Artificial Intelligence

AI, Machine Learning, and Big Data Service With Rapidscale + Logicworks

November 16, 2023

Our November 14th Telarus Tuesday Call welcomed Rapidscale + Logicworks, where they discussed how to propel your business forward with the power of machine learning. Click to access the full recording in Telarus University to learn more about AI, machine learning, and big data service: https://lms.telarusuniversity.com/forums/topic/31582


Doug Miller: Good morning, everyone. This is the Telarus. Tuesday. Call for November fourteenth, 2023. I’m Doug Miller, Director of Order Experience and the voice of Telarus, still feeling proud and patriotic about all of the wonderful celebrations this past weekend in honor of Veterans Day, the Us. Holiday observed each year on 1111, November eleventh.

Doug Miller: honoring all those who served in the United States armed forces, and the date of that came about due to many countries celebrating the end of World War one, as major hostilities were formally ended at the eleventh hour of the eleventh day of the eleventh month of 1918, just over a hundred years ago. It was originally known as Armistice Day, and still is in many of the Allied nations.
Logicworks
Doug Miller: and the observance coincides with Remembrance Day or informally as Poppy Day in Canada and in the Commonwealth Member States. So many of our suppliers, partners, and Telarus employees have served, and we remember and thank you all for your service and sacrifices.

Doug Miller: Here. On last week’s Tuesday call we were joined by VP of advance solutions, for IoT. Graeme Scott for a discussion of IoT. And AI Opportunities and solutions

Doug Miller: Epic iO, Claro Enterprise Solutions, and Abundant IoT. Were all here focusing on those opportunities and solutions. I highly recommend reviewing this discussion as you make plans for your end of year, client conversations and offerings. If you missed last week’s call you can access the recording at any time at Telarus university.com

Doug Miller: to day. We are very excited to be joined by the experts from Logicworks. Their portfolio of public cloud services and solutions is now part of Rapidscale here to day, with a game changing presentation, you need to experience on artificial intelligence, machine learning and big data services powered by Logicworks.

Doug Miller: Propel your client’s business forward with the power of machine learning all of that, and more coming upon to day’s Telarus, Tuesday call.

Doug Miller: We have issued those highly coveted and suitable for framing. Tuesday call Hall passes to Telarus, partners attending the CVX. Channel vision expo industry event this week in Scottsdale, and to those attending the Telarus 3 day ascend Cloud and Cyber Security education event going on in Boston. In fact, one of our presenters joins us, live from there to day.

Doug Miller: Partners there will emerge as wicked smart experts on those opportunity, rich solution categories and transform their sales cycle. So have a great time there in Boston. And, by the way, Chicago, we will see you tomorrow for a I lightning training.

Doug Miller: Well, with just 2 months remaining until year’s end. We remind you that Telarus partners who have sold at least 125 k. In New MRR. This year will join us, for President’s club coming up this spring in Cancun

Doug Miller: and higher sold revenue can earn additional trips. Now you should also know that some of our generous Telarus suppliers have provided phenomenal offers that can also get you qualified with a much lower threshold for selling their specific solutions. So thank you. To each of these Airespring, CBTS. Dialpad Granite.

Doug Miller: Nextiva, NHC.

Doug Miller: Peerless infobip, spectrum and zoom. They have provided these terrific offers. Now, fourth quarter is moving quickly, but there’s still plenty of time to qualify for our Telarus Presidents Club in Cancun. Now, our top 10 selling Telarus partners overall world wide will join the chairman’s club, and they will join us in May for an 8 day tour of Ireland.

Doug Miller: It’s magically delicious. Yeah, I said, that you can get there contact Telarus to day, and we will help you map out your fourth quarter strategy. After all, it’s what we do we help our partners win?

Doug Miller: Well, I hope you’ve taken a look at it. I have. I love this thing. This is the just released Telarus 2023 Technology Trends Report.

Doug Miller: This is phenomenal. It was highly anticipated, it absolutely paid off. It is packed with exclusive insights to help technology advisers better understand the key issues and trends impacting. IT leaders buying decisions over the next 5 years. Now, Telarus has done an amazing amount of research. We are showing you the emerging technologies in which the industry is seeing the most interest.

Doug Miller: you will be able to uncover opportunities, cutting edge solutions, all of which that boost your client’s capabilities. It’s all there in the free Telarus technology trends. Report. You can register and get your copy to day.

Doug Miller: Now, following up on the unprecedented success of Telarus solution view. That’s our intelligent sales assessment platform QSA. Modules. We had cyber security last year we had cloud, and now we’re very excited to announce the launch of our new contact Center, QSA. Module for solution view.

Doug Miller: This powerful new tool has been available for just couple of months now, but partners are already excited about the benefits they’re seeing solution, view contact centre. QSA. Asks your clients questions based on their answers to prior questions, it carefully guides the conversation to ensure a focused, brief discovery of the appropriate contact centre solution.

