UJET Executive Interview
Vasili Triant, Chief Operating Officer, UJET
Click the image below to view the ebook
Sheri Greenhaus, Managing Partner of CrmXchange sat down
with Vasili Triant, Chief Operating Officer at UJET,
a cloud contact center platform for businesses who put trust at the heart of
their customer experience.
Welcome to the CrmXchange interview. To start, can you
tell us a little about UJET and how the company helps the contact center?
We are a cloud contact center as a service application.
We build software that is the interaction medium between you or any end user
and the companies and brands they are dealing with. Every customer order
creates a customer service need. We build all the routing logic, the reporting,
the blending of channel access via mobile and web. To reach a brand, instead of
popping out to an iPhone dialer, or accessing a chat widget, we keep you within
the app you are using. This allows users to connect in the app that they are
using to the brand, which creates a seamless, effortless experience for the
It sounds as if you are app agnostic. If there are new
apps within the next few years, is it safe to assume you'll be able to connect
to those apps as well?
How did you get into this space?
I got in the space starting as a reseller implementing IP-based
phone systems. Back around 2001, when I started that company, I sold to a
hotel. The hotel needed something for their call center. None of my engineers could figure out how to
add what the hotel needed. I looked at how I interact with the hotel and designed
something around the typical customer needs.
Years ago, when you call the hotel, almost 100% of the time,
you could not reach the front desk. There were lots of menu trees; press 1,
press 2, etc. I built something a bit more efficient and took it to the hotel
the next day to implement. They were excited about the application. It was
personal for me. From there, it has taken a life of its own.
Let’s take a look at chat. UJET recently did a study
that claims 80% of consumers who've interacted with Chatbots said they're
frustrated, 78% say they're forced to connect to a human. Why do you think that
First, our research, which shows that Chatbots are basically
ineffective, is really going against the grain of most industry analysts. Automation
and Chatbots are often referred to as AI. It’s supposed to be artificial
intelligence, which is essentially a computer that can act as a human with
logic, reasoning and learning. What has happened in the industry is an AI label
has been attached to Chatbots which is just a legacy IVR tree for chat. That’s
the reason why we have less than optimal experiences.
On one hand, you have analysts saying automation can make things
more efficient; it can save you money, head count, and help with FCR – all
through a Chatbot. In reality it’s a screening mechanism. In my personal life,
I don't think I found one that is effective. Most of the time I'm dealing with Chatbots
that don't touch any of my own customers, whether it's airlines, Peloton or
Think about legacy IVR trees: 1) the idea was deflection and
2) to get you through to the right group. Consumers are put through multiple
trees to find that sub menu that gets them to the right person.
For Chatbots built in the background, it’s a series of
questions posed to the user, then based on their answer, they are given a set
of other responses. By the time the consumer gets through all the menus and
can’t get their issue settled, they are connected to a live agent – who then
starts the entire process over again. Customers get annoyed because it is
usually something simple, like a product return.
Chatbots are just a series of decision trees, they are not
Chatbots are working from a database. It seems like it
should not be too difficult to transfer the information the consumer entered to
the agent. Why is that not done?
The agent and everybody working on live interactions are
typically working within a contact center platform. Chatbots pop up when needed.
Currently, there are approximately 800 to 1000 Chatbot companies globally. These
Chatbots are disjointed; it’s a separate application from the contact center
platform. Once there is a Chat transcript, it needs to be handed off to the
agent. There needs to be deep integration for this to work seamlessly.
Chatbot companies are trying to get their automation
technology in the hands of users. Because of this, it is relatively ‘lightweight’
integrations which creates an ugly handoff.
If the Chatbot and transition is not going to work efficiently,
what is the reason an enterprise chooses to go with that technology? Is it
strictly cost or there is a thought that this Chatbot is ‘good enough’?
There are amazing marketing messages out there. When you
see a siloed demo of a Chatbot they look good with great functionality. Organizations
are told the integration is no problem. I think that's where people make an
assumption, because nowadays there is such an open ecosystem and you think
about the application integration as no big deal. However, it’s when you put 1 and 2 together
that it starts to get much more complicated.
There are levels of application integration: 1) pass off the
user and send the transcript 2) transcript goes to a record 3) the agent has to
go look in the record to find a transcript. By that point, the consumer is already talking
to the agent. There is also the matter of verifying the user as well.
What is the best way to create the chat structure for a
If you want a good Chatbot it must be simple to implement
and simple to administer. And the reality is, if you want to provide better
service with more seamless interactions, it's actually not simple. Simple is creating
menu trees saying, ‘here’s my menu 1 and it trees off to menu 2 and 3 and
here's your 4 buttons and then you're going to pop out to a live agent. That's
simple. Complex is ‘I want to learn’.
