Intellyx Video with Jason English for Zoho Analytics
In this interview, Clarence Rozario, product manager of Zoho’s Business Intelligence Suite, talks with Intellyx analyst Jason English (‘JE’) about their latest Analytics for All release, which is infused with AI for conversational interaction (Ask Zia), high volume data processing, and predictive analytics ML models tuned to the enterprise’s business model and vertical.
Watch below or visit our YouTube channel here: https://youtu.be/4yuarSOSKAM
Complete transcript of the video below:
JE: Well, Hi — I’m Jason English or Jay-E, a Director at Intellyx, the analyst firm dedicated to digital transformation. We’re currently working on our Zoho Development Stories series. And part of that, I’m glad to have on Clarence Rozario from Zoho. Clarence, why don’t you tell us a little bit about your role at Zoho and what you’re doing with analytics?
Clarence: Yeah. Good morning Jason. Thanks for having me on this interview. I’ve been with Zoho for the last 24 years and I head the business intelligence product line, taking care of the entire product suite, that’s what I do. And currently we are working on a new launch of this business analytics product, of the BA product at Zoho. It keeps me busy and it is an exciting time.
JE: Oh yeah. I guess we could start off with basically the evolution of the analytic space. I mean, obviously a trend, analytics have been around for a long time, since the early days of business intelligence and similar tools. But in your opinion, how has the analytic space evolved?
what are some of the new challenges that companies face today for improved visibility or decision support help?
Clarence: Okay, maybe I’ll just start with a small historical context of how this whole space evolved. The last 20, 25 years.
So the first 1.0 version of BI was typically simple reporting tools used along with spreadsheets and some database querying and things like that. Very non-interactive, done by some developers, based on requirements. So that’s how it started, the 1.0 version.
And then the big enterprise BI vendors came into space, the likes of SAP and Business Objects and quite a lot of vendors who came in and then said that for all the enterprise data that is getting collected, we need to have a very sophisticated reporting tool.
They called it reporting. And business intelligence was also coined at the time, because it was giving intelligence on the business data that was collected. But still, it was more developer friendly or IT friendly, because you needed to know coding. You needed to know how to write SQL queries.
The database dialect as a skill. So even if an executive wants a report, they have to get connected with the IT persons to create the reports and then start consuming data. So the creators were IT, and the consumers were the executives or business users. Gut the relevant, it was specifically focused on enterprise business applications.
That’s a core target of the particular product line at this the time. And then in the early part of 2000, the 3.0 version came, where the idea was to really democratize access to insights, make insights accessible to everyone. So the new paradigm of self-service BA tools came into picture.
The focus was to enable any business users to really start analyzing data through easy-to-use self-service tools so that they can get insights, because insights can be a great enabler for all type of business users, not just executives alone.
And that’s when tools like Zoho Analytics and others like Tableau came into being. So we launched our product in 2009, that’s a 3.0 version. And it’s been there almost 15 years now. So now we are in 2024. So I would call the new generation is basically the AI powered BI tools or analytics tool. That’s what the trend is.
It is all about trying to infuse AI into the BI spectrum, because naturally AI plays a significant role in BI, because BI is a place where data gets accumulated from a variety of applications and data sources. So when the data is huge, AI can play a significant role because it can have quite a lot of data to crunch to come up with models and analyze them and give insights.
So, the version 4 of the BA analytics space is something of adopting AI, in a such a way that it can deliver powerful insights on one side. The other is also to accelerate adoption of BA further inside the enterprises or organizations. That’s what is happening today, and it is the trend now.
And that’s what most of the players, including Zoho are trying to achieve in this new era of analytics and BA tools.
JE: Yeah. It’s interesting. We’ve really seen analytics go from this kind of bottom up approach where you had upper management, basically telling underlings, give me some reports.
Generate me some reports from this tool, and now it’s completely switched over on its head. We’re basically at the root level where people are actually doing line of business work. They’re looking to analytics to find out “what should I do next? What can I do to help forward the goals of the company and try to have that effort aligned with what’s actually happening at the management level.”
