Interview with Vidhi Chugh from Reltio

As a Head of Product, an author, a course mentor, and a multiple award winner, you’ve seamlessly managed a diverse range of responsibilities. Could you elaborate on your journey with AI and your experiences at Reltio?

Data is core to AI; and so is my journey. I come from an entrepreneur background where I used to observe my father make business decisions and later map the outcome to declare a good or bad decision.

There were various variables impacting the decisioning as well as the end result. Now, if you relate to Machine Learning – these are our attributes that help surface what has worked in the historical training data.

Those were my early days of understanding cognitive behavior and deciphering the science behind it to learn the underlying phenomenon.

This inquisitive nature led me to see the opportunity and make a career in data – think of how we can make breakthrough innovations at scale leveraging AI. 

I have been a data-centric professional throughout; believing in frugal data science – a journey which becomes easier if only we have the right trusted view of data.

Data tells us the story; but that story can only be right if it is coming from high-quality data. It is this mission of data-centricity that resonates the most in my journey with Reltio.

The AI industry is one the fastest growing industries, what is your approach to staying updated on the latest advancements in AI and ML?

Networking, where industry is headed, exchanging thoughts and ideas, learning through how others are iterating and succeeding

Besides, reading posts through LI, medium, podcasts, following veterans, reading research papers to understand cutting edge innovations

In the context of Reltio’s commitment to transforming siloed data into unified and trusted information, what strategic trends do you foresee in the AI and data management landscape?

AI will play a big role in alleviating all data pains, the underlying theme would be that no data professional would need to look at charts, triage through multiple data sources, do all figuring out on their part – be it related to looking for any data consistencies, being on look out for any data deviations (think of SMEs saying – this is not how my data typically looks like, seems something odd, but what has changed underneath)

AI will resolve all such cognitive overloads, simply because it learns all such patterns from the data. I call it, “making data work for your data”.

The AI models, that have gone through volumes of data, will assist at every step of the data management cycle. It would present recommendations all across, so all data users will have to do is, to leverage their domain knowledge and choose to act on those recommendations.

Given the critical nature of data in Reltio’s offerings, how do you approach ethical considerations in the development and deployment of AI for data unification and management?

Core of ethics is trust – and we at Reltio are leading at this frontier – right from how the data is collected, stored, processed and transformed into a trusted view – ethics are embedded by design into our systems.

We are GDPR compliant and have data residency in place, as part of which, we honor organizational requirements of managing their data within their specific locations.

Our focus on providing trusted data adds another dimension to ethical consideration, allowing the businesses to make decisions on reliable data.

Another key aspect of ethics is it is an organization-wide initiative — it weaves into the fabric of the organization, and leadership plays a key role in prioritizing it across the board. I am proud to mention that we at Reltio promote the ethical lens in all our data management practices.

AI adoption hinges on model explainability — and we are actively working towards that, so that our customers, data stewards, in particular, do not just get the potential matches, but are also presented with the reasoning as to why the machine learning model declared a pair a match.

With a focus on delivering interoperable data, how do you foster collaboration between AI/ML teams and other cross-functional teams within Reltio’s product development structure?

We strongly believe in establishing systems and processes for effective collaboration. In a global team structure like ours, we take measures like:

• Periodic connects within PD team and cross-functional teams to ensure everyone is aligned to the same overarching goal
• Check-ins for brainstorming on blockers and charting the path to success
• We function as core-focussed groups ensuring least resistance, speed and agility
• Having a transparent and autonomous culture adds great value too.
• We promote async communications reducing the overload of meetings
  
  

Have you encountered any challenges in deploying AI solutions within the context of data unification? If so, how were these challenges addressed?

Data unification, in itself, is a very complex space – that has the potential to unlock the true intrinsic value of data for all data-driven companies (which is most of the world now). I can not think of any company that can not benefit from a unified data view.            

Data coming from operational sources arrives in different schema, requires data cleansing mechanisms, need standardization protocols — which quickly becomes overwhelming. We have velocity packs for different entity types across varied market segments like Life Sciences, Healthcare, B2B, B2C and more that accelerate the time to create a consolidate data view and generate value from it thereby.

Getting access to quality, relevant and accurate data to ensure AI learns the right pattern. Ensuring data integrity in continuously flowing data in production is an industry-wide known challenge. Being a data management company, we are able to address it more effectively, as we have deep domain expertise in curating and building world class data products.

Additionally, being a SaaS company, we have a leverage of scale that allows us to handle complex AI workloads in an efficient manner.

With a focus on driving growth and improving efficiency through data unification, what key performance indicators (KPIs) or metrics do you find most valuable in measuring the success of AI and ML features within Reltio’s products?

Helping improve the data stewards’ productivity by 10X – is the linchpin of our success story. Productivity can further be elaborated with:

• Generating high quality matches, saving their time on manual effort spend in analyzing numerous pairs
• While matches are core to our business, there are various data dimensions that are precursor to matches, such as data quality, anomalous records. We assist data stewards in every step of their journey of working with the connected data.
• All of these converge can be tied to a single KPI – the rate at which their backlog of potential matches frees-up
• Lastly, we believe in treating our customers with delightful experience, which includes, providing them with simplified and easy to use UI – that is powered with seamless integrations with varied data sources.
  
  

How do you unwind and rejuvenate to show up better for yourself, your team, and the organization?

I learned early on in my career, that you can not ask for work-life balance, you create that harmony yourself. Hence, I inculcated a habit of being fully present in whatever I do — be it family or work. For example, I never scroll through my phone during family time and likewise at work. I do not work at split attention.