Most AI systems look impressive in controlled environments.

But the real test begins when they meet production data.

In my latest conversation, I sat down with Antonio Bustamante, Co-Founder and CEO of bem on The Ravit Show, to unpack one of the hardest problems in enterprise AI today: making unstructured data reliable enough to power real decisions.

We talked about why extraction alone is not the problem anymore. The harder challenge is what comes after. Validation. Business rules. Schema enforcement. Exception handling. And making sure the system does not quietly fail when confidence drops.

Antonio shared how bem approaches this differently by treating unstructured data pipelines more like critical infrastructure than experimental AI workflows.

Instead of relying on prompts and probabilistic outputs, their focus is on deterministic systems that enforce constraints, match against existing data, and flag uncertainty rather than guessing.

We also explored why unstructured pipelines tend to fail silently, how teams should rethink data quality in a probabilistic world, and where the real intellectual property lies when building AI-powered products today.

What I found most interesting is how this shifts the conversation from model capability to operational trust. Enterprises are not just asking whether AI works.

They are asking whether they can depend on it when money, compliance, or customers are involved.

If you are building AI products, data platforms, or automation workflows, this discussion goes deeper than the usual AI hype cycle.

Watch the full interview below and share your biggest takeaway.

#data #ai #bem #theravitshow

They built one of the fastest growing newsletter platforms for creators.

I just published my conversation with Tyler Denk, Co-Founder & CEO of beehiiv on The Ravit Show and this one is fully practical.

Instead of talking about newsletters in theory, we went inside his own newsletter, Big Desk Energy, and broke down how he actually runs it.

On screen.

Step by step.

No fluff.

We covered:

* How Big Desk Energy crossed 118,000 subscribers

* What is automated vs what still needs direct input

* Website builder from landing page to archive

* What happens the second someone signs up

* How signup data powers smarter automations

* Welcome journeys and re-engagement campaigns explained simply

* Mistakes new creators make with automations

* The post editor and what actually improves open rates

* Growth tools like Boosts, Marketplace, referrals, and recommendations

* Monetization through digital products, paid subscriptions, and ads

* Why beehiiv charges 0% platform fees on paid subscriptions

Personally, this conversation meant a lot to me.

I use beehiiv myself to run our newsletter with 137,000 subscribers. It powers everything behind the scenes for us. Publishing, segmentation, automations, monetization.

This episode felt like opening the backend of a real newsletter business and understanding how to think about it as a system, not just emails.

If you are serious about building a newsletter in 2026, this is not inspiration content. It is execution content.

We also have a special offer for The Ravit Show community:

30% off for your first three months

Code: RAVIT30

This link auto-applies the discount:

https://beehiiv.link/7twzg8

I built 137K subscribers on beehiiv.

Now you can see exactly how the CEO himself runs it.

#data #ai #newsletter #beehiiv #creators #community #automations #theravitshow

Let's go 2026!!!! This Enterprise series is huge. Loved recording this one with real leaders who are using GenAI and leveraging right Data. Everyone is talking about GenAI in customer service. Very few are talking about what actually works in production. Next week, we are going live with a real, honest conversation on how enterprises are using GenAI with customer data to transform customer service.

This is not about theory.

This is not about shiny demos.

We will talk about:

- What GenAI chatbots and virtual agents are actually doing in production

- Why most pilots fail when they hit real enterprise data

- The kinds of customer questions that create the most value

- How teams think about security, privacy, real-time access, and scale

The business impact leaders really care about like resolution time, agent efficiency, and customer satisfaction

Joining me are Ronen Schwartz and Miguel Navarro, who are working hands-on with these challenges inside large financial institutions.

If you are building GenAI for customer-facing use cases, or planning to, this session will help you avoid costly mistakes and focus on what actually matters.

Customer service is being redefined.

But only for teams that get data right.

#data #ai #genai #agents #governance #dataquality #k2view #theravitshow

Once teams start trusting AI to explain systems and guide decisions, expectations change. The conversation shifts from understanding to action. In Part 2 of the podcast, we go deeper into what happens next in the AI journey on the mainframe.We talk about trust and why it is the real blocker to scaling AI.How organizations move from alerts and recommendations to agentic workflows.And what safe autonomy actually looks like in real enterprise environments.Liat Sokolov from BMC Software shares how trust in AI develops over time and why it cannot be rushed.Anthony DiStauro from BMC Software breaks down how AI agents and orchestration change day-to-day mainframe operations without removing human accountability.This episode is about delegation, scale, and using AI in a way that strengthens reliability rather than putting it at risk.#data #ai #mainframe #agents #bmi #mainframe #skills #bmc #theravitshow

🚨Breaking: Big announcements by MongoDB. I just spoke to Pete Johnson, Field CTO for AI at MongoDB on The Ravit Show. This conversation cuts through the noise around AI and focuses on what actually works in production!!!!