So yeah, we’ve definitely seen the space turned on its head in the last 10 years. And it’s great to see this trend. I know part of your slogan is, you have “CRM For All” and you have other aspects of your suite, but what does “Analytics for All” mean to Zoho, and how has your approach to analytics changed in the last few years?
Clarence: See when we started on this particular product line, our focus is to really deliver a self-service analytics platform. Where it’ll be an enabler for all types of business users, not just the executives, or not the IT function. So our mantra has been to really offer analytics as a key enabler for all types of business users.
That’s why we call it Analytics-for-All. Right from the front-end users to the executors. Everyone should have access to insights. And how can we build tools and capability that will make it happen, we term ourselves as a platform, which is providing analytics for all types of business users.
That’s one dimension. The other dimension, when you get into the enterprises— Especially the BA personas, data personas. We also wanted to give a platform such that all types of data personas will also be in a position to use the platform for getting insights. Be it your data engineers who are trying to get data, analyze, set up data models.
Or the business analysts or data analysts who analyze data, or data scientists and data scientists who build ML models. Right on the platform for specific analytics workloads. So even that segment of the audience should also get addressed by the platform offering an extensibility capability that we offer.
There are two dimensions here. All types of business users and all types of data personas should be addressed by the platform so they feel empowered to get insights faster.
JE: Yeah, there’s definitely an ‘any business user mode’ and there’s an ‘expert mode’ where, if somebody is familiar, if they want to use Python or write their own queries, they can still do that or they can have some limited guidance.
The analytics tools, helping them extend the existing sets of queries and iterate on those or maybe pull things together in a way that they used to have to make all these custom joins and things that just aren’t are necessary. It’s just toil. It doesn’t really add any value to the analytics process.
Clarence: I would say yeah, that’s part of the process to get insights.
JE: Zoho already has a great reputation for helping startups really get to go from startups into a larger size companies, because they can build their business around the whole suite of tools. Obviously you’re starting to have a lot of different types of customers, maybe some who come in, the mid-to-large size companies that might already be running at a billion or two billion dollars and maybe have a lot of different analytics packages that are already in place.
They already have existing core systems. They probably have SAP in place or some Oracle systems and they probably have Tableau or some other reporting tools and, and all these things. So how can Zoho help those customers?
Clarence: That’s a good question. Actually what we today offer as a BA platform, it’s an end-to-end BA platform that addresses all the business all workflows with respect to BI.
The data integration where you try to integrate data from the data sources, the data management and preparation area, which helps people to prepare high quality data. Then in terms of analysis, where you can help people to analyze and come up with reports, dashboards, and KPIs. And also look at even collaboration and access control.
The entire spectrum of what you try to do with a BA workflow is covered as part of the platform. We are an end-to-end BA platform. Right. And we are also introducing a data science block today in the new version that is coming up this month. Which will help even data analysts to build ML models for their analytics workload right inside the single platform.
It’s an end-to-end integrated BA platform. So that’s a great enabler. Offered in multiple deployment models. Be it in the cloud or on premise or private cloud. So all this flexibility is offered.
So if you look at today’s enterprise they do deploy multiple BI tools, they do deploy multiple data tools.
The vendors typically have a siloed approach. They might be strong in visualization. They might be strong in data management. They might be strong in data science. So a typical enterprise is trying to pick and choose the strong vendors in each one of the areas.
But that adds up to the complexity, because of the fact that you have to manage multiple tools and to really get that entire workflow done and get the insights as an outcome. So that always adds complexity. Or although initially it might be a great enabler, over time, in terms of cost, in terms of complexity, in terms of maintaining the system over time, it really becomes a great burden.
So that’s why our approach is, we want to have the notion of self service because always cemented. We want to really make it easy for people to adopt. At the same time, offer a very comprehensive BA platform so that the complexity can be solved because it’s easy to use. And the other thing is they don’t have to pick and choose multiple vendors for different functionality in the BA workflow.
They can use this, that’s a unique differentiator. The ease of use, being a self-service platform and an end-to-end BA platform. That is one way by which we try to position within the enterprise who are already using multiple tools, but they are really burdened by complexity, burdened by the cost in terms of managing and also by the cost they are trying to pay to the vendors, because over time they keep paying more and more.