Doug Miller: Now that market is huge. Right now, contact centre is big business. You can capitalize on it. This new tool is available now.

Doug Miller: and there’s a recorded launch call which you can access at Telarusunversity.com so check that out and prepare now to cash in on those lucrative contact centre opportunities with help from Telarus. Amazing technology at your finger tips.

Doug Miller: Well, Telarus and our suppliers offer many powerful trading opportunities throughout the year. And here are 2 we are featuring this week that we hope you’ll plan to join if your spouse says, well, gee! Why don’t you hang the Christmas lights? Oh, darn! I’ve got a Telarus training. I need to attend tomorrow Wednesday at 11 am Eastern plan to join the fully managed cloud and Security options provider performiv

Doug Miller: for their 2023 wrapped webinar. This will recap the year’s most exciting trends. You’ll learn how to unwrap success and set yourself up for big wins in 2024

Doug Miller: performive will discuss the it, channel the in demand solutions, and show you the strategies you can use to find success both now and then. In 2024 there will be an interactive panel discussion and answers to attendee questions. It’s the performive 2023 wrapped Webinar. It happens Wednesday at 11 Am. Eastern.

Doug Miller: and then on Thursday, November sixteenth, at one PM. Eastern Epic iO. Will present an exclusive look into their cutting edge solution, deep insights for connectivity which launches next month in December. This will be a true game changer for your customers. It’s a revolutionary platform that promises to transform how your customers manage

Doug Miller: their connectivity. Data. Epic iO’s deep insights is an open, a I plus IoT. Software suite

Doug Miller: that combines and enriches video external feeds and sensor data all into a seamless operational experience. This will be your exclusive preview coming up on Thursday from Epic I/O. It happens Thursday at one PM. Eastern.

Doug Miller: if you’re interested in making a little money, we are also featuring the following pair of tremendous promotional and spiff opportunities. first, up from 8 by 8 it is your chance to earn fourth quarter spiffs

Doug Miller: of up to 10 XMRR. That’s how many fingers I have. 10

Doug Miller: with 8 by 8 Uks and C. Kaz, and more with the add ons, new customer sales. Earn 4 XMR. Spiffs, plus

Doug Miller: 2 more for sea, Caz plus 2 more for qualifying U. Cas, and seek as deals

Doug Miller: plus 2 more for annual prepay and for existing customer sales. We haven’t forgotten you, either. There’s 4 XMR. Spiffs for qualifying. See cas plus 2 x more for annual prepay.

Doug Miller: and those 8 x 8 add ons, I talked about one x of intelligent customer, assistant, digital MR. And one x of conversation IQMR. R. If you know, you know, hurry! These 8 x 8 spiffs of up to 10 x

Doug Miller: are valid until December 30 first, that is such big money.

Doug Miller: And how about this offer? From T

Doug Miller: on their safe and secure plus bundle promotion?

Doug Miller: The bundles had a price change, and it’s been extended through March 2024. The price change was down by the way, these 3, bundled at and T services include managed at and T. Internet Backup

Doug Miller: Dns, security advanced with mobile protection for 2 devices and 3 licenses of At and T end point protection with sentinel. One.

Doug Miller: 25%. Savings are available on these bundles, and they are now available through March.

Doug Miller: So keep your clients safe and secure with these bundled offers from at and T.

Doug Miller: Now those are just 2. But we’ve got many available spiffs and promotions. Currently, in effect, you can see all of them at our website, promos.tellerus.com.

Doug Miller: I did that in 10 min really getting good at this. Maybe they should keep me around for a while.

Doug Miller: Your questions and comments are welcome in the chat window, for to day’s presenters. We are very excited to be joined by the experts from Logicworks. Their portfolio of public cloud services and solutions is now part of Rapidscale.

Doug Miller: and they are here to day with a game changing presentation. You really need to experience, we’re going to be talking about artificial intelligence, machine learning and big data services powered by Logicworks. These are some of the most compelling technologies in to day’s market

Doug Miller: and the certified A. IML. Engineers and data scientists can help your clients chart an a. IML. Path forward or simply optimize an existing infrastructure. Rapidscale empowers them to improve operations, create exciting new customer experiences and accelerate business performance.

Doug Miller: Rolla Dali is an A. IML. Architect and engineer, with Logicworks and Marin Wangard is the national programme manager at Rapidscale, and they both join us now with details and examples of these phenomenal new solutions. Hey, Rolla Marin! Welcome to the Tuesday call. We’re so glad to have you here today.

Maren: Thank you so much, Doug. We are so grateful to be here, and to have the opportunity to go ahead and highlight this very important part of our portfolio. AI, ml, is quite the buzz these days, and I’m really excited to have Rolla on here. Who is the expert I will defer to her to go through the content. But please ask questions as it comes up. We will gladly answer those and provide more insight. Rolla.

Rola: Thank you. Thank you, Mary. Thank you, Doug. I’m happy.

