Soham Mazumdar is the Co-Founder and CEO of WisdomAI, an organization on the forefront of AI-driven options. Previous to founding WisdomAI in 2023, he was Co-Founder and Chief Architect at Rubrik, the place he performed a key function in scaling the corporate over a 9-year interval. Soham beforehand held engineering management roles at Fb and Google, the place he contributed to core search infrastructure and was acknowledged with the Google Founder’s Award. He additionally co-founded Tagtile, a cellular loyalty platform acquired by Fb. With 20 years of expertise in software program structure and AI innovation, Soham is a seasoned entrepreneur and technologist based mostly within the San Francisco Bay Space.
WisdomAI is an AI-native enterprise intelligence platform that helps enterprises entry real-time, correct insights by integrating structured and unstructured knowledge via its proprietary “Information Cloth.” The platform powers specialised AI brokers that curate knowledge context, reply enterprise questions in pure language, and proactively floor traits or dangers—with out producing hallucinated content material. Not like conventional BI instruments, WisdomAI makes use of generative AI strictly for question era, guaranteeing excessive accuracy and reliability. It integrates with current knowledge ecosystems and helps enterprise-grade safety, with early adoption by main companies like Cisco and ConocoPhillips.
You co-founded Rubrik and helped scale it into a serious enterprise success. What impressed you to depart in 2023 and construct WisdomAI—and was there a specific second that clarified this new course?
The enterprise knowledge inefficiency drawback was staring me proper within the face. Throughout my time at Rubrik, I witnessed firsthand how Fortune 500 corporations had been drowning in knowledge however ravenous for insights. Even with all of the infrastructure we constructed, lower than 20% of enterprise customers truly had the best entry and know-how to make use of knowledge successfully of their each day work. It was an enormous, systemic drawback that nobody was actually fixing.
I am additionally a builder by nature – you possibly can see it in my path from Google to Tagtile to Rubrik and now WisdomAI. I get energized by taking up elementary challenges and constructing options from the bottom up. After serving to scale Rubrik to enterprise success, I felt that entrepreneurial pull once more to deal with one thing equally bold.
Final however not least, the AI alternative was not possible to disregard. By 2023, it grew to become clear that AI may lastly bridge that hole between knowledge availability and knowledge usability. The timing felt excellent to construct one thing that might democratize knowledge insights for each enterprise person, not simply the technical few.
The second of readability got here after I realized we may mix the whole lot I might realized about enterprise knowledge infrastructure at Rubrik with the transformative potential of AI to unravel this elementary inefficiency drawback.
WisdomAI introduces a “Information Cloth” and a set of AI brokers. Are you able to break down how this technique works collectively to maneuver past conventional BI dashboards?
We have constructed an agentic knowledge insights platform that works with knowledge the place it’s – structured, unstructured, and even “soiled” knowledge. Moderately than asking analytics groups to run experiences, enterprise managers can straight ask questions and drill into particulars. Our platform will be educated on any knowledge warehousing system by analyzing question logs.
We’re appropriate with main cloud knowledge providers like Snowflake, Microsoft Cloth, Google’s BigQuery, Amazon’s Redshift, Databricks, and Postgres and in addition simply doc codecs like excel, PDF, powerpoint and so on.
Not like standard instruments designed primarily for analysts, our conversational interface empowers enterprise customers to get solutions straight, whereas our multi-agent structure allows complicated queries throughout numerous knowledge programs.
You have emphasised that WisdomAI avoids hallucinations by separating GenAI from reply era. Are you able to clarify how your system makes use of GenAI otherwise—and why that issues for enterprise belief?
Our AI-Prepared Context Mannequin trains on the group’s knowledge to create a common context understanding that solutions questions with excessive semantic accuracy whereas sustaining knowledge privateness and governance. Moreover, we use generative AI to formulate well-scoped queries that enable us to extract knowledge from the completely different programs, versus feeding uncooked knowledge into the LLMs. That is essential for addressing hallucination and security issues with LLMs.
You coined the time period “Agentic Information Insights Platform.” How is agentic intelligence completely different from conventional analytics instruments and even customary LLM-based assistants?
Conventional BI stacks gradual decision-making as a result of each query has to combat its approach via disconnected knowledge silos and a relay workforce of specialists. When a chief income officer must know learn how to shut the quarter, the reply sometimes passes via half a dozen fingers—analysts wrangling CRM extracts, knowledge engineers stitching recordsdata collectively, and dashboard builders refreshing experiences—turning a easy question right into a multi-day mission.
Our platform breaks down these silos and places the total depth of knowledge one keystroke away, so the CRO can drill from headline metrics all the way in which to row-level element in seconds.
No ready within the analyst queue, no predefined dashboards that may’t sustain with new questions—simply true self-service insights delivered on the pace the enterprise strikes.
How do you guarantee WisdomAI adapts to the distinctive knowledge vocabulary and construction of every enterprise? What function does human enter play in refining the Information Cloth?
Working with knowledge the place and the way it’s – that is basically the holy grail for enterprise enterprise intelligence. Conventional programs aren’t constructed to deal with unstructured knowledge or “soiled” knowledge with typos and errors. When data exists throughout different sources – databases, paperwork, telemetry knowledge – organizations wrestle to combine this data cohesively.
With out capabilities to deal with these numerous knowledge sorts, worthwhile context stays remoted in separate programs. Our platform will be educated on any knowledge warehousing system by analyzing question logs, permitting it to adapt to every group’s distinctive knowledge vocabulary and construction.