So that is one key way by which we try to showcase our differentiation into the enterprise who are already into BI. That’s one thing.
The other dimension is, if you look at our platform, although it’s an end-to-end BI platform, it’s also a composable platform. It has different building blocks, offered as an integrated platform, as well as, as a separate component, if you want to consume them as a separate unit. Some of the independent components that we offer, makes us this entire BI platform, includes the data preparation and management block, which helps you to get data.
By prepare, what I mean is cleanse the data. Transform the data, enrich the data so that you can have high quality data that you can set up. That preparation component alone can be consumed. If you have a requirement where you want a tool or a product, to really get high quality data from the data sources that you might have.
For example, you might have data that is coming from an ERP system, or that may be coming from a CRM system, but they might have quality issues. They have to undergo some transformation and enrichment. So you need a very strong data preparation component. You can also use this component, the data preparation component that is offered as part of the BA platform, to do this job and then push the data, the outcome, the data that is coming out of the system into your operating system like a database or a data warehouse, or it could be going back to your ERP systems or CRM system.
So we can use this component as an independent service to really pull in data, prepare, enrich, transform and push it back to your operation system. That could be your ERPs, CRMs, as well as databases. That’s another way by which they can start consuming or using so on.
The other composable component is what we call Embedded BI. If you look at enterprises, typically they have quite a lot of custom applications they build. And they want analytics as part of their application.
For that, they cannot go and invest on writing the entire stack of analytics or implementing the entire stack of analytics to really power insights in their custom application they may be doing. They may be building a small inventory application. They might be building an order tracking application.
That is, there’s quite a lot of applications that they do. So what we also offer is what we call Embedded BI component. Where the entire power of the BI platform can be taken, rebranded and integrated as part of your custom application, to drive or deliver insights in the context of the application workflow.
That’s another component that we offer. On one side we have a very end to end BA platform, which is very easy to use, right, which confers the entire power of whatever that you need as part of the BA platform. On the other dimension, we also have composable components or ‘Lego blocks’ that you can take up and independently use it for different functional requirements.
So that is how we are an enabler for enterprises who are already into the BI or who might be already using multiple BI tools, but still have the gaps.
JE: Yeah. That’s a good thing. I think a lot of existing packages that have been out there, it’s almost like they want to brand the experience.
And that creates a little bit of a problem as far as adoption, because really you want the whole process to seem like your own. Especially if you’re offering it to your customers, you’d want that experience to seem very seamless to the end customer, if you are the one offering that value to their customers if that’s what they want. Interesting part because you you’ve mentioned AI a few times.
So how is it really infused into the suite? And do the analytics users know that they’re using AI, or, at what touch points is AI coming into play?
Clarence: Actually our objective since we started integrating AI into the BA platform has been that we want to take the power and complexity of AI, and then blend it as part of the BA platform that we are offering, so that the end users are not bogged down by the complexity, but they get the essence, the entire power and benefit of AI in a seamless way. We don’t want to push down the complexity of AI to our end users.
We want to just get the benefit out of it. That’s what our objective has been. That’s our approach because as I mentioned, we are a self-service BA platform. And our key differentiator is ease of use–we want everyone to adopt the BA platform.
This is a base vision that we always align ourselves to. So what we continue to do is adopt AI across the entire spectrum of all the components inside the BA platform in such a way that it powers the functionality and makes it easy for people to adopt.
So that could be in the data preparation layer, for example, when you’re bringing data. And you want to cleanse the data. So can I use AI to make you do the job much easier? Okay. By something like an assistant where you can say ‘fix all the quality issues on my email columns, remove all the duplicates.’
It could be a small command that you make, but it can understand what you’re asking and then go and automate the entire job and fix the issues that are there in an email column. That’s an infusion of an AI assistant to make it happen.
The other thing is when you start analyzing data.
Can I have an AI bot to keep asking questions? “Show me the sales trend for the last 24 months.” It might be a question that you’re asking. And it will understand your question and give you a nice chart, which will show you the sales across all the last 24 months as a nice lined graph. This is what as a company we call Ask Zia.
