Ep 8: Real-world Generative AI for Complex Support
According to BCG, one-third of executives have already increased AI investments in response to the emergence of GenAI. Even with that investment, user skill gaps and organizational data complexity slow GenAI adoption and value realization. To be efficient, today’s support teams need generative AI solutions built to help them solve complex problems. They also need enterprise-grade solutions that plug into their existing tech stacks, which are available today. This webinar addresses how complex enterprises are approaching GenAI to quickly realize value. Joe Andrews and Karan Sood discuss the top challenges and questions on the minds of organizations considering AI: - What is the maturity state of LLMs today? - How to choose between Public vs. Private LLMs - What are the GenAI use cases enabled by SupportLogic currently? - Why deploy today when the technology is continuously improving?
Welcome. Good morning. Good afternoon. Good evening.
Wherever you are joining us from today, my name is Joe Andrews,
Chief Marketing Officer at Support Logic.
And I'm joined by my esteemed colleague,
Caron Sude, Chief Product Officer. Welcome, Caron.
Thanks, Joe. Thanks for inviting me.
So glad to have you here. We are talking today about a really
important topic, generative AI solutions for complex support.
And we are talking about how companies are getting real value
from that today. And I'm really excited about this conversation.
You and I have spent a lot of time talking about building,
delivering solutions. And so we have some great insights to share.
I think collectively, you and I have spoken with, you know,
hundreds and hundreds of customers, companies in the market,
people who are facing some of the challenges in looking at generative
AI and the broader set of AI solutions today to help them out.
So before we get started, I just want to go over a couple of
administrative things. You'll, if this is your first time joining
a support logic webinar event using the Goldcast platform,
you'll notice the controls on the right. We have a chat dialogue
going. We have Q&A. So anytime during the webinar,
please enter your questions. We will get to them if it makes
sense at the time or at the end. And then we have some docs
available, some resources in the docs section that you can
check out. So we will come back to that and we will do a
couple of polls as well during the event to fit the conversation.
So why don't we kick off? Starting with a little fun word
cloud. There are a lot of terms out there being thrown about.
These are some of the most popular at the moment. And two of
them that really pop into focus are Gen AI and LLM. And we're
going to spend a lot of time focused on those things today.
Want to start with, I think everyone here is familiar with
the Gartner hype cycle. We are at the peak right now for
generative AI on the hype cycle. Just as a background,
Gartner plots new technologies across the cycle and there's
sort of a hype that happens where you see the peak and then you
enter after that this trough of disillusionment when companies
sort of have skepticism and are facing the realities of these
new technologies before you enter this period of normalcy,
normalcy they call the plateau. And it's really been a year,
maybe a year and a quarter since chat GPT 3.5 came out. And
just a year ago, we were having conversations and people, it
was brand new. People were sort of thinking about it. Not really
sure what to make of it. We were all sort of testing it in our
personal lives. But the question of how does this apply to
business and how does this apply to enterprise business? And in
our case, the world of post sales customer experience, those
questions all have been surfaced in the last year. And we're
making some great headway. We're going to dive into that today.
But it is hype. It currently is in a hype cycle.
And so I think we have all seen hype cycles for the different
technology projects in the past years. And there's always,
you know, a day of reckoning, so to say, right? However, I feel
the technology provides such a step function improvement in
making the enterprise work close so much better. And that's, you
know, a contributing factor to the growing popularity and also
the inflated expectations, I would say.
Yeah, well said. And to compliment that hype stage that
we're in, I think it's fair to say that the data shows clearly
the opportunity and impact will be incredibly large. Just a
couple of stats here. You know, this first one from BCG, if you
think about all of the hours per year, worldwide, 14
billion that customers spend contacting service. I tried to
imagine what that's like. It's like taking every person in the
United States and having them spend a full work week entirely
on the phone with with service or contacting them.
Tremendous amount of surface area, so to speak, to to attack with
solutions that help that problem set. You see the massive
trillions contribution to the global economy. There's always a
conversation around or concern raised around, Hey, will jobs
go away? We know with any disruptive technology, jobs do go
away. But many more jobs in this case are believed to will be
created as a result. So it's about shifting the skill sets and
we'll come back to that skill set. And then there's a strong
correlation with CX leaders met, you know, many of you on the
session today are in this space. Believe that there's a
correlation to success in post sales customer experience with
AI solutions.
