Reliable data is the foundation of confident leadership, yet in too many organizations, it remains an invisible liability. Silent errors, stale records, and undetected inconsistencies quietly undermine strategies, derail analytics, and compromise AI models, often surfacing only after the damage is done.

Anomalo is solving this challenge at its core. Co-founded by Elliot Shmukler, CEO, and Jeremy Stanley, Chief Scientist, the company is redefining data quality through powerful AI-driven intelligence. Its autonomous platform scales seamlessly across the largest data warehouses and lakehouses, continuously monitoring, investigating, and resolving issues, from obvious errors to subtle “unknown unknowns” before they can impact critical business decisions.

Going far beyond traditional rule-based checks, Anomalo acts as a true autonomous data system for the agentic enterprise. It doesn’t just alert; it understands context, surfaces what truly matters, and builds deep, proactive trust in the data that powers analytics, models, and strategic choices.

Rooted in values of ownership, radical transparency, customer focus, and intelligent execution, Anomalo empowers organizations to move from constant firefighting to lasting data confidence. In an era where every major move depends on data integrity, Anomalo gives leaders the clarity, reliability, and peace of mind to innovate faster, decide bolder, and build lasting competitive advantage.

At CIO Bulletin, we had the distinct honor of interviewing Elliot Shmukler, CEO of Anomalo. He offered powerful insights into how his company, guided by a culture of radical transparency and deep customer focus, is setting a new standard for data integrity.

Interview Highlights

What was the key moment or challenge at Instacart that led to the founding of Anomalo, and how did that experience shape the company’s vision for autonomous data systems?

Before founding Anomalo, I was the Chief Growth Officer at Instacart, and my cofounder, Jeremy Stanley, was the VP of Data Science. Instacart, at the time, and likely still today, was an intensely data-driven organization. Every major decision was made using data, and the team was full of machine learning experts and people with PhDs in logistics optimization. Data was very important and was constantly used to run the business.

So, it was particularly surprising when data quality would let us down. One particularly memorable event was when I arrived at the Instacart offices one morning, only to be told that orders from Costco, an important retailer on the Instacart platform, were down by 50%. The entire product, engineering, and analytics team was in fire-fighting mode—trying to figure out what was going on. They were checking everything: Was our checkout flow working? Did we break the search experience somehow? And on and on.

Eventually, we figured it out, and it ended up just being a data quality issue. The way the Instacart platform works is that it relies on feeds from retailers like Costco to tell Instacart which products are available to be sold. And the feed from Costco the night before had omitted, for reasons unknown, all the items from the meat department.

The Instacart systems did what they were supposed to do and made no meat items available for delivery from Costco. And of course, meat is a big part of typical grocery orders, so no meats meant a 50% decline in orders.

Now, any intelligent system checking for data quality would have spotted this obvious issue in the Costco feed. But we had a more traditional, rules-based data quality system at Instacart, where you had to define rules for your data and essentially tell the system what issues you wanted it to find. No one anticipated that one day we’d get a meat-less feed from Costco, so no one thought to write a rule to check for that condition.

It made me realize at the time that there had to be a better way to assess data quality. And that’s why we founded Anomalo. Anomalo doesn’t require rules to check your data but is able to spot a wide range of issues using machine learning and AI. And so, issues that you don’t anticipate (like no meats in the Costco feed) will still be caught by Anomalo.

Anomalo positions itself as the autonomous data system for the agentic enterprise. What do you believe truly sets Anomalo apart from traditional data quality and observability tools in the market?

Anomalo simply offers an astounding level of automation for data quality and data observability. You don’t have to write any rules or set any thresholds to spot issues with Anomalo. And if an issue arises, we will help you determine the root cause automatically (often now with the help of our conversational analytics AI agent—AIDA). We’re even launching an AI agent soon that will automatically respond to all Anomalo alerts by assessing their importance and taking the appropriate actions automatically.

The result is that we are saving our enterprise customers millions of human hours of work to ensure high-quality data for whatever data-driven use cases they have (such as business analytics or AI).

Your platform combines proprietary data profiling and prediction engine with Agentic AI. How does this unique foundation allow Anomalo to deliver insights and monitoring at massive scale without requiring manual rules or prompts?

This is the core of what makes our approach work and also why it's difficult to replicate. We have a proprietary data profiling and prediction engine that understands data at the content level. It's not just checking metadata or freshness. It's building statistical models of every column, every distribution, and every relationship in your data, and doing it continuously across billions of rows. This is a foundation that took years to build and is trained on enterprise-scale data.

Now, here's where the convergence happened. For years, machine learning could detect statistically meaningful patterns like anomalies, trends, or shifts at a scale no human team could match. But the outputs were cryptic. You got a signal, but explaining what it meant required a human analyst to investigate. Then large language models arrived. Suddenly, you could take a statistical finding, for example, a 12% drop in conversion rate concentrated in mobile users in three markets, correlating with a schema change from three days ago, and explain it in plain English with context, likely causes, and recommended next steps. That combination of statistical detection plus natural language reasoning is the engine of self-driving data. Our agents don't need a prompt to start working. The monitoring finds something, the agent investigates it, and the insight gets delivered. No one has to ask for it.

With nine specialized agents working as one autonomous system, how are you seeing data teams’ day-to-day work transform when they adopt the Anomalo platform?

