Home > Columns > Executive Interviews

ASAPP Executive Interview

Click image below to view the ebook

Sheri Greenhaus, Managing Partner, CrmXchange recently had a discussion with Macario Namie, ASAPP Chief Strategy Officer, on the benefits and advantages of employing an ASAPP solution in the contact center.

What can you tell us about ASAPP? 

Our primary purpose is to make people more productive at the work that they do. We believe that artificial intelligence and all its resulting disciplines can be used to augment work in ways that can really transform productivity.

What we've witnessed, particularly in the CX and contact center domain, is that AI has been largely relegated to – “I have a bot.”  The bot was largely used to keep the customer away from the agent. We don’t agree. We believe that the use of AI machine learning can be used to augment the work that people do to increase productivity.

During our webcasts we often ask if their agents are coming back to the office, hybrid, or work from home.  The majority say they are hybrid or work from home.  Almost no one mentions everyone is back in the office.

Yes, we are seeing the same thing … it's the new reality.  Organizations need to rethink their training, coaching, and culture. We hear some companies that believe it is temporary, but I don’t think so.

The move towards hybrid or permanent work from home has really shifted from ‘this is just an emergency temporary procedure’ to ‘this is the new reality’.

How is ASAPP making agents more productive and how are you different from companies say that claim the same about their solution?

We work to understand how agents operate in their daily roles, meaning how do they make decisions? How do they understand what are the right responses and the right ways to solve problems? We use technology to observe and learn how people actually work.

We learn from what the best agents are doing. We then take what is learned and provide that back in the form of suggestions on what to say and do during a live conversation for all other agents.  We observe the best agents and use that information to train our machine learning models.

When you say you observe what the best agents do, how do you define best? Does the customer decide what's best?

The customer and company decide the best.  That’s the beauty of machine learning, you can actually handle any degree of complexity. It could be best at sales, sales of additional accessories, customer service, or more technical support. It could be a combination of the balance between handle time and CSAT scores or the balance between handle time and first -all resolution (FCR) resolution. Our customers define what best really means.  

It sounds like your customers define the metrics and goals for the metrics and look at the agents that are best meeting those metrics.

That's exactly right.  We're learning from what those agents are saying and doing and translate that back into action.  The system is constantly learning.  What is best can change over time.  We've designed our technology to have a continuous loop, retraining the models weekly if not daily.  As we learn what best looks like, ASAPP AI services adapts.

Does the system provide a guide for the agent?

It differs slightly based on whether you're engaging in a voice call, or if you're engaging in a chat, or what ASAPP AI service the customer has chosen to deploy.  AutoCompose is an offering for chat agents. During the conversation agents are given informed, and tailored choices for a response.

Most organizations have a knowledge base.  The highest we’ve seen is about 20% of messages come from that canned response library. On a typical chat exchange an agent will send on average 10 messages. Two out of those 10 messages will be coming from the canned response.  Those tend to be the simple responses, because that's where you can program in advance, such as “thank you for contacting,” “how can I help you?”  Simple things that are used the most.

Agents have to write out a response for 8 out of 10 messages.  Typing, on average, is 21 seconds.  AutoCompose looks at the conversation in real-time.  It sees what the customer is saying and suggests to the agent a minimum of three responses.  What to say changes depending on how the conversation is progressing.

We can automatically complete what they are typing, fix grammatical and spelling errors during the course of the conversation. Agents that use ASAPP are getting 7 out of 10 messages coming from the system.  They are getting information from that hard middle part of a chat conversation. We give the agent suggestions how to respond, but the Agents are in charge.  All this time, the system is learning what Agents are doing.

ASAPP is relieving the workload on agents to make them more productive. The agents go from an average of 21 seconds to respond to a chat down to 3 seconds.  The quality of response standard is also much higher. There is now consistency among an agent pool. Agents are using phrases that we know are accepted by the organization. We have similar capabilities for voice agents too.

Tell me about AutoSummary.

AutoSummary is an API that will summarize the call. Wrap up notes are done automatically.  The summary can be read and analyzed by an agent and for data purposes.