You have described WisdomAI’s growth course of as ‘vibe coding’—constructing product experiences straight in code first, then iterating via real-world use. What benefits has this method given you in comparison with conventional product design?
“Vibe coding” is a major shift in how software program is constructed the place builders leverage the ability of AI instruments to generate code just by describing the specified performance in pure language. It’s like an clever assistant that does what you need the software program to do, and it writes the code for you. This dramatically reduces the guide time and effort historically required for coding.
For years, the creation of digital merchandise has largely adopted a well-recognized script: meticulously plan the product and UX design, then execute the event, and iterate based mostly on suggestions. The logic was clear as a result of investing in design upfront minimizes pricey rework through the costlier and time-consuming growth section. However what occurs when the price and time to execute that growth drastically shrinks? This functionality flips the normal growth sequence on its head. Abruptly, builders can begin constructing practical software program based mostly on a high-level understanding of the necessities, even earlier than detailed product and UX designs are finalized.
With the pace of AI code era, the hassle concerned in creating exhaustive upfront designs can, in sure contexts, grow to be comparatively extra time-consuming than getting a fundamental, practical model of the software program up and operating. The brand new paradigm on the earth of vibe coding turns into: execute (code with AI), then adapt (design and refine).
This method permits for extremely early person validation of the core ideas. Think about getting suggestions on the precise performance of a characteristic earlier than investing closely in detailed visible designs. This will result in extra user-centric designs, because the design course of is straight knowledgeable by how customers work together with a tangible product.
At WisdomAI, we actively embrace AI code era. We have discovered that by embracing fast preliminary growth, we are able to shortly take a look at core functionalities and collect invaluable person suggestions early within the course of, dwell on the product. This enables our design workforce to then concentrate on refining the person expertise and visible design based mostly on real-world utilization, resulting in more practical and user-loved merchandise, sooner.
From gross sales and advertising to manufacturing and buyer success, WisdomAI targets a large spectrum of enterprise use instances. Which verticals have seen the quickest adoption—and what use instances have shocked you of their influence?
We have seen transformative outcomes with a number of prospects. For F500 oil and fuel firm, ConocoPhillips, drilling engineers and operators now use our platform to question complicated effectively knowledge straight in pure language. Earlier than WisdomAI, these engineers wanted technical assist for even fundamental operational questions on effectively standing or job efficiency. Now they’ll immediately entry this data whereas concurrently evaluating in opposition to greatest practices of their drilling manuals—all via the identical conversational interface. They evaluated quite a few AI distributors in a six-month course of, and our resolution delivered a 50% accuracy enchancment over the closest competitor.
At a hyper development Cyber Safety firm Descope, WisdomAI is used as a digital knowledge analyst for Gross sales and Finance. We diminished report creation time from 2-3 days to only 2-3 hours—a 90% lower. This remodeled their weekly gross sales conferences from data-gathering workout routines to technique periods targeted on actionable insights. As their CRO notes, “Knowledge AI brings knowledge to my fingertips. It actually democratizes the info, bringing me the ability to go reply questions and transfer on with my day, fairly than outline your query, look forward to any individual to construct that reply, after which get it in 5 days.” This skill to make data-driven selections with unprecedented pace has been notably essential for a fast-growing firm within the aggressive id administration market.
A sensible instance: A chief income officer asks, “How am I going to shut my quarter?” Our platform instantly presents an inventory of pending offers to concentrate on, together with data on what’s delaying each – reminiscent of particular questions prospects are ready to have answered. This occurs with 5 keystrokes as an alternative of 5 specialists and days of delay.
Many corporations at present are overloaded with dashboards, experiences, and siloed instruments. What are the commonest misconceptions enterprises have about enterprise intelligence at present?
Organizations sit on troves of knowledge but wrestle to leverage this knowledge for fast decision-making. The problem is not nearly having knowledge, however working with it in its pure state – which regularly consists of “soiled” knowledge not cleaned of typos or errors. Corporations make investments closely in infrastructure however face bottlenecks with inflexible dashboards, poor knowledge hygiene, and siloed data. Most enterprises want specialised groups to run experiences, creating vital delays when enterprise leaders want solutions shortly. The interface the place folks devour knowledge stays outdated regardless of developments in cloud knowledge engines and knowledge science.
Do you view WisdomAI as augmenting or ultimately changing current BI instruments like Tableau or Looker? How do you match into the broader enterprise knowledge stack?
We’re appropriate with main cloud knowledge providers like Snowflake, Microsoft Cloth, Google’s BigQuery, Amazon’s Redshift, Databricks, and Postgres and in addition simply doc codecs like excel, PDF, powerpoint and so on. Our method transforms the interface the place folks devour knowledge, which has remained outdated regardless of developments in cloud knowledge engines and knowledge science.
Trying forward, the place do you see WisdomAI in 5 years—and the way do you see the idea of “agentic intelligence” evolving throughout the enterprise panorama?
The way forward for analytics is shifting from specialist-driven experiences to self-service intelligence accessible to everybody. BI instruments have been round for 20+ years, however adoption hasn’t even reached 20% of firm staff. In the meantime, in simply twelve months, 60% of office customers adopted ChatGPT, many utilizing it for knowledge evaluation. This dramatic distinction reveals the potential for conversational interfaces to extend adoption.
We’re seeing a elementary shift the place all staff can straight interrogate knowledge with out technical expertise. The longer term will mix the computational energy of AI with pure human interplay, permitting insights to seek out customers proactively fairly than requiring them to hunt via dashboards.
Thanks for the nice interview, readers who want to study extra ought to go to WisdomAI.
