Founded in 1996, Zoho Corporation is an Indian multinational technology firm developing web-based business tools and is headquartered in Chennai, India. Additionally, the company has a presence in 11 locations with offices across the globe. Zoho Corp recently announced its plan to foray into capital goods manufacturing with capital investment to set up an R&D centre in the Kongu belt region (comprising present-day parts of western Tamil Nadu, southeastern Karnataka and eastern Kerala). 

The company offers smart software for businesses to grow. To understand better, Analytics India Magazine got in touch with Clarence Rozario, Director of Product Management at Zoho Corp. 

AIM: How does Zoho implement AI/ML/analytics to drive business growth? Can you share customer success stories or use cases?

Clarence: At Zoho, we build the software required. In this case, we have Zoho Analytics, which is a more than a decade-old self-service BI & analytics platform. Typically, we don’t take up projects for implementing analytics in our customer organisations. That said, we help prospective customers build POC prototypes using Zoho Analytics and assist with troubleshooting and providing technical support to our customers. We also have a partner network that assists our customers with implementation work.

As far as the role of AI in our Business Intelligence (BI) Suite is concerned, both Zoho Analytics and Zoho DataPrep use AL/ML technologies. Zoho Analytics has ‘Zia Insights’ that leverages the capabilities of our AI assistant Zia to provide automated insights on business data through smart narratives.

• Users can also converse with Zia and pose questions. Zia understands the context. Let us say you ask Zia, ‘Give me recurring revenue for India this year’, Zia will get you the related number for this year and the specified region. Then, when you subsequently say ‘by product’ or ‘by month’, Zia understands you are actually asking ‘Give me recurring revenue for India by product’ (or by month) and comes up with the appropriate report. This can be looked upon as one example of the power of conversational AI in analytics. 
• Another AI-backed capability we offer for users is predictive analytics in the form of custom forecast models. We also provide an AI-assisted automated generation of business insight reports. Furthermore, Zoho DataPrep includes an AI-powered transformation engine that provides intelligent suggestions to enrich data with 250+ transforms, such as sentiment analysis, keyword extraction, language detection, and more.

We have more than 13,000 paying customers worldwide. One can read some of their customer success stories here.

AIM: Can you share with us some of the new innovations and additions you have made recently?

Clarence: Zoho has launched a set of updates for its flagship offering, Zoho One—the operating system for business. The updates include six new apps, three new services and seven major platform enhancements in Zoho One. Zoho One leverages the capabilities of Zia, Zoho’s AI assistant, and Zoho Analytics to allow its users to provide and predict insights across the organisation, enabling confident decision-making. New innovations and additions in Zoho One on this front include:

• Zoho DataPrep: Our AI and machine learning-driven self-service data preparation tool that was launched in July 2021 is also among the six new apps included in Zoho One. If the users are preparing data from third-party apps or any other sources of their choice, DataPrep, powered by machine learning, is there to help business users model, integrate, cleanse, enrich, transform and even catalogue data before making it ready for analysis.
• Work Graph:  This new back-end service from Zoho is an industry-first for business software, helps in mapping interactions between people, resources, processes and systems by studying signals and their strength across the board to build a business-wide work graph — specific to each individual working within the organisation. The outcome of a work graph can be observed in users’ day-to-day productivity across multiple apps.
• Embedded and Conversational BI: Zoho One features both embedded and conversational analytics to enable decision-makers to drill down into their data and glean cross-departmental insights, all with the help of natural language commands using Zia Insights. We are providing more than 1,500 pre-built analytics reports and dashboards, good enough to make critical business decisions with greater accuracy and speed.
• Enterprise Search: Zoho’s actionable, organisation-wide search, which is powered by Zia, can now easily understand natural language requests. Moreover, the natural language-powered search will lead to more accurate data discovery across teams and functions.

AIM: What kind of skills do you generally look for while hiring data scientists/ML engineers?

Clarence: At Zoho, we typically hire fresh graduates out of college. We also have an in-house training programme, Zoho Schools of Learning, wherein high school and diploma graduates are inducted and trained for 18-months, after which they are hired. There’s also a lot of on-the-job learning that happens during which they further develop any specific skillsets required for the role. So, lateral entries of people with experience are minimal. 

In general, working with AI/ML requires at minimum a good understanding of multiple programming languages like Java/Python as well as technical concepts like applied mathematics and statistics.  

AIM: What programming languages and tool stacks are used? What kind of deployment challenges wrt AI and ML does your team face?

Clarence: Java and Python are the predominantly used programming languages by our data science team. Different tools are used for different purposes. Apache OpenNLP and Apache Spark are used widely, to name a few.

Zoho being a SaaS business apps provider, all our 50+ apps are hosted in the cloud. Our cloud stack is proprietary, and we host our apps using our own data centres. So, we do not face deployment challenges. Instead, we have incremental builds going every week—sometimes every day—simultaneously for multiple products.

AIM: What do you prefer — your own in-house R&D team or outsourcing the innovation?

Clarence: At Zoho, we have always believed in developing all our apps in-house. We are a very patient, engineering-driven company and believe in staying for the long haul.

For Zoho Analytics, we have had our R&D team working on it since 2006. Then in 2009, we started off with Zoho Reports, a reporting service in the cloud. Now, with the release of Zoho Analytics 5.0 (available in the cloud, on-premise, and embedded) and Zoho DataPrep in July 2021, we have a well-rounded BI and Analytics platform. We strongly believe this was possible because of having our R&D team in-house.

AIM: How do you see the landscape of AI/ML evolving in India with regards to your domain?

Clarence: Regardless of our domain, we feel that the AI/ML landscape in India is evolving fast. We expect it to evolve even faster because:

• the entry-level barrier for AI/ML is coming down, 
• good frameworks will become more commonly available, 
• an increasing number of companies will start using AI/ML to solve their business problems.