Yeah. So I’m here like, Doc said. And, Marian, I am a machine learning architect at Rapidscale, and I will be taking you through some of our offering and our expertise.

Rola: Through Ml, and big data next.

Rola: So in terms of agenda for today. I’m gonna start a little bit by talking about the machine learning life cycle just to give you a little bit of a background. I’ll talk a little bit about our team at Logicworks a little bit of our framework. I’ll talk about but a few use cases that. We’ve helped clients with, and then we’ll go through next steps. I’m trying I’ve intentionally left a lot of time for questions. Because I think that’s how we can customize this to your interests next.

Rola: So in terms of machine life, cycle machine like solve machine learning is a general purpose technology that can help solve a lot of problems. And the type of problems that it can solve are increasing by the day. Just to show you a little bit about how the process goes.

Rola: We first start by defining a problem. Right? We we wanna understand

Rola: the business, what it needs the problem itself, and how we would want to solve it. And so we

Rola: are able to frame that problem as to an analytics framework. So if you come as a business. And you say, well, I wanna increase revenue. There’s a lot of ways for you to increase revenue. You can increase gross sales, you can increase margins. You could decrease churn. There’s a lot of ways to go about the same task.

Rola: And so there is a step to understand exactly how we want to tackle that problem. Once we do that. And we decide on the metrics. We want to use the type of problems, the type.

Rola: the type of machine learning. We want to use. Then we go in and go through a data prep.

Rola: phase.

Rola: So we collect data, we engine, we do some data, engineering, some feature engineering if required, some data labeling and we set up the data in the way that it needs to be to serve us.

Rola: Then we go into model development. We select the model. We train the model, evaluate the model, and when the model starts to perform well, we deploy it and we serve it, and then we go into maintenance and monitoring, and often this is a an iterative cycle, which is why it’s showing us a not

Rola: in here. But this is to show that there is quite a bit of expertise and skills that are required to take machine learning all the way from a business model to the customers

Rola: next.

Rola: And so in our on our team Logicworks has curated a a team of diverse skills. We’ve got 7 people on the team with very different backgrounds. My background comes from academia. I spent 17 years in academic research

Rola: studying brain development.

Rola: and I’m interested in solving problems with machine learning. Amir is our Ml’s architect. He is. He was an aws employee for a leading long time. So he’s our aws expert on data and machine learning.

Rola: Dan has his master’s in large language models. So he’s our generative AI expert.

Rola: Elmer is a devops engineer who helps us build all the pipes that are required to support these systems.

Rola: Will is our data architect, big data architect. He thinks a lot about organizational maturity in data. And Mml as well. Austin is our data scientist.

Rola: and he likes to to think of of the business and how to help businesses. And Michael Yang, of course, who is our lead? Is an Ml. Ops, expert with enough breath to to cover the whole life cycle.

Rola: I’m showing that showing you this because it is very important when you go into Ml, production to have a team that can span the whole end to end life cycle within that team. Across this team we’ve got about 18 aws certifications 6 and a half masters, a Phd, and 60 years of collective experience in data. And AI

Rola: next.

Rola: So I’ll just talk a little bit about our Logicworks next, please.

Rola: We’ve curated a a framework, a Logicworks made of 9 steps divided in in 4 phases. So in blue, we’ve got our strategy phase

Rola: in a navy. We’ve got our validation phase. We’ve got the build phase in orange. And then the green in opt for optimization. So typically when we have again. And this stems from the la, the machine learning life cycle so that we typically start. If a company has not yet started their machine learning journey. We start with determining the business goals again. Tech serves the business you do not want

Rola: a system that works in isolation and then doesn’t solve your problem as a business. So we start there. We understand what the the business goals, what problem needs to be solved, how best to tackle that problem.

Rola: Then we go into step number 2, where we frame that problem. The problem, of course, is in business terms, and we need to bring it down to an analytics framework with metrics and numbers. And and so we have a face for that.

Rola: Then there’s a data sourcing phase where we set up the data, and then we’ve got a Poc where we run

Rola: a miniature model to see whether the idea is viable. There’s a lot of things that can go wrong in machine learning. It could be that the data is not good enough. It could be that the framing has not been done well. And so, because there’s an intense investment for Ml. In terms of cost and skill required. A poc can help you determine the viability of an Id.

Rola: Yeah, before you start to invest

Rola: and infrastructure. So the poc is very important. When the poc passes, it tells us that everything before that was done well. And so we can start building

Rola: the system for production.

Rola: Let machine learning like any other software system. Requires infrastructure. They, it requires compute storage, networking security. And so on. So we start to build these foundational pieces.

Rola: And because of the importance of data and machine learning. We’ve got a a special phase for just the data infrastructure. So we can build data, lakes, warehouses, lake houses and help you structure your data. in your organization.