I agree with everything that you said, do and I think, you
know, I know, you know, AI replacing jobs is is a sensitive
topic. But a statement, you know, from a Harvard professor that
one stuck with me is AI won't replace humans, but humans with
AI will replace humans without AI. That's such a profound thought
actually. And I think that's going to be a theme of, you know,
what we're going to be talking about today, the impact that
this technology can actually make in the enterprise context in the
enterprise workflows is going to be, I think, completely
transformational.
Totally agree.
One more thing to share is that we've seen in just the last six
months, that the perception of JNI has dramatically shifted
among the C suite. And we're seeing now what we're calling this
tailwind of demand, where companies are, you know, looking
to deploy AI and generative AI specifically in the enterprise
and looking for where are the best opportunities to do that. And
just fall of last year, there was a big concern from the C suite
about adoption because they didn't fully understand it, they
didn't understand the implications. And they were not
supporting its rollout. And today they are the key supporters,
right? Two and three are saying it's going to be the most
disruptive technology in the next five years. And fully a third
are saying they're already increasing their investments
due to the emergence. So that's a big shift. Now, it's still
perception, right? Reality has a long tailwind to go. I know,
you know, I come from the cloud infrastructure space previously,
we talked about cloud, you know, 20 years ago or 15 years ago,
and it's taken a long time to get to cloud. So this will be a
long journey. I think, you know, someone famously said that
we in business, we tend to overestimate the short term
benefits or changes of a new technology and we underestimate
the long term. So I think that's definitely clear here as well.
So, Karin, you know, first sort of basic setup question is in,
I'm sure a lot of people here are wondering this or
maybe have different points of view, but how should we think
about generative AI compared to large language models?
LMS. Love it. Love that we are starting with a very basic
still. And I'll keep it simple to begin with because we're going
to go into the details, you know, as we progress in the
webinar. So generative AI is, for me, a subset of the broader
AI technology that excels in content generation abilities.
And that content could be text, it could be images, videos,
code, music. LLMs are a further subset of generative AI that
essentially use deep learning algorithms and are tailored to
perform language related tasks such as X generation language
translation, natural language understanding and text
analysis. So I would say, yeah, generative AI being a subset
of the broader AI and LMS being a further subset of gen AI,
tailored for performing language related tasks.
Great. I think that's a perfect setup. And we're going to be
talking about both interchangeably, but it's helpful to
understand how they fit together as a starting point. So let's
go deeper and deeper. Another question that people will have
about is about the maturity of LLMs. And this is a little bit
of an eye chart, but I thought this was useful context to show
just how quickly that curve has risen to the upper right in the
last few years of computational evolution around the LLMs.
And so curious to get your thoughts here, Karan.
Yeah, I mean, there's no doubt Joe that the technology has
evolved very rapidly. I would say in the last six to 12 months,
every new week, a new foundational model is being
announced. You know, the joke was every time OpenAI has a demo
day, they kill a few hundred startups or reset a few thousand
roadmaps at least, right? Also another measure of maturity,
at least from my perspective is GPT 3.5. That was announced, I
think sometime in November of 2023. It had SAT score of 1260.
GPT 4.0, which was announced just six months later, had a
score of 1410. That's a 150 point increment in just over six
months. And of course, GPT 4.0 had so many more capabilities
for example, it became multi model. So I would say on the
technology side, it actually has been evolving very, very fast.
The second element of this is actually also, as you pointed
out earlier on the adoption side, now almost all companies that
we speak to are, I would say, exploring usage of large language
models in some ways. 2023 in some ways was a year of pilot
projects for a lot of companies. And 2024 is the year when a lot
of the same companies are now rolling out their first
projects in a production cycle kind of a setting, right? So I
would say definitely the technology is maturing, as well as
the adoption of the technologies is maturing as well. And on the
adoption side, I would say in my opinion, one of the reasons
why that has happened is because of the growing awareness and
the understanding of the different levels of sophistication
when it comes to implementing a large language model. You can
start with the crawl phase of just creating a wrapper around
the foundational LLM to let's say the next step being, you know,
influencing the output of the model using prompt engineering or
techniques like RAG to the advanced stage, for example,
being completely building, fine tuning and running your own
model, right? So I think a lot of these different elements from
an education and awareness perspective have also eased the
adoption because people have understood that we can crawl,
walk and run and don't go to the final stage at the very
beginning.
That's a profound point I think you just made, which is there
are really two vectors to this. It's not just the
computational power in the advancements of the technology,
technological capabilities themselves, but it's also addressing
the ease of use, right? And you can't have one without the
other to drive success.