Ask is basically asking questions to our Zoho Intelligent Assistant. Zia is ‘Zoho Intelligent Assistant.’ So it’s an AI component, or what we’d say, an AI Assistant, embedded across the multiple workflows of BI. Right in the data integration layer or the visualization analysis layer, you can use Ask Zia to ask questions and get insights.
That’s one aspect. The other aspect is if you’re looking at a chart. If you’re looking at a chart or a report or a dashboard, you would like to get the top insights of the report. That the report is trying to convey to you. But that might take some time. You have to keep analyzing the chart or a report and then say, okay, this is the upward trend, this is the downward trend, these are the changes, this is the percentage of change.
All these things need some analysis. You have to put down the pen and paper to come up with the insight that the chart is trying to convey. So to make the job easier, right, we are using AI there, where we are using natural language generation, where looking at the chart, looking at the data, We’ll try to bubble up the top insights that you might be interested to know from the report.
So that makes it easy for you to get the insight that you required immediately, and how do we do that? We analyze the entire data, look at the trends inside the chart or a report, and then say, these are the top things that you might be interested in. This is the percentage of change between the last month and this month, which is not directly shown in the chart, but this is something that you might be interested to know.
As a nice narration in English slides format. So it’s simple, or verbose. You can go through it and you will get to know the insight that is there behind the chart. Right. So that’s also an AI model at the backend. It’s called Natural Language Generation. It’s a GenAI component where it is generating content.
Right. What I talked about is the previous case is basically Ask Zia, which is, once again, a generative content where you are asking questions and then getting insights like similar to a Chat GPT. Yeah. You are asking. I’m getting insights as answer. The second thing, what I’m talking about is the narration engine, where you’re saying, I give you a chart, give me a nice summary of the top insights.
You give a big paragraph of text to Chat GPT and say, give me a summary of the document. It’s another way of trying to use generative AI in the context of the BA platform. So the way we are trying to do is, Adopt AI, especially generative AI and the latest AI capabilities, in such a way that it will empower users to make decisions faster.
Our goal is to really accelerate adoption of analytics using AI. That’s the way we are blending. That is one side of it. The other side is also basically, apart from trying to give you a generative AI capability, we also give you prepackaged ML models, like for example, you might want to do prediction.
You might want to do sentiment analysis of your data. So can I give you prepackaged ML models that are available so that you don’t have to write the models for yourself. So that you can do a predictive forecasting, or you might do an anomaly detection on the data that is there. We have to find out outliers from the data that is coming up.
There are quite a lot of prepackaged ML models that we try to give to the users so that they don’t have to worry about writing a model to really do a prediction to do an anomaly detection of the data. This is another dimension by which we make it easy.
This is all about making it easy for people to use insights or do analysis through the prepackaged generative AI components or ML models that we offer. That is one dimension.
The other dimension, there will be always pro developers who will say, “okay give me all the things, but we are not sufficient. I want to write my own ML models. I want to write my own code. I’m a hands-on guy.”
I have special requirements which are not solved by the BA platform directly. We also give a pro-developer-friendly studio inside the product so that they can build their own ML models right inside the platform, without trying to buy another tool or use another component to do ML models, which are specific for the enterprise requirements.
On one side, we predominantly focus on bringing AI, especially generative AI models into the platform to make it easy for the users to adopt our BA platform. On the other side, we also enable users to build custom ML models through a pro code, pro developer friendly model so that even if the platform does not address all your requirements pre built, you can still extend the platform and get your requirements solved through AI.
So these are two ways.
JE: Yeah, so there’s a probably a much smaller set of use cases out there that people would need to customize just for a specific business when most of the pages in the playbook for analysis are pretty much well known. Now, there’s a lot of different types of standard reports.
And then, what becomes harder there is maintaining the context of a report as it evolves. So that if I’m asking a question, could you help me generate this report? Then you want to drill into the details of it. And then you want to do refinement.
You don’t want each one of these queries to be a separate workflow. Because then you basically lose context of what you’re trying to explore. And that’s what I think is very powerful about the AI model associated with analytics is that you can get this context that continues forward as you do your analysis.