Rola: there’s a phase 4. There’s a step for Ml operations. So we set up all the Ml, Ops

Rola: automations. and we can set up visualization and reporting

Rola: to monitor the business impact of the model once it’s in production

Rola: and then at the end. We have an optimization phase. We’ve helped many companies. Optimize cost security compute latency and so on. This tends to be. If if you’re of course, a company that hasn’t started. It’s machine learning journey. We recommend that you start at the beginning and and go through

Rola: of course, this is very flexible. So if you’re a company that already has done a Poc, you’re really happy with your model, you wanna productionize it. We’re happy to help you there. So we’ll start with the build. If you’re a company that is still looking at data governance. And how do you set up your data before you start your machine learning, we can

Rola: walk, we can go with step number 6 and help you. Set up your data. And if you’re a company that has

Rola: production productionize your model, and you’re seeing some sort of issues with it. It’s too expensive. It’s too slow. It’s not optimal. We can definitely help you with optimizing your production loads

Rola: next.

Rola: So this is some of this is an extension of the phases that I talked to before. This is. These are just an example. So in the strategy phase, we tend to

Rola: produce an assessment report that tells you where you are as a business some solution. Architect diagrams, total cost of ownership for you to know

Rola: how much it would cost. And then we can scope for the Poc in terms of the validation phase. These are just some examples of things we’ve done. So we have experience with image and videos with speech text, Llms, chat bots and this is by no means complete. These are just examples

Rola: for the build phase. We’ve got experience with data leaks. Etls, ml, apps Vi tools. And again, for optimizations, we’ve helped companies optimize security latency cost.

Rola: Next.

Rola: if you’re interested in in generative AI generative. AI has seen a huge increase in popularity in the last year, because of chat. Gp.

Rola: we’ve developed a tool called genflow. Gen. Flow is our pla, a platform that we developed here and Logicworks. That is no code. And it’s self hosted. So if you’re interested in large language models, building your own chat, bot

Rola: producing your own summaries or so on, but don’t have the expertise in coding these models or dealing with them. You could use Gen. Flow to help you it it

Rola: it has a graphic user interface. And then you go through and it helps you. Set up these systems with minimal code.

Rola: Actually, no code. The other thing that’s important about gen flow is that when you’re using a genitive AI model, often you are

Rola: calling a model through an Api, and that means that your data and that your prompt are leaving your own domain.

Rola: And that also means that the model is not yours. It’s not under your control. So the the nice thing about Gen. Flow is it is the model is self hosted. So you can choose any model from hugging face that you’re interested in and hosted inside your Vpc. Which gives you control over the model. But also, the data doesn’t leave your domain.

Rola: Gent flow has a lot of features so it can help

Rola: with model alignment. If you want to fine tune your model. It. It helps with guard railing to protect against certain outputs model evaluation. It can help you with deployment.

Rola: It does a a low testing on your model, and it recommends the exact compute instance on aws, so that you’re not over provisioning, for example. So Gen. Flow is is one of our products that can definitely help you if you’re interested in generative AI next

Rola: again, like I said, machine learning is a generative general purpose technology. It has really been a game changer to a lot of vertical and a lot of businesses, and these are some of the verticals that we have helped clients

Rola: from. So we’ve got a lot of experience with healthcare, digital health genomics.

Rola: in fact, I have a doctorate actually in from in healthcare and genomics. Hospitality, IT. Finance, legal aviation, science, scientific field. So whatever it is that your vertical is we can definitely help

Rola: next.

Rola: So I’ve chosen a few use cases to just show you some of the things that we’ve done. I’ve chosen 3 from different verticals just to illustrate.

Rola: Some of the type of problems, perhaps, that that can be solved

Rola: next.

Rola: So there’s a lot of text here. Oh, I had highlighted things in red for for, but I think those have been lost. But anyway, I’ll try to explain some of this without, you know, the text can be too much, but I kept it there because we’re sharing the slides. This is a client that is in the shipping and

Rola: supply chain vertical, although before you get started on this, I want my 7 extra points for having looked up the word demurage, and I know what it means now

Rola: I well, I’ll I’ll explain.

Rola: Oh, but good for you! Yes, what it is actually is that so once ships come into the the port they have to. There’s a fee for that. Pause that that that

Rola: for staying in the dock, and that those are called damage fees, and they’re quite costly. They’re about 30,000 an hour. And so our client wanted us to wanted to reduce the charges.

Rola: because, of course, ships can be delayed by weather, by by sea conditions, by things happening on the ship.

Rola: And so our customer wanted us to build a model that predicts the arrival of the ship. And so we went in. We collected data for weather data, sea condition data, ais, data, satellite data data from the ship. And then we built a deep learning model that can predict the arrival of the ship.

Rola: And we could we could actually predict the arrival of the ship within minutes, and so that kind of saved our client half a million dollars per year in fees.