Exactly.
Another question that often comes up is public versus private
LLMs, right? We're all familiar personally with, you know,
any of the public LLMs where we're, you know, performing content
generation tasks and asks and it's pulling from the large public
corpus. The enterprise side has, you know, some concerns, which
we'll get to a little later. But let's talk about how we should
think about them differently. And also how to think about that
crawl, walk, run that you just mentioned.
Exactly. So, so as I was saying, I think there are, you know,
different levels of sophistication when it comes to how you
could implement large language models. Now, there are, you know,
maybe three to five different ways, but let's say to keep it
simple, the beginner mode, for me, is basically just a wrapper
around a foundational LLM model. Now, that may, in certain
enterprise use cases be enough. But at least from what we found
out, talking to a lot of our customers, just a basic
foundational model that is just trained on any public data,
is never going to be sufficient. So the next level up from
there, let's call it, you know, intermediate mode is when you
provide certain instructions and context through something that
we call as prompt engineering. So you have a better control over
the model output. And this could be in terms of the formatting,
it could be in terms of the style. And when we'll talk about a
couple of use cases, for example, case summarization for the
model, and our, in our presentation, we will go deeper on
this. But let's say you have the foundational model, the next
step up from there is, you know, you use prompt engineering to
influence the output of the model. The next level up from
there could be there are techniques that you can use. For
example, one of the most common ones is the trivial augmented
generation rag, where you could add your knowledge and data
context to the model. So you can use your own proper iterative
data and train the model so that the output is restricted to the
context that you've provided to the model. Right? That's the next
level of sophistication. And then I would say the level up from
that, let's call it the advanced mode for me is when you
essentially build fine tune, run, and probably even orchestrate
a series of your own custom private LLM models. So these are
for me at a very high level, the different levels of
sophistication. And for the different use cases, one or the
other may have to be chosen. And of course, there are advantages
and disadvantages with each of the approaches.
Thank you, Kern. So can we walk through the trade offs a little
bit on each side? I know they're bullets here kind of highlighting
it. But how should I think about it? Maybe if you start with
phase one? Yeah, yeah. Yeah, I would say, you know, I'm in
there are there are a few different vectors. There is, for
example, customization and flexibility, there was
scalability and performance. Also in certain regulated
industries, you know, privacy and security is more important than
certain other industries. And then of course, on the other end
of the spectrum, there is, you know, there is always cost to
development, there is the cost to integration. So from my
perspective, I would say,
things that are the factors that kind of favor a private LLM,
for example, are when you need extreme customization and
flexibility. And this is because public LLMs kind of come with
predefined architecture and parameters, limiting the
customization options. And in contrast, the private LLMs offer
the organizations the flexibility to tailor the models
according to their specific needs in cooperating even the main
specific knowledge and fine tuning the parameters, specifically
for a particularly use case. So I would say for the use cases
and organizations where customization and flexibility is
the most important thing, you know, maybe phase three or phase
four is the way to go. Same with scalability and performance, I
would say private LLMs, you know, can be optimized for scalability
that can be tailored to the specific workload requirements of
the use case. So again, if scalability and performance for a
particular use case or for a particular organization is the
key, maybe phase three and four are kind of more important. On
the other end, you know, with how fast the technology is evolving,
you know, every every new week, a new foundational model being
announced, developing and maintaining private LLMs can be
actually very resource intensive, requiring significant
investment in research in infrastructure and talent. So I
would say it's not a one size for tall, you do not always need
phase three and phase four for all use cases, maybe in certain
cases, just a foundational model, a wrapper around it, the prompt
engineering, which is phase one and phase two, probably is going
to be good enough. So I think in the end, it all depends on,
you know, for the different considerations that you may have
depending on the use case, and the use case could also be
specific to each customer, you may have to choose whether you
use, you know, a public SLM or public LLM with, I would say a
little bit of prompt engineering or most likely more advanced,
which is then going towards the direction of fine tuning and
private LLMs.
Thank you for unpacking that, Karan. So I would just want to
say for everyone who's joining us, there's complexity and we love
having these conversations with, you know, folks like yourselves
who are joining about considering different paths. And as you've
heard, it's very nuanced, depending on your own situation.
So we would love to have follow up conversations with you about
where you are and what are you trying to achieve and figure out
the best path to do that. It's a great segue into support
logic's approach and how we differentiate with our own
generative AI solutions. So, you know, there's kind of four key
areas around this infrastructure and security on sort of the
outer perimeter and then domain deep support and post sales
CX domain expertise and then our ability to fine tune the
models. But, you know, let's dive into that a little bit, Karan.
