Concentric AI develops data security governance technology that helps organizations discover, understand, govern, and protect sensitive data across cloud, on-premises, and SaaS environments. Its platform uses proprietary AI and language models to understand data content and context rather than relying primarily on traditional pattern matching, and its Semantic DLP capabilities extend that intelligence to public AI applications, including shadow AI use. Pulse 2.0 interviewed Madhu Shashanka, Concentric AI Chief Scientist and Co-Founder, to learn more about the company’s origins, its Semantic Intelligence technology, and how AI is reshaping data security.
Madhu Shashanka’s Background
When asked about his background, Shashanka shared:
I can distill my entire background into two words: luck and intelligence.
And when I say intelligence, I’m not talking about my own! I’m talking about the topic I have always been interested in.
I became fascinated with AI during my bachelor’s degree, pursued neuroscience in grad school so I could better understand human intelligence, and have had the good fortune of applying AI for real-world applications across a variety of verticals and diverse settings, from corporate research labs, large enterprises, and technology startups.
How Concentric AI Started
When asked how the idea for Concentric AI came together, Shashanka explained:
I knew Concentric AI co-founder & CEO Karthik Krishnan from a prior cybersecurity startup where he headed products, and I led machine learning.
As the company grew, I left and moved to Austin to kickstart ML and AI efforts within Charles Schwab.
Before long, Karthik reached out to discuss a gap he had been thinking about in the market for protecting sensitive enterprise data, especially unstructured data such as documents, emails, etc.
It came at the right time, as deep learning as a research field was making massive strides in processing and understanding text.
When I built a prototype leveraging the state-of-the-art at the time and saw the results, I knew right away this was something I had to do.
And the rest is history.
Favorite Memories
When asked about his favorite memory working for the company, Shashanka said:
Too many to choose from!
There are several events etched in my brain, like the day we got our first check, the day we opened our first office, the day we closed our first deal, or even the day I got mugged in San Francisco walking back to my hotel after a founder working dinner months before we formally started the company.
But my favorite memories are all with the people who joined us on our journey.
Every person, current and former, played such an important role, and I’ll cherish those memories forever.
Core Products And Features
When asked about Concentric AI’s core products and features, Shashanka explained:
Semantic Intelligence is the foundation of Concentric AI’s data security governance platform.
It enables organizations to discover and understand sensitive data across cloud, on-premises, and SaaS repositories.
Unlike traditional approaches that rely primarily on regex, trainable classifiers, or other pattern-matching techniques, our proprietary AI and language models understand both the content and context of data, helping security teams identify sensitive information more accurately and precisely.
The platform continuously monitors data for risk and enables security teams to take action, including moving, deleting or archiving, and classifying data, as well as managing permissions.
This helps organizations reduce unnecessary data, address exposure, and maintain stronger governance as their data environments evolve.
Semantic DLP extends that intelligence to public AI applications.
It gives organizations visibility into the AI tools their employees are using, including unsanctioned or shadow AI, and identifies the users and applications presenting the greatest risk.
Security teams can establish granular guardrails for sensitive data shared through AI prompts with inline categorization and policies that can warn, block, or redact sensitive information before it leaves the organization.
Prompts, responses, and policy violations are also tracked, providing the context needed to investigate incidents and support compliance.
How AI Is Reshaping Data Security
When asked about challenges in the sector and how Concentric AI is addressing them, Shashanka said:
The answer to both questions is the same: AI!
The entire cybersecurity domain has been turned upside down because of AI at an unimaginable pace.
AI not only gives bad actors a new powerful weapon but also massively exacerbates risk, especially risk around sensitive data.
But AI is also a big part of the answer to these challenges it has created.
We are at the forefront of leveraging the latest advancements and are continually innovating on behalf of our customers to keep their data secure, wherever it lives, however it travels, and whatever AI does with it.
How The Technology Has Evolved
When asked how Concentric AI’s technology has evolved since launching, Shashanka explained:
I firmly believe in first-principles thinking and relentless innovation to continually get better.