Rola: Next

Rola: this client is in the healthcare

Rola: vertical the client is a medical clinic. That had many doctors, and of course, when when doctors meet a patient there are a lot of mundane, repetitive automated tasks that need to be done right? So there needs to be a report. The doctor needs to leave notes and so on for the future. And so

Rola: we were. The the client wanted us to build a model to automate a lot of these tasks between doctors, visits. And so what we ended up doing is that we the doctor would dictate their notes. Because voice notes are easier than typing and and much less time consuming. And then we built a model, a system that takes these notes. It transcribes them through Amazon medical transcribe into text.

Rola: and then that text is sent to GPT. 3 to create a report. And of course, what that does is it reduces the administrative load on doctors, on medical professionals, but it also produces very consistent reports for each patient across doctors for a clinic.

Rola: Next.

Rola: This is my last one. This is from the hospitality industry. Our client is 8 specialized tech company that serves restaurants. to become more data driven. And what they needed is to help restaurants

Rola: be more prepared by predicting sales for the week. So that how does that help? Well, if the restaurant has an idea of the amount of customers that would come in, they would be able to hire enough staff or not, depending on what they need. Buy enough

Rola: raw material to prepare and and make enough prep, right? So that kind of helps in operations. And so what we did again, we collected a historic data patterns from across the restaurants. And we built a model we. We used Amazon forecast, and we were able to predict sales up to a week. And of course, these models, these predictions are based restaurant by restaurant right? So they’re very custom to each restaurant and their historic

Rola: information.

Rola: And of course, that kind of we build everything. These are scalable systems. That kind of

Rola: help as business grows

Rola: next.

Rola: So what’s next? Well, we have

Rola: a $0 rapid data acceleration workshop. We’re happy to help. We’re happy to sit with

Rola: you at no cost to you to ha! Kind of help. You see where Ml. Can help your business.

Rola: And chart the path for you to do so.

Rola: The only thing I say, I think I think machine learning has proven its value in a way that you know II didn’t start by saying how machine learning is important. I think that’s self evident today. What I would say, though, that if, as a business, you wait for your competitors to prove value in your specific domain, you might be playing catch up for a really long time. And so start

Rola: experimenting with machine learning. You can small start small. You don’t have to start big. You can start with small pocs, with with just conversations of how and where Ml can help you. But definitely start because machine learning will transform

Rola: not only our businesses but our society, in in ways we have yet to imagine. Right

Rola: next.

Rola: And that’s it, that’s all I had.

Rola: We can take some questions.

Doug Miller: Sounds great, Rolla. That’s a very compelling presentation. Just to remind everyone we’re joined here today by Rola Dolly. She is an a I. And Ml. Architect and engineer with Logicworks which became part of Rapidscale

Doug Miller: marin Wangard is here as well. She’s the National program manager at Rapidscale. We’re talking about aim L and big data today. And you’ve probably heard a lot of things that you haven’t heard before. And I’m very excited about this. It’s time that we get farther into this each week. Rolla, in your first slide you introduced a number of people that are helping to power all of this.

Doug Miller: and I know that there were probably 7 or 8 partners out there who said, Oh, this is where I go. Get a coke or something, because they’re just introducing the staff they were introducing your staff.

Doug Miller: I this is the part that always amazes me is that when we come on and introduce people like this, who are architects, who are engineers, who have the type of educational and training background that they have. These are people that Tellerus partners have just essentially hired to help them in designing these solutions for their clients. What a powerful set of people that you have standing behind

Doug Miller: these products. And I just wanted you to emphasize again. Why it is that it’s so important that you have that partners are generally not going to understand how this comes together, but you’ve just provided them a team of people who does

Rola: well, exactly that. So I started by explaining. So the the machine learning lifecycle that I started with is a really summarized life cycle. It takes a a lot of skill

Rola: for a good machine learning team. And the point here is that

Rola: you can have a company that has 2 or 3 people who call themselves data scientists. You do want to know

Rola: that that team skills spans end to end right? It’s a large, it’s a long, it’s a multi step process. You want to know that you have skills in every part of that life cycle. You also want to know that these people, you know, are experienced. They have the education they need to have and are willing are able to help you in the ways that they need to.

Maren: Hey, Doug, if and rol, if I could just add to that the other piece? I think that’s really important, too, is we also have the folks on staff. If you, if you uncover an opportunity that you think might be a good fit for this. We do have this sales journey to take you through. As well to try to figure this out with the engineering team, so we will co-sell with you and make sure that we’re taking into account

Maren: all of the business goals and the outcomes that we’re looking for. All we’re looking for you to do is qualify, find opportunities, and we will be there with you every step of the way.

Doug Miller: So for those that are unfamiliar with the terminology, and I had to look up some of it, not only to merge, but there were some other words that you’d used as well for folks that are a little bit unfamiliar. Let’s talk a little bit about a a data lake versus warehousing, and some of the other common terms that we use in designing a solution for an aim. L, opportunity.

Rola: So the data, of course, is becoming bigger and bigger by the day. And this is what powers a lot of machine learning solutions. Of course, that that puts a strain of how do I put that data? And how do I organize that data?