Absolutely. I've been first and foremost, since there is, as we
discussed, there is no, like, one large language model that
will solve for all enterprise use cases. So what we have done
as both logic is we've invested passively in building, I would
say a robust LLM infrastructure that combines public LLMs,
private small language models and private large language
models. And we can seamlessly switch between several public
LLMs and soon to be announced private LLMs as well for the
different use cases. So I would say first, you know, the way
we differentiate is because of the infrastructure that we've
created around the technology piece itself. Secondly, I would
say we have tons of domain expertise within support logic,
both on the technology side, as well as on the support domain
side. And this is extremely important, in my opinion. As we
said, as new foundational models and domain specific models
are announced every week, we have a dedicated team of ML
engineers that are continuously evaluating them. And if someone
wants to build a private LLM from scratch, you will need a
dedicated team of ML engineers to research them, to test them,
to build them to train them to deploy them. And that requires a
lot of deep expertise and ML resources. And on on the domain
expertise side as well, you know, we take the burden off you,
for example, we were talking about prompt engineering, it's
extremely important to pass the right context and instructions
to a foundational LLM model to get the desired output. The best
example that I have for this one is, you know, one of the use
cases that we do is case summarization. Now, there is
never going to be one case summarization, which is going to
be, you know, best for every user persona. Think about if it
is a case summarization for a handoff from one agent to the
other, the focus has to be on the most recent events on the case,
what are the immediate next steps? Whereas if the user persona
is a support manager, then the case somebody has to focus more
on what has happened on the case from the time it actually was
logged, who are the people that were involved. And for a
support manager, recommendations on what the swarming team should
look like. Whereas if the case summarization is for an
executive, maybe there, you know, the focus is a little bit
more on what's the overall sentiment on the case and at an
account level. Now, this requires different kind of prompt
engineering that could be needed for the different user
personas. And that's something that support logic has been
doing for many, many years. All of us are, you know, domain
experts in the support domain. So I would say the second level
of differentiation is because of the domain expertise that we
bring to the table. Third, I would say, as you go up the level
of sophistication in terms of the implementation, we are also
able to use techniques like retrieval augmented generation
rag to train models on customers proprietary data. And we can
also actually fine tune the deeper layers of the models by
further training them on smaller and specific data sets that we
can get from customers. So I would say leveling it up, even
the fine tuning is something that we can help do for our
customers. And of course, last but not least, we take security
very, very seriously for all public LLMs, we use enterprise
level safeguards to protect the data from being stored in
subprocessors, something that we've heard from our customers
and prospects. And on top of that, we've also built an
additional layer of reduction, which kind of removes all the
personal information from the data that is being stored. So I
would say across these four different vectors, there are
certain things that we do that kind of take the burden of our
customers, and they can rely on our expertise. And the fact
that we have dedicated engineers who do this on a full time, full
day basis.
I would just add to that, if you look at our customers, and we'll
show some later on, many of them are in the complex
infrastructure and or security space, or in the data and
application space where all of these requirements are
critical. And it's it's because of the domain experience and our
ability to do something for them that is not within their core
competency that that they've decided to partner with us. So we're
very proud of that. And, you know, any concerns that you may
have around any of these vectors, we're happy to dive into at a
deeper level. So let's pivot a little bit and talk about the
use cases. We have, you know, we, as I said, we have a number of
customers that are doing a whole bunch of amazing things. We
also talk with, you know, hundreds and hundreds of companies
out there in the market. And this is what we hear. They are
prioritizing. And we would also love to hear, you know, get your
opinion, we have a poll that's opened up. So we have kind of
three basic areas, the agent productivity side of it, the
core support operations efficiency, and then quality
monitoring and coaching and we'll dive into a little bit of
detail of those. But if you go to the poll, tell us what is top
of mind for your priorities. We'll give that a few minutes. While
people are doing that, we can, we can dive in a little bit. So
Karen, if you would, let's talk a little bit at first about
agent productivity. I think this is the one that, you know,
companies we speak with first talk about, because that idea of
using generative AI to create a response to summarize the case
to, you know, deliver guidance on next best actions, etc. Let's
talk a little bit about that and how the gen AI technologies are
able to help with that. Absolutely. And if I may take even a
step back, so I would say, you know, we at support logic have
been doing AI from the time we started, especially predictive
AI. We actually even started using technologies like like
BERT, I would say way back when it was launched in 2018, 2019. And
this was, you know, one of the first LLM's way before, you
know, the hype that we've seen in the last six or 12 months, and
at least at support logic, we fundamentally believe that
combining predictive AI with generative AI is the real game
changer. So what you see on the slide from Joe is, of course,
we will, we will, we will kind of dig deeper into the gen AI use
cases. But all the other things that we do with predictive AI
combine with gen AI for us is the real game changer.
