AiThority Interview with Venkat Rangan, Co-founder and CTO at Clari

 Tell us about your career journey. What led you to AI?

I got my start in tech as a Software Developer. After completing my undergrad in Engineering and a graduate degree in Computer Science, I joined a product company, building a bundled hardware product with a significant software component.

I first realized the power of AI when I founded Clearwell Systems in 2004 with Andy Byrne, who is also my CEO and Co-founder at Clari today. The goal behind Clearwell was to automate the back-office work related to litigation and securities issues by applying AI. Our technology was able to analyze vast amounts of unstructured data to augment search and retrieval in the context of electronic discovery. We quickly achieved a run rate of $100M and were acquired by Symantec in 2011.

This first foray into AI made us realize Machine Learning could be brought to areas that were ripe for transformation. That’s when we honed in on Revenue Operations, the most critical B2B business process. Revenue teams — Sales, Marketing, and Customer Success — were disconnected, operating in their own silos and misaligned throughout the buyer’s journey. Because traditional CRM systems depended on manual data entry, Revenue teams were operating without visibility into Sales and customer activity, and therefore couldn’t effectively prioritize Sales opportunities, manage pipeline risk or make accurate forecasts. And this is still true for many organizations today.

To sum it up, the genesis of Clari was a recognition that revenue is a process, and a methodical approach to applying Analytics and AI, that has the potential to fundamentally change the way companies manage their revenue generation activities.

We hear about Revenue Operations a lot these days in the context of growing a company. What is the challenge here and how is AI helping?

Growing revenue is a fundamental need for every company. The more we interact with our customers, the more we see companies looking at revenue as a process, as opposed to just an outcome.

The fact is, most companies strive to create new paths for generating revenue, but each new approach involves a critical set of decisions. Even companies with established revenue streams have to answer questions like: Should we invest in a new Marketing campaign to generate top of the funnel for the next quarter? Should we institute an incentive program to increase sales through indirect channels? Each of these decisions requires data analysis and careful consideration. And only by analyzing past outcomes can companies gain real-time visibility to create an action plan.

As I said previously, revenue teams often tend to operate in their own silos, with their specific tech stack and staff, resulting in disconnected operations. This is why Revenue Operations is a true challenge for many companies — and it’s a technology problem as much as it is organizational. Each team, rep and individual generates tons of business activity data and signals (both internal and external) daily, which is difficult to analyze especially in the absence of historical data.

Given the volume and complexity of data, modern AI algorithms like Clari’s can detect patterns that are relevant and surface them as key insights. This is especially important if these analyses have to be performed repetitively, and are needed by every individual in the organization. A well-designed data acquisition pipeline combined with feature engineering and automated training of models, generation of predictions, and delivery of relevant insights up and down the hierarchy in an organization, is extremely critical for realizing the full value of AI.

Clari specifically addresses forecasting. What makes this a critical component in the revenue chain?

At Clari, we believe making revenue predictable is critical to the long-term success of a business. If you have a reliable forecast, cross-functional teams can take that number and design their programs with greater predictability. AI-enabled forecasting enables teams to predict with confidence where you are likely to end 30 days, 60 days into the quarter. We see forecasting as a combination of achievement towards a goal, accurate visibility to inventory, and accurate predictions.

The criticality of this becomes evident at the end of the quarter. The few deals that you were able to identify early in the quarter that helped make quota — and the programs you put in place midway through the quarter that converted a best-case deal into a committed and closed deal — all highlight the importance of granularly understanding the forecast and having the entire team engage across every aspect of the revenue process.

Read More: AiThority Interview with Seth Siegel, North American Leader Artificial Intelligence Consulting at Infosys

How would a company use Clari’s technology?

Clari’s AI and Automation capabilities help companies manage critical revenue growth and retention processes such as current quarter forecasting, out quarter pipeline creation, deal and rep inspection, account and rep activity tracking, churn, upsell, cross-sell, and more.

As a first step, Clari automatically captures all of a company’s contact, business activity, and prospect engagement data through a range of data integrations with dozens of systems, including CRM (Salesforce, Microsoft Dynamics 365); Email (Office 365, Gmail); Marketing Automation systems (Marketo, HubSpot); Sales Enablement systems (Highspot, Chorus) and more. We then load this data into Clari’s Time Series Data Hub, a proprietary time-series database technology that automatically captures and snapshots Sales, Marketing and Buyer activity signals.

Our Time Series Data Hub implements a data warehouse dimension model that allows slicing and dicing the data across many facets. In addition, Clari’s AI and Machine Learning technology processes historical data and builds appropriate models that capture the behavior of both the customer’s journey as well as your internal team’s handling of the revenue process.

The real-time nature of the system provides insights that are ‘fresh’ which enable teams to make revenue projections and identify areas where Sales might be at risk or where a customer might be open to making additional purchases. It gives insight into critical information like whether a deal is at risk; where the quarter stands; how much pipeline they need; etc.

What other areas in Revenue Operations are being transformed by AI/ML?

Besides forecasting, a number of areas of revenue operations are benefiting from the ubiquitous availability of AI. Lead scoring is a well-developed area where AI technologies process large volumes of buyer intent when they visit public web properties, respond to marketing programs, attend webinars or download assets. Simply pre-qualifying leads and ranking them provides a much more targeted list to work with for inside sales teams looking to set up the first sales meeting.

Another area that we see emerging is the assignment of Sales quotas — fair and equitable — by considering territories, existing prime accounts, and emerging competitive threats. We also see AI applications leverage historical data and provide solutions in areas such as finance and territory planning.

 What areas outside of Revenue Operations is AI/ML impacting that you’re particularly excited about?

Conversational AI, whether implemented in Chatbots or Intelligent Voice Assistants, seem to be becoming commonplace. Significant advances in Voice recognition combined with query understanding and search/retrieval have produced remarkable practical applications with the likes of Siri, Alexa and Google Home solving real-world needs. This can have a massive impact on digital reach to underserved populations, and bring the benefits of the digital economy into their lives.

Read More: AiThority Interview with Nicole Silver, Vice President of Marketing at Button

What do you look for when building an AI/ML engineering team?

At Clari, we place a lot of emphasis on teamwork. As the complexity of data and associated AI problem space expands, we look for team members who can work in a cross-functional setting with product managers, designers and software engineers. Many of the AI/ML technologies have moved from the research phase into more broadly available toolkits, in packages like TensorFlow or Amazon SageMaker.

So, we expect Data Scientists to understand the overall business problem we are solving, analyze available data, evaluate technologies, identify toolkits, and build robust, scalable and effective solutions, deployable in a production setting. We also look for data scientists who understand engineering complexities in delivering the solution practically.

 What advice would you have for a budding AL/ML practitioner?

I think successful AI applications get their beginnings from deeply understanding the use cases where these insights can be utilized. In addition to developing skills in pure AI techniques like Deep Learning, it is helpful to understand deployment delivery vehicles and ways in which AI insights can be incorporated practically. In the process, you will get to analyze the data availability, feature engineering, model selection, performance, interpretability of results and on-going improvements.

Tag the one person in the industry whose answers to these questions you would love to read:

I would be interested in thoughts on these topics from Andrew Ng, who has been my idol, and a pioneer in expanding the understanding of Machine Learning and AI among the community.

Thank you, Venkat! That was fun and hope to see you back on AiThority soon.