Since we launched before people knew what LLMs were, AI has dramatically improved in capability, and we have leveraged these capabilities every step of the way to push the boundaries of the value we can create for customers.
And along that journey, we have developed and continue to develop a significant body of IP that differentiates us in the market.
The latest example is the recently announced availability of image models to detect certain sensitive documents such as driver’s licenses, passport pages, etc., with distinctive visual signatures.
We have several exciting new features in development that I can’t wait to announce in the coming weeks and months.
Customer Success Stories
When asked to share specific customer success stories, Shashanka highlighted several examples:
Concentric AI helps organizations across more than a dozen industries discover, govern, and protect data across on-premises, cloud, and SaaS environments.
The results range from uncovering sensitive data that traditional tools missed to dramatically reducing the time required to investigate insider risk and govern AI use.
Here are several examples.
Public Health: A regional public health agency processing thousands of COVID tests each day was running six servers to support a competitor’s data classification tool, yet they still lacked visibility into inappropriate access to protected health information, or PHI.
During a proof of concept, Concentric AI identified unprotected health data residing on shared drives.
The agency replaced the six-server deployment with a single server and reduced the time spent on compliance and data governance by more than 75%.
Technology: A 10,000-person electronic systems design company had previously experienced trade-secret loss involving departing employees.
Investigating potential insider activity across SIEM and endpoint logs required the equivalent of 10 person-months of effort.
Concentric AI identified a senior engineer uploading proprietary algorithms to cloud storage during their final week with the company.
Insider-risk investigations that previously took 30 to 60 days were reduced to hours, giving the security team a far faster way to identify and investigate potentially risky activity.
Asset Management: A 400-person investment management firm had banned AI tools company-wide after a peer organization experienced a confidential-data leak involving a public chatbot.
Concentric AI provided visibility into how employees were using Copilot, ChatGPT, and Perplexity, including the types of sensitive data being shared.
The organization could then establish controls and approval workflows to identify and address risky uploads.
Instead of relying on a blanket ban, the firm gained a practical way to govern AI use while allowing employees to use these tools with appropriate safeguards.
Insurance: A national benefits carrier and absence-management administrator processes millions of disability and leave claims, many of which arrive as scanned, handwritten PDF forms.
Concentric AI identified U.S. Social Security numbers in handwritten, scanned documents, a task that traditional pattern-matching, text-extraction, and OCR technologies could not reliably handle.”
The organization gained visibility into sensitive information hidden within unstructured and difficult-to-read documents, strengthening its ability to govern data that conventional detection approaches could miss.
Competitive Differentiation
When asked what differentiates Concentric AI from its competition, Shashanka explained:
I cannot truly and faithfully answer this question without knowing details about my competitors.
As a builder, that is secondary to me. My singular focus is on addressing real challenges and delivering tangible business outcomes that create value to our customers.
Having said that, what I do know is what our customers tell us.
Several of them have come to us having been disappointed with other products, whether it’s from legacy players such as Varonis, platform vendors such as Microsoft,, or any number of existing vendors in adjacent markets with “me too” offerings who treat data and AI security as bolt-on additions.
And that has everything to do with what makes our company special: our people, period.
Future Goals
When asked about Concentric AI’s future goals, Shashanka said:
We want to be the tool and solution of choice for data and AI security governance, across whatever type of data, structured, unstructured, and anything in between; across however it is being used, at rest, in motion, or in use; and being processed by any human or machine entities, whether users, internal or external, or AI assistants, or fully autonomous AI agents.
Coherence And The Organizational Impact Of Agentic AI
When invited to discuss another topic, Shashanka highlighted his forthcoming book:
I’d like to share a personal project I’m excited about, my forthcoming book titled “Coherence: The Competitive Advantage AI Can’t Buy.”
It is about the organizational implications of agentic AI, aimed at business leaders.
It argues that as AI makes execution cheap, the scarce resource becomes organizational coherence — the capacity to deploy many autonomous systems without descending into chaos — and that this is now the real competitive constraint.
It is slated for release on November 17th. Follow along for updates on the book’s website at coherise.com/books/coherence.