Thank you for starting there. By the way, just to guide us. So we
just closed the poll. About 50% have said their top area of
priorities in this core support operations efficiency, 40%
agent productivity, 10% quality monitoring coaching. So we'll
love that. cater our content accordingly. Yeah, yeah. And I
would say, you know, when it comes to the gen AI use cases, I
personally like to use some kind of a taxonomic categories to kind
of group even the different use cases, right? A few examples of
these categories could be, you know, Jenny, I can help you
summarize content. Jenny, I can help you transform content.
Jenny, I can help you create new content, or it can help you
retrieve content. So I think these are the broader categories.
And within each of these categories, we have different use
cases across the breadth of our product portfolio as as Joey
were talking about. Now, for example, under the content
summarization, we have use cases like case summarization and
account summarization. And again, as I was mentioning, this
could even be broken down into case summarization based on a
specific user persona case summarization for an agent kind
of going through the handover from one agent to the other,
being one of the use cases versus case summarization for a
support manager versus case summarization for an executive.
On the content generation side, we have use cases like something
that we we plan to build in the second half of this year is
knowledge article generation using Jenny I for content
transformation. We have use cases like translation assist
translating from one language to the other. Most of our customers
are global customers that deal with cases coming to them in
different languages. And even before we can do the sentiment
detection on the case, the case may have to be translated for
the agent for the support organization, let's say into
English. So I would say on the content transformation side
translation, it's just as one of the use cases that we support
today. The other one is response assist, where let's say after
the case has been summarized, also helping an agent or maybe
even a support manager compose a response that could be sent out
either internally or could be sent out directly to the
customer as well. And then on the content retrieval side, we have
use cases like troubleshoot assist, and something that we're
currently working on is natural language powered analytics.
Where for example, you can query our platform based on natural
language instead of prebuilt charts coming from us to you. You
could on the fly in real time create your own chats using
natural language. So these are, I would say some of the examples.
I'm just going to expand it to some of some of the new
capabilities coming as you're referring them. Yeah, absolutely.
Anything to talk about on the core support operations efficiency
side, you know, just we tend to think of that as sort of more of
the summarization retrieval, helping to guide workflow and
providing, you know, the analytics that help the entire
operation get more efficient. Exactly. And I think one of the
other use cases that we, we are currently working on in our
in our current releases, besides the the case summarization,
also summarizing based on all the post sales interactions that we
get into our platform, summarizing at an account or a
customer level. And not just summarizing, actually, it could
also be providing insights about, you know, like for a case, there
is a likelihood to escalate. There could also be a likelihood
to turn on the account side. So some of the things that we are
also planning to build in the second half of the year are kind
of moving or kind of adding to our case centric view to kind of
also becoming more customer in account centric and providing
count, churn insights or account summarization and stuff like
that as well.
Fantastic. And for the people who joined us who are interested
in quality monitoring and coaching, just a brief mention of
that, I think there are some powerful use cases here being
able to automatically cue all customer interactions right
against a rubric and against compliance standards based, you
know, in some industries, that's very important. Being able to
sort of surface back coaching insights and analytics that help
not only your frontline agents, but the entire org get better.
We've seen some powerful results here. So if that's
something you're interested in, or if you have questions about
that, you know, feel free to add that into the Q&A.
And I would say just to add to that one, I would say a lot of
our customers are looking to move to move from, you know, a
very manual QA process to an automated process. And I would
say through JNI, you could completely reimagine, you know,
for example, the automatic QA process as well, where the score
cards and the automatic grading could happen based on
or leveraging the JNI technology. So I would say immense
opportunities to kind of completely transform the auto QA
process using JNI as a technology.
Fantastic. All right, let's move ahead. One of the other
topics that folks are interested in covering are the
challenges, right? What is getting in the way of enterprise
AI and LLM adoption? And let's go a little deep on these and
talk about what they are and how to address them.