It helps you make decisions because it’s maintaining that that context. And you can always return to the workflow again later. Or tell it to start over if you’re if you’re not heading down the right path. But it’s an interesting capability that’s baked into your platform, I’d say .
So how, in your experience, have customers dealt with this huge incoming wave of all analytics, data metrics, both from IT and from the business side, that are just washing over them? I mean, there’s definitely a volume problem and a signal to noise ratio problem.
So how are you helping customers out there deal with this high volume and like seemingly stochastic randomness that happens when you look at so much incoming data?
Clarence: See, there is a data velocity and diversity problem. Because there is data that has been generated or collected by businesses today and is almost doubling every year.
We see it as an important challenge, being a vendor, observing how customers are facing challenges. One, data diversity and data velocity is an important problem. When you talk about diversity, the variety of business applications from which data is generated is increasing.
By one count, they say a particular organization uses anywhere between 50 to 100 business applications to really run their operations. So that keeps collecting data. And that keeps collecting. And not only business applications, but also machines generate data. Okay. Then the variety of data also includes not only structured data, but you’d also have unstructured data, you have streamed data, you have your video content, quite a lot of input that is coming into the businesses.
There’s a velocity problem. There’s a diversity problem. How are you trying to help them? How as a product vendor, what has been our take? So one thing that we try to do is strengthen our data integration layer. Because at the end of it, the BA platform should be in a position to consume the data that has been thrown at it.
And that means it has to have a variety of sources from which it can get data from. Be it from your databases, be it from your data lakes, be it from your business applications. Be it from your files and feeds. It should be in a position to get data from a variety of sources.
We have a very comprehensive data integration, which offers connectors to a variety of data sets: your databases, data lakes, business applications. And a lot more. So more than 500 different end points from which you can get data immediately without doing much work. So you connect to it, it’ll start consuming the data.
That’s one thing. The second thing is if you look at business applications like ERPs, like NetSuite or Salesforce or Zendesk and quite a lot of popular business applications. We not only help people to get data, we also help them to model the data automatically because when you get the data, if you start, if you want to really analyze the data and get to insights.
You have to properly structure the data that has been fetched from the systems. So we go deeper, not only as a BA vendor help people to get data, which is what most of the vendors do. They give connectivity, but they stop there. You have to do the modeling. You have to really structure the data in a logical way.
There’s a complicated job. When you bring data from something like Salesforce or NetSuite, we get data, and we automatically model them. And not only stop with that, we also give you prepackaged reports and dashboards for that particular departmental vertical.
For example, let’s say you’re bringing it from Salesforce. As soon as you integrate with data, I will give you hundreds of reports and dashboards for sales analytics. So that you can jumpstart yourself in terms of analyzing the data immediately. Without knowing, without worrying about ‘what metric should I start looking at?’
How can I report, what dashboard should I create? Then the next thing, what we need to do is assume that you are bringing data from two different applications. You’re bringing sales data from your Salesforce CRM and your help desk ticketing system data from your Zendesk system.
These are two different applications. Coming from two different vendors. It means it adds another level of complexity, you have to bring data, but you have to know how these two data connect to each other or talk to each other because these are two different applications, two different models.
So what we do as a vendor. Given that we in Zoho build quite a lot of business applications, we understand the models. We understand how multiple applications talk with each other in terms of the model layer. So what we try to do, is we do the plumbing automatically. We do the connectivity between the Salesforce data and the Zendesk data automatically.
We call them data blending, because of the fact that we do smart blending without the users having to worry about that. You can create a simple report where you can look at a nice graph, where you want to plot the number of the amount of revenue that you’re making from a single account, along with the number of tickets that the particular account has raised to you, in a single chart, across months.
Where the amount of revenue is, the revenue data coming from Salesforce, whereas the help desk ticketing system information is coming from Zendesk. blended together, seen in a single shot. This is a simple example. That is the level of depth that we do, to help people to get end-to-end business insights automatically.