Ask any data analyst or data engineer what they actually do all day, and beneath the polished job description you'll find a lot of the same things, such as monitoring pipelines for failures, writing SQL queries to answer one-off business questions, building dashboards someone requested, triaging data quality issues that shouldn't have happened in the first place, and explaining what a metric means for the fourteenth time this quarter. None of that is what they were hired to do. It's the operational tax of running a data organization.

When teams adopt Anomalo, that operational layer starts running itself. Our Table Observability and Data Quality agents handle continuous monitoring. AIDA, our conversational analytics agent, fields the analytical questions that used to consume analyst hours, because now anyone can get answers in minutes through natural language. The Data Insights Agent proactively surfaces interesting changes and trends before anyone thinks to ask, delivering analyst-grade reports automatically. I've seen it catch internal performance issues on our own platform that no one had flagged. The net effect is that data teams shift from reactive firefighting to strategic work. Our customers analyze more than ten billion rows daily and have saved millions of hours through autonomous monitoring. That's time given back to the work that actually matters.

Many enterprises today struggle with unreliable data feeding their analytics and AI initiatives. How is Anomalo helping organizations build and maintain deep trust in their data?

Fundamentally, deploying Anomalo within your organization means that someone (or, in this case, an AI system) is going to be watching your data for any changes or issues that might break the analytics or AI workloads that run on top of the data.

Without this kind of monitoring, you simply have to trust that your data is correct and hope for the best, or you have to spend hours manually verifying and investigating any dataset for issues before you use it.

With Anomalo monitoring in place, you know that if anything odd happens to your data, you will know about it. And what’s more, you know that issues will be detected quickly—before they can cause significant damage to downstream use cases.

Some of my favorite stories from the early days of Anomalo were situations where data teams got an alert from Anomalo and could resolve the issue before anyone else in the organization even noticed that there was a problem. That kind of capability builds a lot of trust in both the data and the data team managing it.

And this is precisely why industry leaders, including Atlassian, Block, Casey's, and Notion, rely on Anomalo.

Several powerful agents in your suite are marked “Coming Soon,” such as the Data Issue First Responder and Dashboarding & Reporting Agent. Which of these developments are you most excited about, and why should clients be particularly enthusiastic?

The Data Issue First Responder Agent is the one I'm most excited about because it addresses fundamental issues with alerting systems like Anomalo—the problem that sometimes you just get too many alerts.

We work very hard at Anomalo to make sure every alert we send is real and significant. But even so, the more data you use and the more monitoring you do, the more alerts you will generate. That means more time that humans need to take to process alerts and perform the right investigation, triage, and resolution steps.

Well, what if humans didn’t have to receive and process alerts? What if an AI agent could do that for you? That’s what the Data Issues First Responder does.

It receives every alert, investigates the issue, assesses impact, and follows your organization's established runbooks and policies to take appropriate action, including escalating to the right person when needed. That alone will save data teams an enormous amount of time on alert processing and will remove one of the biggest worries about monitoring data.

Ultimately, every agent we build is designed to give data teams superpowers that pull them out of the mundane, repetitive work that consumes their days and frees them to do the high-value strategic work that actually moves the business forward. The Data Issue First Responder Agent is a great example of this.

What is next for Anomalo in 2026 and beyond? How do you envision the evolution of “self-driving data” and its role in the agentic enterprise?

We introduced the category of Self-Driving Data in April 2026. The name is deliberate. Just like a self-driving car doesn't ask you to monitor the road, a self-driving data system doesn't ask your team to monitor every table. You set the destination with your business priorities, the policies, the metrics that matter, and then the system gets you there.

Self-driving cars didn't happen because one piece of technology improved. They happened when everything came together across computer vision, sensor fusion, real-time decision-making, and map data. That same convergence has now happened for enterprise data. Over the rest of 2026, we're shipping the remaining agents—the First Responder, Dashboarding, KPI Monitoring, and Experiment Evaluation—plus full MCP and API integration, so organizations can plug Anomalo's context and tools into their own custom agents and frameworks.

Looking further ahead, I believe the organizations that reach fully autonomous data operations in the next 18 to 24 months will build a compounding advantage that their competitors simply cannot replicate. They'll see things their competitors can't see and act on them before anyone else knows to look. The window to build that advantage is shorter than most people realize.

If you had to summarize Anomalo’s ultimate promise to forward-looking enterprises in one powerful statement, what would it be?

Every data team I've ever seen spends the majority of its time on work that doesn't require human judgment. We're building the system that finally changes that, where your data monitors itself, understands itself, and acts on what it finds, so your team can focus on the work that actually requires human judgment.

The Visionary Leader Upfront

Elliot Shmukler is the Co-Founder and CEO of Anomalo. Before founding the company, Elliot served as Chief Growth Officer at Instacart and held senior product leadership roles at successful Silicon Valley companies like Wealthfront and LinkedIn. He cofounded Anomalo with Jeremy Stanley, former VP of Data Science at Instacart, with a shared conviction that machine learning and AI could eliminate the manual burden of enterprise data quality.

Under Elliot's leadership, Anomalo has grown to serve some of the world's most data-intensive enterprises—including Atlassian, Block, Casey's, and Notion—has earned recognition from Gartner, and has pioneered the category of Self-Driving Data. Anomalo is backed by Databricks Ventures, Snowflake Ventures, SignalFire, Norwest Venture Partners, Foundation Capital, Two Sigma Ventures, First Round Capital, Smith Point Capital, and Village Global.

“We’re building a system where data monitors itself, understands itself, and acts on what it finds, so teams can focus on the work that actually requires human judgment.”