That’s the true definition of AI. It actually learns, without human
intervention. The result is similar to a human-to-human conversation.
It's taking past interactions and saying something there
didn't work. What was the learning? Is it identifying which interaction was
successful? So now for the next interactions, the model will be changed. The
Chabot understands that sentiment, emotion and specific words used in a
specific order means that consumer is or is not frustrated.
Some companies are choosing to just have chat, no
agents. That can cause some real customer service issues. Chat may be less
expensive up front. Is that the primary reason for chat?
Everybody gets stuck on cost per transaction for
customer service. A few analysts started this back in the early 2000s thinking
chat is going to replace voice. Voice is too expensive and everybody is going
to want to go to chat. But the reality is a conversation is fluid and flexible;
all kinds of things change. It’s important to understand the tone of voice.
There may be technologies that are coming that will usurp chat.
Frontier Airlines recently announced that they are going
to only chat.
When I saw that, the first thing I thought is that
Frontier is trying to remove cost. They were having a challenge with their
voice agents and I think they specifically stated their hold times were too
long. It was an internal issue. Instead of fixing the issue, they got rid of
the entire thing making it a customer problem. The danger is, the customer may
now decide to call another airline. The only people that are going to fly
Frontier are the ones who really want that low-cost airline coming from a
specific destination, like their Denver hub. If I am flying from another
airport, why would I choose them? I'm
just going to fly with somebody that provides better service.
Maybe over time things shift. Generations cause that shift. I
think what this industry is missing is that chat isn't the future: textual
communication through a plethora of applications is the future.
The trend is to meet the customer where they are. Are they
in Snapchat or Instagram, or some other application? Our partnership with Google
has opened my eyes. Forget WhatsApp, WeChat, SMS etc. Where do most people go? They start either in
the mobile app of a company, or they go to Google search. A consumer can bypass how they reach a
company. Between a brand and consumer, if someone goes to Google, they can
actually speak words into the Google search and a contact us button could pop
up through something like SMS or whatever app the customer is on. It's amazing
what could actually be improved in a very short period of time. Around 90% of
customer service inquiries start with a Google search.
You were saying things will be different. It's not
necessarily going to be chat, or the chat that we know of today.
We are looking at the number of different communication
means that people are using. The next generation are using many different apps.
They're using short code words, not full written explanations. They may want to
escalate, but they want to escalate between multiple different apps and
channels versus escalate to voice chat. Chat is a dedicated textual experience
versus being in an app and message from that app. Today it is handled by doing
something like an Instagram or WhatsApp integration. At some level, you have to
be able to support all these applications seamlessly. And if you continue to
try to build an integration by app it will take forever. Chat will stay there
but the ‘new’ generation with a dedicated chat bubble will change.
Technology will get more conversational; the tech can listen
and learn. It can listen to your entire phrase. Technology will be more
expensive but it will provide a more personalized experience.
Where does ChatGPT fit into all of this? It is all over.
I read two different viewpoints: it’s great and it’s not ready. Which is it?
ChatGPT is a product from OpenAI. I'll call it advanced
conversational intelligence for chat which can learn and figure out phrases and
conversations to give you the right responses. It will get better over time
because it's going to learn; it's going to have more data points and more
information. Chat is a menu tree of options, ChatGPT actually learns.
ChatGPT will become not just a product from one company; there
are other companies that are doing learning textual conversations, where it
learns and actually has a human conversation which can understand sentiment in real
There is going to be more of an emphasis on results and how
your customers feel. Companies need to
test their own Chatbot – they need to see how they feel at the end of placing
order and think, is that how you want your customers to feel?
I think we'll see some of it during this recession. Companies
realize that they need automation that works while looking at financial
constraints. We will see a strong focus on having great customer experience
through the right tools and allowing your customers to connect to you in any
way they want. We will be out of the
phase where the customer experience does not matter.
You mentioned recession. Some say we're in it, some
people say we're going into one and some say we will have a light recession.
Let’s use the premise that we will be in some form of recession. What do
contact centers need to do to prepare for that?
First, companies need to look their technology stack. The
focus is now on the agent. Easiest thing to do is remove bodies, but there is a
lot of money wasted on their technology stack - in the millions in most
Boom time is over. During boom times, we assimilate lots of
applications for very niche things and then we bolt them all together because
money was available. We start stacking it all together. Those integrations cost
money and people's time.