Yeah. So Joe, I think even before the five categories that I
mentioned here, I unlocked the blog that you had there, you
know, you were talking about, you know, it's for any
technology, not even JNI, it's important to define the
problem first. JNI could be a solution to some problems. But
it should not be the end goal. In fact, misusing JNI
diminishes the value of AI in an organization in my opinion. So
so getting the problem first is the most important thing. But
specifically speaking about just sorry, Karl, just quickly
mentioned the poll is open. We want to hear your opinion in the
audience. What are the top challenges you're facing today?
Pick up to two.
All right, so specifically speaking about the JNI challenges
and the pitfalls from a technology perspective. Personally, I
would say, according to me, hallucination is one of the
biggest ones that I've heard from our customers. One of the
challenges with large language models is that they have a
nancy to confidently bullshit or hallucinate stuff, right? I
mean, hallucinations are essentially output that while
appearing possible, are not based on factual animation. And
this becomes particularly significant in an enterprise
context where accuracy and fact based decision making is
paramount. And the principal challenge arises from the fact
that these models, especially foundational public large
language models, they are trained on massive data sets that
has an advantage because it kind of gives them the breadth. But
they're often derived from the internet or extensor, you know,
text, corpora, and if the sole state I encapsulate certain
bias, the model will learn and possibly even magnify that
bias, right? So I would say hallucination for me is one of
the biggest challenges when adopting a technology like this.
Of course, there are ways around it. For example, as we were
talking about using technologies like rag fine tuning, the large
language model are exactly to counter problems like
hallucination. So I would say hallucination for me is one.
Data complexity is the other one key to successful fine
tuning lies in the preparation and the pre processing of the
domain specific data sets, which is critical to ensure, you
know, compatibility with the model and with the task at hand,
whatever is the use case. And this is where as we're saying at
support logic, you know, we have the expertise to be able to
deal with domain specific data to be able to train language
models based on your proper right tree data so that we can,
you know, make the complexity of data a little bit easier for
our customers. Third, I would say is till in several
organization, not just in the regulated industries, but even
generally across many customers, data privacy and security
concerns are still all valid foundational element models are
trained on massive amounts of public data. And, you know, for
a large language model to become helpful in the enterprise
context, they need to be retrained on proper right free data
that may contain some sensitive personal data, financial
information, or or any confidential stuff, right? So to
mitigate these concerns, enterprises must ensure that their
data is adequately secured. There is a reduction if needed is
happening on all of that data. So I would say data privacy and
security is definitely one of the other things. Skill gap, we
talked about it in one of the earlier slides as well on the
technology side, new models are being announced every week.
It's extremely difficult to keep pace with it unless you have a
dedicated team of machine learning engineers and deciding
which models are best suited for which use cases. Yeah. And on
top of that, you know, you, you know, also have the main
specific models. And that expertise basically is also not
very easy. And that's still gap, I would say is one of the
other challenges that I've seen where just using a foundational
large language model, but not having then the expertise to find
you know, not having the domain expertise does not give you the
output that you were expecting for a particular use case. Now,
this is not the problem with the technology. This is that you
don't have the right skill sets that you need to apply on top of
the technology to get the best out of it. Right. So I would say
those are some of the challenges that I can think of. It would be
also interested to see, you know, on the voting, which ones
scored. Yeah. So those just came in. Thanks, Karen. So the top
vote getters were data privacy and security concerns, probably
about 25% that is top two. And then integration with existing
systems and workflows. I think this is
Yeah, I can I can talk about this one as well a little bit. I
would say, you know, like any other technology, Jenny, I is a
solution and not the end goal. So before diving into the AI
implementation, it's it's critical to first define the
business problem that you're trying to solve for. And when we
think of Janae use cases at support logic, we take a very
user persona and a jobs to be done based approach. And we think
about the end to end workflow that way, you know, whether it's
predictive AI, whether it's generative AI is not really an
afterthought. But it is embedded into the core workflow itself.
For example, we dog food, our Janae with our proper right
free predictive AI to get more out of generative AI. So I would
say integration of Janae into the end to end workflow is extremely
important as well.
Gotcha. Just want to state there's some questions coming in.
Avi, you've got a couple in there, we're going to come back to
those really soon. We have about five more minutes of content.
And then we want to get to Q&A. So for others, if you have a
question, please enter it in the Q&A. And let's move along. This
is an important topic, though, is addressing these challenges
and just a couple of stats from a survey from IBM in terms of,
you know, the propensity for the skills gap and the data
complexity to be issues. It's pretty significant among
concerns among IT leaders who are faced with deploying these
solutions. Another question that comes up is, why deploy now?