When the agent is chatting, does the system pop up different responses? Does it provide the best response first?

We limit choice to 3 responses as to not overwhelm the agent.  The system will present what it thinks is number 1 on the top.  We provide responses that have slightly different semantic meanings.  We understand the context of the conversation, how it’s progressing and make changes.  Our ranking techniques present what we think as best and also provide two alternatives which are slightly different in terms of their meaning.

How does JourneyInsight help?

With JourneyInsight, it is the agent’s journey ASAPP is analyzing; using machine learning to understand at scale what agents are doing and saying. Speech analytics analyzes the conversation, desktop analytics analyzes click streams, and JourneyInsight brings them together. We look at the entirety of the customer and agent conversation, everything that the agent did during that conversation, at every point, as a conversation progresses.

The conversation allows us to understand the intent of the customer call.  We know enough about the agents to know what team they’re on, what their skills are, etc. Now, we can start to look at the different systems the agents are using, number of clicks.  Analyzing the agent path lets us identify operational inefficiencies in the system, so companies can establish more efficient and productive processes.

What is ASAPP’s analytics?

Our research team developed a degree of analytics we haven't seen anyone else do by bringing together: Words. Intent, and Context.  We care a lot more about the agent experience; what’s their points of frustration? How many clicks do they have to take, etc.  All of that is analyzed.

Our process starts with observing. In addition to analytics, we use what is observed for real time guidance down the road. RPA technologies today are designed to understand process flow. We create a spotlight on workflows throughout the contact center. 

What type of reports alter management thinking?

We find most organizations have no idea how efficient or inefficient their processes are.  JouneyInsight finds the holes and provides reporting that illuminates the inefficiencies.  For example, we have standard reports that look at which agent teams are actually efficient and which are not.

What do companies find most helpful?

We see two big actions taken from reports that we provide:
1. An analysis of what self-service automation should be, or what the automation roadmap should look like.  They have a better understanding of what’s causing a lot of volume where the customers can serve themselves. That change takes call volume right off the top.

2. Every organization we work with has some kind of roadmap for the agent desktop environment.  We are influencing that road map by illuminating where people are getting stuck, where people are hovering and what is causing agent frustration. 

What is the profile of a company that can use your solution?

We serve companies that have anywhere from 200 agents to 50,000 agents, as labor is now a significant portion of costs.  We ask companies if they have tapped out on automation opportunities and recognize that they have hit a plateau on what can be automated. We ask if they have attrition problems, issues with keeping employees happy, or if their productivity metrics are stagnant.

For the majority of companies that we talk to the answer is yes.

How do you help with attrition?

The attrition rates are much higher than what they used to be.  We try to speed-up and increase the agents’ time to proficiency and provide a better experience to reduce the stress factor. We get new hires more productive faster. 

Agents generally want to help people. They're service oriented.  They may be getting yelled at all day and don’t get support from their existing systems. If we can eliminate that stress inside the organization, the agent will show empathy and serve the customer.

Looking at the microcosm of an individual interaction, the stress level goes up tremendously once they understand what the customer's problem is.  They're trying to navigate a labyrinth of internal systems and processes to solve the customer’s problem but they are stuck. Now, repeat those 20 times a day, 5 days a week, it's just too much.

That’s where ASAPP can help. Three years ago, the only conversations discussed was how bots can eliminate the human element. Companies are now realizing that bots don’t solve all issues.  Human agents are still needed to solve increasingly complex problems.  Over the last 10 years, companies have not invested in innovations to make the agents successful.  Now they realize that they need to help agents do well.

What else should we know about ASAPP?

There is a lot of noise in the market about artificial intelligence.  Everybody has AI on their signage in a convention hall. The strongest way to discern what is the real difference is to look at the results and outcomes people are getting. We are very proud of the companies that we serve today. We're very proud of the case studies. We use the phrase ‘transformative results’ a lot. We increased concurrency for one of our customers by 64%. We reduced the actual workload on chat agents, using AutoCompose by 26%. That's real transformative results. We like to show companies what we have delivered before to let them know what can be done.

Just look at our results.