more efficient. Now, CFOs of public companies have to find a way to cut costs. In the last 30 days we are seeing tremendous demand to do things more efficiently.

Think about how expensive a customer service team is.  The manpower, QA, the coupons, the cost of customer churn.

There's also a unique opportunity here. There are fewer sales agents as more transactions move online.  The only time you interact with the brand is when something bad happens.  That's an enormous opportunity for brands to say, let me provide a great experience – and perhaps an upsell too.

What do you think will happen to agents, will there be terminations?

What's unique here is the headcount churn in this department tends to be higher than all other departments. The life expectancy of an agent is not high, so you don't necessarily have layoffs. 

Think about the extra people you need for the holidays.  Do you need to hire a lot more people or can you figure out a way to not have to temporarily bring on an extra 200? Is there a way to keep it streamlined? There are so many things that can be optimized. Our main goal is to make customer service personnel more valuable, more efficient, and more influential within the organization.  Loris can help you do that, help you save dollars and potentially help you increase the direct revenue impact of a customer service team. There’s a lot of things that can be optimized here, that doesn’t necessarily mean you’ll need to fire the extra 200 people.

When companies come to you, do they have something in mind that they want to optimize or do you have somebody from the organization sit down, have a conversation and say this is where we should start?

It varies.  The category is new. It seems like we're in a crowded space, but no other company is doing this. A lot of people come to us thinking they need to do something with AI, but they are more familiar with deflection and chatbots.

After they sit down with us, there are two paths that most of our clients take to get started with Loris.

One, is they want immediate access to first message and last message sentiment and visibility into 100% of customer conversations - not just those with CSAT survey results.  They love our dashboard that can reduce QA time, costs and more easily surface the voice of customer insights that can inform everything from support policy to greater product decisions. 

 

 

The second is they want immediate workflow efficiencies with our real-time agent guidance that is either language based (especially for those clients with outsourced agent teams) or policy based (when they are hoping to test new policies, monitor agent policy adherence more easily, and reduce customer churn threats). 

What workflow are you optimizing? Is it the customer journey?

Yes, we’re optimizing the customer journey for any aspect that involves human support agents.  The latest policy suggest guidance that we're releasing could be game changing. And it’s just the tip of the iceberg.  We’re going to be able to optimize almost every single action agents take based on intent and sentiment, do it live, test it and make sure that everything is optimized.

Is it telling the agent they can choose from a menu of things because of the work flow you've provided?

Yes.

Part of the issue that I see with some companies is being able to distinguish what you are doing with AI. The fact is that AI is fine as long as it works. Companies want their problem solved. 

The problem with most companies is that they constantly come up with new policies and they have to make sure that agents are adhering to the policies, and they have to stay within budget. To do so, there's a lot of things that have to be tweaked and rolled out to agents constantly to determine what works. If you can let us do that for you, and then surface the results to decision makers, they can spend more time being the pilot of this cockpit, deciding which knobs to turn as handle rolling out guidance to agents on an interaction level basis, in real time. . Forget how you deliver it, if it’s AI, NLP, we’re solving for a real problem customer support leaders deal with every day.

If I'm coming over to your booth at a conference, what are the questions you would ask me?

What are your biggest pain points? What are you trying to solve for this year? What's top of mind? What are you looking for? What have you seen work? What do you think could work for you? For us, it’s all about identifying if you have a problem we can help to solve. 

We also recommend that if you are looking for new solutions to directly help with customer churn threats, we’d love to walk you through our policy suggest features and calculate potential impact it could have on cost savings for your bottom line. 

For companies that are getting started with you, what are the steps you take them through?

Onboarding with us is really easy. It goes through a quick tutorial.  Agents usually only need half an hour, team leads maybe an hour.  The Loris Chrome extension is downloaded and then they're off to the races within a couple days without any coding or tech support required from their internal engineering resources.