Rola: And traditionally, a lot of data was dumped into data warehouses. These are massive relational data stores. Things. Think, think, think, oracle. Think Microsoft? You know, think redshift on on aws, for example. So these are massive relational data stores that allow massive parallel processing for you to access your data.

Rola: Of course, a lot of data today is not in the relational form. And so other models have emerged. The data lake emerged for partly for nonstructured data, but also because data sizes are so huge that data warehouses

Rola: are not always optimal. A data leak is a is A is a data store. Where you put your data? But of course, there are ways to index that data to know where it is. You know, if you if you have 1 TB of data.

Rola: looking for one file is is looking for for a needle in a haystack. So there’s specific ways in which you structure data. You federate data for people to know where the data is and how and so on. So the the words I use is these are 3 paradigms in data, storage data leaks.

Rola: data warehouses

Rola: and a new paradigm, a lake house which takes a little bit of a data what the data lake is. And a little bit what a warehouse is.

Doug Miller: So that helps with some of the terminology. Let’s talk about those data sets. So we had a couple of questions from partners Matthew asked in the Q. And a window is credit card database information such as the merchant processing servicing that many businesses do. Is that a good data set useful in some way for this type of solution development?

Doug Miller: Dennis joined here a little bit late to the call. But is also asking in the in the restaurant sales AI solution, what sort of data services or sources are being used there to provide the data needed.

Rola: So data. So Po, what what machine learning does is it tries to understand patterns in the data, right? There are patterns in data around us all the time. We are as a society. We’re moving as a society, as a business.

Rola: We’re, we’re we’re

Rola: living and that produces data. And that data captures patterns. And what machine learning tries to do is capture patterns in that data.

Rola: So any data. That is captured well enough can be a source of machine learning.

Rola: of course. What matters is, what? What do you want to know.

Rola: What is it that you would like to predict? What is it that you want to know? Is credit card information a good source for machine learning. Yes, it is. It is actually used every every time. Use wiper card. There’s an anomaly. There’s a machine learning alum, anomaly, detection. Mo model that is running in the background to make sure that that’s you. That’s why. A lot of our transactions today are pretty smooth

Rola: not because there’s less fraud, but because there’s machine learning models behind every swipe, making sure calculating what the risk of fraud is, and and if there is fraud, then you get a message that says, is that you? So yes.

Rola: finance data and and and credit card data can be a good source. In the restaurant business. It it was historic data of how many people they served in any particular day. And and what and what they had how many servers they had, and and so on. So it was data captured by the restaurant for day. The operations.

Doug Miller: This is going to be a little bit specific, but Jesse Wallace asked a question, and want to try to make something a little bit more out of it. Here we talked a little bit about the medical and healthcare industry with electronic medical record transcription.

Doug Miller: Jesse is asking a little bit deeper question if there were a machine learning opportunity that required a much larger data set. And he’s thinking of a very large

Doug Miller: amount of let’s call it DNA information informatics is the type of technology that we’re talking about here and what’s being made available

Doug Miller: large enough to handle a large data set such as this, that would also require some very specific applications, and probably a high degree of security. As well offload those to azure or other types of sources

Doug Miller: where that’s going to be used. How? I guess the larger question here is, how large and how specific can these solutions be designed depending on the industry and the opportunity?

Rola: So there is no upward limit. Just to give you an example of, you can run a model that that you know, that has megabytes of data if you need

Rola: But the models that are running today, for example, Chat Gpt is a model that has 1.8 trillion parameters

Rola: and is running on the corpus of most everything that’s on the Internet. They haven’t released what? What exactly it is. But these are petabytes and petabytes of data. So there is no upper limit

Rola: to the data itself or the model. Of course, the bigger the data, then, the longer the model would need to train. But there are ways to do that these days. You can accelerate you can, you know there’s compute instances, the Gpus. There’s a whole lot of ways

Rola: for us to deal with that. So there today there is no upper limit, not neither to data nor to models, but of course, the the bigger the data and the bigger the models, then the the bigger the cost and the bigger the the longer the time it takes to set these systems up.

Doug Miller: Many of our partners, of course, are being asked questions about AI in general terms. They’re being asked very specific questions about generative AI specifically, and how clients should be considering using that at this point. So for those that want to dip their toe in the water a little bit, but maybe aren’t ready to dive in completely, and may quite frankly have some fears and concerns about introducing

Doug Miller: generative AI applications into their business. What is a good way for a Telarus partner to have a conversation with their client about possible baby step applications for generative AI in their clients. Business

Rola: so generally. Yes, has become very, very popular. I think the the it said, there’s 1,500% increase since Chat Gp was released in November last year. It is very powerful technology, which is great. But again, things like genflow allow you to host the model and the data in your own domain, thus reducing the security risks. And

Rola: you know. So we’re seeing a lot of companies that are building not only chat bots to to automate certain things, but also assistance that help you go through confluence pages that are within your own company. And so there’s a lot of ways where you can safely experiment with these things without releasing your data. Without

Rola: you can re, you can start internally within your own company, within your own employees by automating a lot of things internally, and then when you’re comfortable enough, you can find solutions

Rola: for your customers. But there are a lot of ways today to start small to ha again, Gen. Flow. No code required no security risks, because it’s in your domain. And if you talk to

Rola: anybody at Logicworks. We’re happy to help you through these.