Right. We looked at this exponential curve of technology
and computational improvements. You know, why not wait? Why should
we deploy now?
Yeah, that's that's one of my favorite ones. You know, I have
you, Jenny, I as a fundamental shift in how we will deliver
interfaces. This technology provides such a major step
improvement in interface, making the enterprise workflows so much
better compared to what it was without the technology, right?
So it's on the same scale as if not more, I would say, worldwide
web three, three decades back or mobile phones to decades back,
just like it was never a good advice to not go online and
leverage the internet. In fact, the companies that aggressively
embrace that change were the ones that actually succeeded. The
same way I would say it does not help to sit on the sidelines
with this technology and wait for it to see how it plays out,
right? Earlier doctors will likely will be in a leadership
position as well. And the earlier you get involved, the more
you will learn and the more it will put you in a better
position, as the technology evolves and develops.
I would add one thing, Karan, so with previous technologies,
there was, you know, a major forklift sort of transformation
upgrade. There's a lot of the, you know, infrastructure and
foundation that needs to be built and enterprises could not
undertake these changes lightly, whether it was moving from
on-prem deployments of applications to the cloud, etc.
In this case, think of it as a sidecar. There are, you know,
abilities to take advantage and deploy some of these use cases
very quickly, like in, you know, a month and a half, 45 days to
deploy some of these use cases. And it's not negatively impacting
your core operations. And, you know, partners like support
logic would be doing the heavy lifting. And then you're able
to get immediate results that help you inform the rest of your
roadmap and in your plans. So it's sort of a parallel effort
versus I need to make one big decision. And then it's going to
be a major move.
Wonderful said, 100% agree.
Okay. So I want to come back to sort of the benefits. We've
talked a lot about the use cases, the challenges, because we
want to be very transparent and face those upfront. But it's
important to consider that whether you're looking at predictive
AI solutions, which have been around from support logic for a
lot longer or newer generative AI solutions, these benefits are
very compelling across the organization, right, whether it's
the core support team operations management, you know,
frontline agents, executives on the left, customer success, even
the GTM functions, you know, top center or product and
engineering, IT BizOps, CXOs, finance. There are benefits that
accrue company wide. And one of the questions we always get is,
you know, which one should I prioritize? If I'm trying to make
the business case, if I'm a support executive or customer
success executive, how do I compete with corporate IT looking
at say 14 different AI projects? And so this is one of the best
ones that delivers value for post CX. And we see that I'm just
going to show a few stats here across these vectors from
customer experience, right, in terms of reducing escalations,
reproving CSAT, MPS, customer retention, removing friction,
improving experience, and then operational efficiency, which is
all about, you know, better, faster, cheaper. And you see the
results here, one of the customer customers that we're
proud of and did a case study webinar a couple weeks ago with
TSIA is Informatica. They were able to prioritize implementing
support logic. And this actually augmented a build initiative
that they had going on for many years. And they brought us in
because of the deep domain expertise. And they were able to
very quickly reduce the overall number of cases, which took,
you know, removed the drag on their internal resources that they
were able to redeploy to other initiatives in their portfolio
of projects across post sales CX. So all of these customer case
studies are on our website. We would be happy to walk through
them with you, but they're very compelling. So any any final
points, Karen, I want to show a couple of resources, and then
we'll address some questions in the time that we have left.
No, nothing from my side, Joe, we can go to the resources and
then the questions.
Fantastic. So one of the one of the things that we're proud of
at support logic is we have been building over the last couple
of years, a community we call SX live for support experience.
And we do a series of events and we have a lot of virtual
content as like the webinar that we're on now, where you can go
and learn and you can contribute. So we encourage you to check
that out as a resource for yourself. We are also doing a
series of city tours across the US this year. We just completed
two in Redmond together with Microsoft and in Austin, Texas,
and you see the upcoming calendar for that. So great way for
you to network. These are not support logic infomercials.
These are conversations and panels with, you know, thought
leaders and practitioners who are sharing their best practices
and it's a great way for you to learn. And then the other
thing is if you want to see more in terms of what we've done
and see the actual product, we have several ways for you to do
that. If you go to our website, you can request a one on one
demo. You can join a weekly group demo. We have every Friday.
And then there's also an upcoming webinar where we're going to go
deeper on what constitutes enterprise grade AI for the
category because as we saw, there is concern around privacy
and security, rightfully so. And we're going to dive into that
in addition to the other requirements that constitute
enterprise grade. So with that, I want to open it up for some
questions. Thanks for in please enter if you haven't had a
chance. So we're going to get to Avi who had a couple questions
first. When do you think RAG will be available in our JNai models?