So that’s how we try to help businesses who are getting data from business applications. On one side, we give you connectivity. The other side, when you’re getting data from business applications, we go deep and call this basically unified business analytics because we unify data. And also help people to get a unified understanding of how the data is really playing across to get them the perspective in terms of end-to-end business insights, as soon as the data comes into the system.
So that is the other way we try to help people to manage the data. And the third dimension, in case we don’t help you with all the connectivity options or going deeper, we also help them to build their own connectivity options easily. With no code wizards where the people need not have to write code.
They can use a small no-code assistant. Or even build a custom data connectivity option to bring data from their data source. That could be your drives, it could be your database, or it could be your applications. So in an easy way they can try to bring in data.
And as soon as the data comes in, as I mentioned already, we have that modeling capability automatically. We have the blending capability automatically. So that users need not have to worry about how to really model them, how to really blend the data, and how to really analyze data. So it will be assisted as you’re trying to bring even your custom data source.
That’s the level of what is the simplest sophistication which is brought into the platform. People can manage the complexity in terms of not only the diversity of data, but also the volume of data that is being processed.
A small point to add with respect to the data velocity.
We have also built a very deep data warehousing engine as part of the platform. It can process and store huge amount of data in the order of billions of records and terabytes of data. That’s a component that is inbuilt as part of the platform. So even if you’re trying to push billions of records to the system or in terms of terabytes and petabytes of data, you can still crunch the data and help you to analyze that.
JE: Right. So, to really manage the huge volume of data that is getting generated, that’s how be an enabler to solve this complexity.
Yeah, it’s very cool to have so much integration capability, but you also have the complex backend to handle the heavy load part of it if needed.
So Clarence, how can customers take advantage of Zoho Analytics or the new Zoho analyti cs, how do they get it and activate some of these capabilities?
Clarence: Yeah, actually Zoho Analytics can be accessed from our website https://zoho.com/analytics. That’s our website. So they can go there and then start sign up there.
If they want to use the hosted cloud version of the product, they can sign up and start with a free trial for 15 days. Immediately. Try out all the capabilities that I talked about. It’s a single platform. So we don’t want to restrict people to try it out. And even if they want to have an extension in terms of where they want it valid for one month, two months, we always give a free extension for people to play around the product.
That’s how they can get started. And if they want the on-premise, or on our private cloud deployment, they can also download the product because that we give multiple distributions. If you want to use Zoho’s cloud, you can sign up, and start using it.
If you want a on-premise edition, you can always download and that again gives you a free trial for you to keep trying out the product. And both these distributions are equivalent because all the capabilities that you see in the cloud is also available on premise. There is no disparity between the capabilities.
And we also give you a, always free edition. There’s a limited free edition, which you can always use. It’s a freemium model. You can always try the product any for any time longer. And if you want to upgrade to a paid edition anytime, you can also do that.
So that’s how people can try it out. And the new version that I’ve talked about, especially the the big launch that we are making as part of this 2024 release. It’s getting launched on September 12th, which will be available across all regions in the cloud.
To follow it up, will also have an on premise edition by October end with its new capabilities. Available. So they can get started and the new version is available as a beta.
In case you want to try a beta version of the new capabilities, it is available right now. As you ge t inside the product, you can try the beta of the new versions. In case you want to try out the new capabilities that are coming up as part of the the new version.
JE: All right. Thanks, Clarence.
Yeah, it sounds like, once again, Zoho makes it easy to get access to some really advanced functionality here. Alright, well, it’s been very great talking to you, Clarence, and I invite anybody to go check out the Zoho Analytics website there just visit https://zoho.com/analytics.
Download it, try the demo. It’s free to do, and if you’re an existing customer of the Zoho One suite, you’ll also probably see that module in there as we speak. Sounds great. Well, thanks for joining me, everybody. And it’s been great to have another Zoho development stories Intellyx video with you.
So take care and I’ll, I’ll see you later.
Clarence: Thank you, Jason. Thanks for having me on the call. Thank you.
©2024 Intellyx B.V. Intellyx is editorially responsible for this content, and no AI was used to generate text or ideas. At the time of recording, Zoho is an Intellyx subscriber.