Doug Miller: Martin. I know you wanted to get in on this too.

Maren: Well, I you know I would also add to that, too, to kind of take it back up to an over level view, too, I think you know, with our product set at Rapidscale, including the AI and Ml. The biggest thing to remember is, we want to talk to customers about business outcomes. So what business outcome are they looking to solve. And then what we typically do then, is we take that business outcome, and we do an assessment of everything overall

Maren: and look for what products do we have to solve those issues? Whether it’s AI, whether it’s disaster, recovery, infrastructure, all of that stuff. We have such a broad portfolio of multi cloud solutions. So I would start with the overall business outcome conversation, and that get with our team about what we can solve for introduce us to the customer. We’re very good at looking at the whole picture. We are not

Maren: a quote shop in the sense of Hey, if you send me an email and say, Hey, I want to quote disaster, recovery. That’s not how we work. We work from a business outcome perspective. And you know, if it’s AI specific you want to look at will get you to that team and kind of go through what you’re looking to solve for and get you into that process right? Everything is business outcomes, though. Just remember that anything in this space business outcome business outcome business outcome.

Doug Miller: I’m glad you mentioned that. Matthew peck and Pi asked a question here which I think is appropriate. The questions have to be different, too, in terms of leading into an opportunity or sales like this? It’s not, you know. Hey? Can I tell you about AI as a service as he mentions there? But it’s as you mentioned. Let’s sit down and talk about the particular outcomes that your business is looking for, and how these technologies may be able to help you achieve that. Am I getting that correct

Maren: absolutely, I mean. And it’s like I said, our overall portfolio is so impressive, and the people that have the most success with us are the ones that bring the overall situation to us and allow us time to ask questions and really make sure. You know, I think it’s also important of note, Doug, I think you know, our 93% Nps score as voted by 1,500 agents is huge.

Maren: I also think our 4.8 Csat score is huge, too. We are a trusted partner with people that want to go down that solution selling business outcome discussion route.

Doug Miller: So as we’re getting close to the end of our time. Here we’re getting a lot of application questions that are coming in from partners. I’ll take a quick look at a couple of them here. Armizad, just sent us a note here, the call center with 300 agents. How do we approach them? What’s the value? Prop? Is it productivity cost savings business outcome? What will it cost, Roi?

Doug Miller: I’m guessing it’s all of those. But how do you start the conversation to get that sit down to talk about potential outcomes. Have you found something that’s particularly successful in this or other areas?

Rola: I think the where the barrier is for a lot of people a lot of people know. AI is powerful today, and but the barrier is well, what kind of do for me? Right? What are the possible

Rola: solutions that apply to my business and for for the call. For example, the call center, we see a lot of

Rola: people that a lot of companies that use AI as assistive.

Rola: You can use AI as a an assistive measure where you’ve got an AI running in the background, and the call. The the call agent is is running the show, but they have. They can tap into the system for things that they don’t know, or or to accelerate Job.

Rola: There’s a lot of call centers, actually, that where you’re fronted by the machine learning system, and they do everything except when they have a low confidence when the machine learning system doesn’t know how to do something, then it pings a human in the background to take over. And so, though then you lower your staff.

Rola: there’s different combinations. There’s different things to be done but anything that a human can do, a machine learning system is very likely to be able to do and then the other thing is that

Rola: machine learning scales much better than humans. Right? You have a good model.

Rola: I should start by saying, machine learning is not always the best solution. Everywhere it works it, it can work anywhere. It’s not always the best investment and return roi. So you have to think, what is the Roi? Is it worth that I invest? I make that investment. There are 3 cases at least where I tell people. It’s a no brainer. AI is a no brainer. When you have scale, humans do not scale as well as machine. So.

Rola: if you have a call center, if your business booms to double, you have to hire double those people, train double those people and find infrastructure for double. If you have a machine learning model that’s already running.

Rola: you can scale your compute and inference, and you’re good to go

Rola: when there are trends that change. So when some, when a, when a business depends a lot on changing patterns in the world like social media trends or things like that that’s really hard to do for humans to do. But machine learning does that really? Well? And when there’s really really complex tasks. So there’s a lot of

Rola: things that are really affected by so many things that it’s hard for us humans to predict all the time. And so if there is enough data, machine learning does really well in those situations.

Rola: But the machine learning can automate. A lot of tasks that humans do it can augment a lot of things we can do. Is it worth it? That’s always a question to ask. Yes, if you have scale. Yes, if you have complexity. Yes, if you have changing trends.