I recently looked at over our vendor list and was shocked
how many SaaS applications we had just started stacking up. We need Salesforce and
an application on top of it because Salesforce doesn't do forecasting better,
and another application on top of that. Before you know it, you have 20 applications
because a standard out of the box Salesforce doesn't give you what you need.
We turned it off to see how people feel. It’s possible to
change core pieces of your stack, get more from one vendor who is going to give
you a better ROI and bolt in automation to get some of the FCR though automated
Contact centers need to look at modernizing their stack to
save money. Modernize your processes, so you're giving customers a better
For a company that's looking to consolidate and find a
better option, what's the first thing that they need to do? Do they look at
what is not working well?
I would look at what is possible. What exists that can
transform my specific business and my industry and what tools do I need to include?
Normally, I would say an integrator or technology provider can do an
assessment. However, most are still motivated around selling more bits. Because
of that, you may have an assessment that tells you to replace one piece rather
than think through what is needed over the years.
I would tell the business owner to take themselves
completely out of their business position as a business manager and think about
what matters to them. Look at tech and decide if it helps the customer. I'm
probably hurting some people's feelings, but for tech-like speech analytics –
is it worth it? It tells you after the fact that the interaction didn't go well
and helps with analyzing issues and coaching, but it did not help with that particular
For virtual agents, I would think through interactions you are
trying to automate. I’d make sure that the ones I already know don't need to be
automated are sent to a live agent immediately. You would be saving money going
directly to the agent by not paying Chatbot ‘time’.
FCR is always important. It’s better to solve the problem the
first time rather than make the customer follow up. If an application is to be
eliminated, it should not be voice. Eliminate email, one of the worst service
experiences as a channel. There is no SLA, it can take days between a response.
Get down to the core pieces.
You need automation – you need routing between channels, you
need reporting. You need to make it easy for the end users and then try to
figure out what are all the things you've added on top of that.
When you're looking at the overall picture of what needs
to be done to provide the customer with the optimal experience, who within the
company should be involved?
The VP of Customer Experience, Chief Customer Officer, and
the Marketing VP because they're the ones reaching out touching customers and marketing
usually has a pretty good beat on what people are frustrated about. Bring in
your support - whoever is actually on the front line of support. The CFO is
involved for the financial buy in.
Sometimes the problem is that too many people get focused on
cost. What we really need to move to is looking at true ROI models and check-ins.
For example, if something is more expensive but in the long run can save 500,000
dollars more, which is the better spend? Then there needs to be a check in 6
months later. Was the spend correct? We need to look closer at models and see
what moves the needle with customers and saves on cost.
What surprises do you see coming during the next year?
There’s a lot of surprises going on in the vendor world
of this space. Some of them have now started to play out. You are going to see new
players in this space. New players that see the value of customer experience
and what it means to their other applications. There are folks that have a lot
of knowledge in the real world of AI. There will be a shift in the contact
center premise world and the proliferation of other channels of communication.
I think the main thing is going to be some of these new
entrants coming in. Much of the old guard in this space are changing.
Experiences are going to become more conversational and you'll see a weed out
of things like the bad experiences that you and I are having.
Is there anything you think our audience should know that
we did not cover?
When we started the conversation, you said that you have
had bad experiences. I think your audience has to realize that these experiences
aren't great, but people are using them because you stuck a gate in front of it
and you just have to go through the gate.
Many times, I've asked companies what in chat is working for
them. They mention deflection. But are they tracking whether that user came back
in through another channel or did a reorder? The reality is they are not
tracking. The reality is they track the end of the session only. Companies need
to see how many times the customer came back because they did not have their
question answered. They should also assess what the chat live agent provides as
a solution to the customer problem and what the phone agent provides. Does one
know what the other is providing, is there a single source of truth? This is the reality your customers are facing.
Companies need to reassess their process and their technology stack.
And when nothing works, you go on Twitter and complain,
and then they get back to you.
I hate doing that by the way, because people think Vasili
just goes on Twitter because it'll just solve this problem. It’s not because
I'm a CX executive, I have literally tried everything else. Chatbots right now
are very frustrating, but they don’t have to be.
We were talking about Google. Through cookies you know where
people come from and what they looked for. Now, when that user comes in, I know
how many places they've been, what they've been looking for, if they accessed knowledge-based
articles. It’s a better customer experience if the agent can pick up where the
search left off. “I noticed you're looking how to reset your account password. Is
that the reason you're calling?” By
doing that simple thing, you change the entire conversation. When you already
know what the person wants, the customer thinks you actually know them – and is
provided a better experience.