ASAPP Executive Interview
Chris Arnold, VP, CX Strategy, ASAPP
ASAPP Helps Businesses Empower Agents with Efficacious
Desktop AI Technology
In today’s increasingly complex contact center environment,
attrition has become an even greater threat. Disenchanted employees are more
likely to run out of patience quicker when they feel they are not given proper
support. It’s critical for organizations to provide tools that allow
individuals and organizations to realize their full potential. ASAPP takes an
agent-focused approach, offering research-based AI, machine learning, speech
recognition, robotic process automation, and natural language processing tools
to empower agents and make it easier for them to serve customers.
CrmXchange caught up with Chris Arnold, VP CX Strategy for ASAPP
as the company was preparing to exhibit at CCW in Las Vegas to discuss what
differentiates their AI solutions in the marketplace. ASAPP also struck a
partnership at CCW with TaskUs, a leading contact center provider (BPO) focused
almost entirely on digitally native, hyper-growth companies that won the CCW
Excellence Award for BPO of the Year.
To download the interview, please click the eBook image at the bottom of the screen.
Chris, please tell us a bit about yourself.
As someone who started his career as a phone agent, I was
motivated to join ASAPP by my sense of empathy for those who are trying to
succeed at their very complicated and difficult jobs. I spent much of my career
somewhere between the front lines-- supervisor, contact center director, and
operations-- making numerous technology and process decisions on behalf of
these customer-facing employees. I was responsible for purchasing a lot of
technology over the years and also had the unique perspective of being an ASAPP
customer. I have sat on both sides of the table.
As an ASSAP customer, what was it that attracted you to use
I was initially attracted by their focus on the agent, an
area which has not been developed as much as it should have been. Suppliers
have needed to modernize and bring intelligence…in this case, artificial
intelligence… to the complex contact center space. This should be done across
both the voice and digital channel in service of agents whose job has grown
more difficult over the years. Much of the technology I bought was based on
automation: chatbots, IVRs, conversational AI. But with all that, we
infrequently had the opportunity to put high-functioning technology on the
desktops of the agent that they could use in real time. That is a key
difference that makes ASAPP stand out.
Walk me through how the technology works.
In my world, agents took between 40 and 50 interactions each
day. That could have been in a chat or voice environment; we did both. But the
desktop technology—the actual UI—was ASAPP. While some customers might not
choose to use the ASAPP UI, the agents I worked with were using it. Particularly
in the digital environment, they were using the UI to chat and message with
customers. We were also augmenting the agent by having numerous micro processes
One of our products, Auto Summary, automatically summarizes
the conversation and can save 1-2 minutes per interaction. It is particularly
effective for repetitive, mundane tasks that agents seem to go though in about
every interaction. We try to automate as
much of that as possible. Think about troubleshooting a wireless smart phone
outage where there were a number of back-end systems. Prior to ASAPP, our agents had to manually
seek out information from those 20 systems. With ASAPP sitting on top of the
desktop, all the information was brought In front of the agent, whether it was in
a knowledge base or in an outage management system. Having all the information
at their fingertips allowed the agent to focus on the conversation as opposed
to the transaction.
Is ASAPP monitoring the conversation in real-time?
ASAPP transcribes the conversation in real time which can
take place in a voice environment where it utilizes real-time automatic speech
recognition (ASR), which automatically transcribes the verbal conversation
between the customer and the agent. In a digital environment, the agent already
has the transcript as the conversation is being typed back and forth. You can
take those words from the conversation—what we call ‘entities,’ keywords, or key
phrases—and through machine learning, ASAPP has a high degree of confidence for
bringing up a specific knowledge articles or can invoke a specific action,
automate a workflow, or mine an outage management system to pinpoint that
particular customer’s location. There are applications across all verticals.
How does a company prepare for an ASAPP implementation?
Many customers have antiquated legacy technology in their
technology stack. In many cases, there are APIs available. If you think about
deploying artificial intelligence, specifically machine learning (ML) algorithms,
they function on large quantities of data. It really matters what data a
business feeds through to train these models so there can be continued effective
learning. Ultimately, what gets produced for the agent has a high degree of
efficacy: producing information in real
time to resolve issues, drive higher customer satisfaction, net promoter
scores, and reduce AHT…all of those critical KPIs.
But the business has to make sure that the data the agent
requires is running through the machine learning algorithm in real time is
accurate. There’s so much data in every organization that provides context for
swift resolution, but APIs aren’t always available. So, the most important
thing is ensuring that the data needed to amplify the agent’s work, to feed
these machine learning algorithms, is available.
There were times when we had to take an antiquated legacy
data warehouse stream, normally batch processed, and do work on the back end to
make that data accessible to agents in real time by using an API.
A good way for a company to get started is to take a look at
its data architecture.to see if it can be made available. The conversations
back and forth are easy to obtain and are what is needed to feed the algorithm.
Along with the data, what else do companies need to think
Beyond the data, enterprise companies need to look at all of
the technologies in their stack. A tool like ASAPP can facilitate the conversation
between a customer and agent, expediting the conversation to clarify offers to
customer for increased success. Businesses need to look at what they have and
gauge its efficacy and if it’s doing what is needed. A big part of driving
efficiency and improving customer service is simplifying an overly complicated
Does ASAPP guide the agent with what to say, or does it
provide information on the desktop only?