Doug Miller: So just in general terms, then what would partners expect to tell their clients if they’re asking, you know, how is this sort of a solution priced? What would a client end up paying? Not in terms of dollars, but in terms of? Is it an ongoing, recurring cost? Is it a one time? Consultation fee? How is this priced for a client.

Rola: Marin? Do you want to take that one?

Doug Miller: Sorry I had mute. I had mute on. I mean, you’re gonna love this answer. But it really depends on what we’re trying to solve, for it’s hard to predict the cost until we really get in and look at it.

Maren: yeah.

Doug Miller: So that actually leads in very nicely. Then to this, we want to have those conversations. And you’d mentioned earlier the rapid data acceleration workshops. And I want to talk a little bit about that quickly. Describe just a little bit more in detail what that is, how we initiate that. And then Myra and I know that if partners had additional questions you want them to get in touch with you, too correct.

Maren: Absolutely. I mean, I would say, start with me because I think the the general questions. I think there’s gonna be quite a few. Let me get you to the appropriate person. I don’t want to overwhelm the sales at Logicworks email box, but I will get. We’ll we’ll get my contact information out to everyone with that deck, and then we’ll take it from there also, just real quick to point out Doug.

Maren: We do have a webinar. That is a Rapidscale, webinar, where Michael Yang’s going to be speaking.

That is

Maren: Thursday. That you can also join for more more information as well. So we will have a ton of education on this, but happy to chat about any specific opportunities or questions as we go down this path.

Doug Miller: And then again, the information on the rapid data acceleration workshops. What should partners be looking to to do if they want to attend? That is, is that client facing as well, or is that just for partners? What is that?

Rola: That’s an offering where it’s client by client, this is not, you know. So if you’re interested in in machine learning, and you’re wondering how it can help your business we can set up sessions with

Rola: our engineers to help you. Custom your solution. So this is not a general, you know. This is a client by client session custom for you.

Doug Miller: This is just so exciting. You know, we we hear a lot of different conversations going on about AI generative AI machine learning big data. And quite frankly, there are a lot of people out there who have.

Doug Miller: you know, concerns about it. They don’t understand it. They. They wonder what the applications would be for their clients, and how they can discuss it intelligently. I think, if anything, what you have done is to ease a lot of concerns to day among our partners about. Hey, look! I can have this conversation. I can introduce an opportunity. You have tremendously expert people there who can sit down with my client and help not only design a solution, but reassure them that it will be safe.

Doug Miller: You’ve introduced a, a, a generative, AI solution, Gen. Flow which can be hosted on the customers domain, providing them with outstanding control making it easy for them with no code required. You’ve made this easy. You’ve made it secure and controllable. And I think you’ve gone out of your way to provide expert help

Doug Miller: for partners that wanna have this discussion with their clients. You’ve accomplished a lot here today. This is just tremendous. Partners, then, are able to access the information that you’ve provided today. And we’ll send Marin’s information out with that. If you request that I’ll I’ll throw the last word to you, Rolla and Marin. What do you want partners to go away with today beyond anything I’ve just said. Here.

Maren: Rola, any wrap up that you wanted to provide before I go.

Rola: I you know the same thing I said. At the end of the presentation machine learning is almost everywhere. Today. It’s almost seamless. You don’t feel it, but it is running. Every time you watch a show. There’s a recommender system trying to sell you more. Every time you swipe your card. There’s an anomaly detection making sure it’s you. Every time you look at your phone. There’s a a facial recognition software that is running to make sure it’s you. So machine learning is almost everywhere whether you’re aware of it or not.

Rola: And again, it’s going to transform businesses. And and there’s no better time to start than today. And again you can start small.

Rola: You don’t have to, you know. Start with a big, you can start with small pocs, or at least start to have those conversations of Where does Ml. Fit in my business, or how can I accelerate my business with Ml, and and we’re definitely happy to help.

Maren: Yeah. And I would add to that, too, from a full portfolio perspective. You know, with current economic climate, all of the things that we are seeing, we have a ton of customizable solutions based on the business case. So if you are interested in more information, please do reach out to me. I’ll get you in touch with your local resources as well as resources on the Logicworks. Rapidscale team. Aws, public cloud, private cloud, all of those solutions we have, and we can build something really special for your clients

Doug Miller: wonderful.

Doug Miller: It has been so much fun to have the 2 of you on the call today. Thank you for this very powerful discussion. It’s a technology that I’m becoming more familiar with have a lot of interest in. And I know our partners do as well look forward to hearing back from you in a few months as to how it’s going rolled, Dolly. And and Marnard. Thank you so much. If you have questions, please get them to us. We’ll make sure that they get to them as well.

Doug Miller: Thank you for the powerful presentation, thanks to each of our Telarus partners for attending and participating to day. We will be back next Tuesday with more. I’m Doug Miller. The voice of Telarus have a great selling week. Everyone.