So Avi, we are working on this and we are hoping for this to be
available in the next three months. We're already doing some, I
would say internal pilots at this stage, the use case that we
are starting with when it comes to RAG is the troubleshoot
assist use case where we're kind of trying to use this
technology to help agents find one first troubleshoot the
problem and find right solutions to the problem that have been
described in the case. So that's the use case that we are
starting with. And we're hoping for this to be available at least
for our better customers in the next three months.
And there's a follow up as well from Holly on the RAG topic of
do you have any tips or resources about how to get started?
We are in the process of, since this is still not a generally
available feature in our in our platform, we don't have extensive
documentation as yet, but this is again, something that we are
working on. And we would absolutely love to share that with
you as soon as that's going to be available. But yes, this is
this is exactly in the works.
Great. And then another question from Avi, how are you
measuring toxicity and bias in the gen AI sentiment analysis
use cases?
So just just to maybe understand, is it more in the more
classical?
Sentiment analysis itself, because that probably is not even
related to gen EI.
Obviously, if you could clarify in the comments or the Q&A, the
you know, because sentiment analysis is typically the
predictive AI technologies where we're mining all the signals,
the unstructured data, and then organizing it and bringing the
analytics front and center. So there is, you know, bias and
toxicity, I would say is another word for a signal, which is,
you know, a customer is extremely frustrated or upset. And then
bias is about sort of normalizing the data and bringing forth,
you know, predictions and assessments, level setting, a
customer sentiment based on their normal baseline, not the
entire market.
Have you got me from Ron?
Ron?
No, that's that's absolutely accurate. And maybe Avi, what we
can also do is if you could please drop in your contact details,
we can also schedule a separate session with you where we can
talk about, you know, how for example, to be measured accuracy
and precision when it comes to detecting signals and the
scores that we calculate. It's a deeper conversation and we
would love to have that with you.
Great. And a follow up. Well, Amer says, when will this be
should be available? The last I heard was Q1. I think maybe the
reference is to rag what you were talking about earlier.
So the right.
Correct. So the rag, I would say, as we said, it's in the works,
internal pilots in the next three months, but some of the other
use cases that do not use rag, for example, case summarization,
that we have already opened up for our better customers. And we
have a couple of customers already trying that out and working
with us to kind of fine tune it for their use cases. So not
everything is not available. It's only the rag specific use
cases that is that is planned in the next three months. A lot of
the other things, translation assist, responses, case
summarization, these features are already available in the
platform.
Fantastic. All right. Last call.
Any other questions?
I think Laurie. Yeah. How do you approach or pitch a prospect
that is not on board with AI? What why don't you take that
to do and I'll pitch in?
Yeah. So I think Laurie, we basically followed the framework
of showing benefits and then understanding someone's
objectives, right, the priorities, whether that's, you know, in
the operational side, getting more efficient with existing
resources, improving the experience directly for end
customers or for their internal agents. And as we showed kind of
the three column set of use cases, all of those basically
support those higher level business benefits in some form
or another. In understanding where the prospect or customers
main objectives are, we can then quickly map them to the
solutions and then talk about here's where the AI technology
can benefit them specifically. So that's a that's definitely
a starting point. Obviously, a lot more nuance as you're
addressing, you know, potentially concerns, the challenges
that we talked about. But at the end of the day, any technology
is about solving problems and use cases, you know, better,
faster, cheaper. And then it's inspiring confidence with those
who are going to make that change.
Pretty well said, too. All right. So final one, and then we
have to wrap up from our OK clarification about the question
it's about PII data, will support logic be able to detect PII
and redact the PII? Yes. So with our Q1 release, we have
already. So if you're an existing customer, you may have
already seen some release notes, we are introducing
reduction already in Q1. This is also going to be an ongoing
investment through the rest of the year. So we will add more
capabilities. But basic reduction is already available with our
Q1 release.
All right, we've reached the end. Fantastic session. Appreciate
everyone's input on this and Quran, your fantastic business
partner. And I really enjoyed the conversation. You so much
knowledge to bring in that really appreciate your time here
today. Likewise, Joe, it was my pleasure to be here with you.
Thanks, everyone. Please stay in touch. If you have follow up
questions, feel free to reach out and we look forward to
continuing the conversation.
All right, everyone.