In the light of the pandemic, consumers are rapidly moving
toward using chat or messaging to communicate with companies. ASAPP is nudging
the agent, giving them options on how to reply in real time. These prompts are based
on what best agents have said historically to drive higher satisfaction and
achieve greater efficiency. In a voice environment, there’s a balance to be
struck that doesn’t get the agent frustrated by limiting their options and diminishing
their authenticity with the customer by popping up knowledgebase content that is
relevant and helpful.
Often agents have to click as many as 4 or 5 times to find
an appropriate article and in many cases read through multiple paragraphs while
trying to keep up the conversation. This can be awkward in real time, sometimes
leading the agent to have to put the customer on hold. This is no longer
necessary in this AI Native world where the solution is transcribing the conversation
and bringing forward the answers to the consumer’s questions—and guiding the
agent on what to say but allowing them to do it in their own way. ASAPP
produces a summary of the article in real time with a latency of milliseconds—instead
of minutes—which is imperceptible to the customer on the other end. ASAPP allows
agents to avoid lengthy delays while engaging in a free-flowing, natural
sounding conversation. So, instead of having to focus on getting through the
transaction, agents equipped with ASAPP can focus on servicing the customer
while driving desirable outcomes like first contact resolution.
How does call summarization help the agent?
Call summarization is a feature that is quite important. It
not only listens to the call but can invoke actions, such as when an agent says
they will get back to the customer in three days, it triggers an automated
action so that no promises are broken resolving another common CX issue.
Instead of relying on agents to type two minutes’ worth of notes at the end of
the conversation and hope they remember to follow up, all promised actions have
been captured as the interaction progresses. The company eliminates virtually
all of the call summarization time. The positive effect goes beyond each
individual call; since most agents operate in queues with back-to-back calls,
they don’t always have breathing room to write up cogent notes for each one.
This often has the effect of having them go into every call distracted fueling
In what other ways is ASAPP different from other solutions?
Another attribute which separates ASAPP from other solutions
is that all of the AI models are designed for our customers' data, which is
specific to their operations. We’re also looking at outcomes and we’re taking
every signal to feed those models. The solution currently works in English and
Do you find that agents using the ASAPP solution enjoy their
jobs more and leave less?
40% of contact center workers leave their roles within 12
months. We’re seeing this problem being alleviated among companies using our
technology. High turnover rates are not new. In my 20 years in the contact center space,
it’s always been an issue. But in that timeframe, the environment has become
very different. The modern contact center employee who often comes from the
Twitter generation learns very differently and expectations are higher. The
contact center itself has not always evolved to keep up with the times. So,
from a workforce management perspective, while we talk about work/life balance,
it’s still very regimented. It’s not just the technology side. Businesses have
to be cognizant of the needs of today’s agents to create an environment that
keeps people around longer.
Where does ASAPP help in training agents?
Companies often have training programs that often involve 10-to-12-week
onboarding, feeding prospective agents an overflow of knowledge that they can’t
always assimilate. It’s difficult when agents don’t have technology on their
desktops to help them manage the information. The training program starts them
with a huge brain dump of information which they are supposed to remember in
real time conversations with highly emotive customers. This causes anxiety and
leads to frustration, putting a significant cognitive load on the agent. Most
agents who graduate from these lengthy programs are scared before they even
start. Technology like ASAPP can not only reduce the amount of time needed for
training but also simulate customer interactions in a safe environment. Very few
companies have evolved into using artificial intelligence to update their
training programs to keep up with the changing characteristics of 2021
It seems that this environment also provides more employee
Yes. Having this type of support at hand enables agents to
better own contact center issues. They want to take care of customers without
having to escalate to supervisors or transfer them to a different department.
Not having this kind of empowerment drives friction and fragmentation which can
lead to total frustration, because they do want to be the hero. Instead of
deploying desktop technology for the agent, too many companies have been busy
deploying bots or IVRs that do nothing for empower their representatives.
What was the impetus behind the development of ASAPP?
ASAPP was founded by our CEO, Gustavo Sapoznik, whose
impetus—like so many visionaries in this space—was a terrible customer
experience. In his case it was with a cable company. The communication went on
for three hours unsuccessfully trying to solve a basic problem. While he was on
his lengthy holds, he was googling the world of CX, learning just how big the
problem was and how great an opportunity it presented. His brief research
revealed that it was a $600 billion-dollar worldwide problem. He realized that
artificial intelligence, if skillfully applied to CX, had the potential to
create massive benefits for consumers. He saw the problem as not lying with the
individual agent but with the companies that had not provided their agents with
the right tools to resolve issues. Too many organizations viewed customer
service as a necessary but nettlesome cost of doing business. Out of his own
experience, he was motivated to bring effective AI to fix CX. We now have many
Our focus is on the agent and providing them with
intelligent tools that help with issues like attrition and absenteeism in the
contact center space. Now is the time to
use AI applied to CX in a meaningful way.
Agent assistance has not been a priority in the past. We need to look at what agents are both
saying and doing; marrying those two can help gather insights that AI can
provide to help address the challenges of the agent. Until we throw a lifeline to agents by
assisting them in real time, these issues will get worse.
Click image below to download the ebook