We get into:

- What the Field CTO for AI role really looks like in the field

- How MongoDB thinks about the data and AI intersection

- Which recent MongoDB announcements actually matter

- What enterprises should prioritize as they head into 2026

- Clear advice for developers, data, and DevOps teams building with AI

No hype. No buzzwords. Just practical insight from someone working directly with customers every day.

If you are building, scaling, or planning AI beyond pilots, this episode is for you.

Thank you MongoDB for sponsoring this post!

#data #ai #mongodb #databases #mongodbpartner #agents #agenticai #theravitshow

AI agents did not suddenly appear this year. What changed is that enterprises are finally ready to use them. At AWS re:Invent, I spoke with Ronen Schwartz, CEO of K2view, to talk about what has really shifted in the agent space from last year to this year and why this moment feels different!!!!

Here is what we covered

• What changed in the agent landscape

Last year was about ideas and experiments. This year is about execution. Ronen shared why agents are moving from demos into real enterprise workflows

• K2View on the AWS Marketplace

We talked about what it means for K2View to be available on the marketplace and how customers can make the most of it, from faster onboarding to simpler adoption

• The AWS and K2View partnership

We discussed how the partnership with AWS helps customers deploy agent driven use cases faster while keeping control over their data

• Agentic AI and the re:Invent keynotes

We reflected on the keynote announcements and why agentic AI has become a central theme for AWS and its ecosystem

If you are trying to separate agent hype from what is actually working in production, this conversation brings a grounded perspective!

#data #ai #awsreinvent #agents #agenticai #aws #enterprise #k2view #theravitshow

Cloudera and AWS Partnership is extraordinary!!!! At AWS re: Invent, I stopped by the Cloudera booth to sit down with Michelle H., who leads Global Alliances and Channels worldwide for Cloudera. She sits right at the intersection of Cloudera, AWS, and the customers who are trying to move faster with data and AI.

A few highlights from our conversation

• Cloudera x AWS as a strategic bet

This is not a surface-level partnership. There is a Strategic Collaboration Agreement in place with AWS that goes deep into joint roadmaps and hands-on POC workshops, all focused on delivering more value for customers.

• AI, but with control and flexibility

Customers want AI, but they also want to de-risk it. Cloudera is helping with offerings like AI Studios, while still giving customers an open base and control over their data in their own AWS environment.

• Hybrid as a real differentiator

Cloudera is becoming a true “bridge to the cloud” for many teams, helping them move workloads between cloud, on-prem, and even the edge instead of locking them into a single pattern.

• Real impact, not just slides

We talked about work with Mercy Corps, where data and AI are helping save lives in crisis zones like Sudan and Gaza, and about a major pharma company running 100-plus GenAI use cases on Cloudera AI to bring drugs to market faster.

• Innovation and sovereignty

Cloudera is also leaning into Sovereign Cloud, including being a launch partner for the EU Sovereign Cloud in AMIA and working in regions like the Kingdom of Saudi Arabia.

Michelle’s advice to enterprise leaders was simple and practical: lean on your partners as trusted advisors, and do not wait for the perfect plan. Get started today with one use case and learn from it.

#data #ai #awsreinvent #aws #agents #awspartners #agenticai #theravitshow

At AWS re: Invent this month, I spoke to one and only Ed Huang, Co-Founder and CTO of TiDB, powered by PingCAP, to talk about something many teams are quietly struggling with.

The last few years in data were all about unbundling.

- One database for transactions.

- One for analytics.

- One for vectors.

That worked when workloads were separate.

But AI changed the rules.

As Ed put it, today the same application, or even a single AI agent, jumps across transactions, analytics, and vector search in milliseconds. No developer wants to manage five databases. Agents cannot do that at all.

That is why PingCAP is betting on unification with TiDB X.

One system that can:

- Scale like an analytical warehouse

- Behave like a transactional database

- Support vectors and search natively

- Run on object storage with elastic compute

Keep a single SQL interface with strong guarantees

This does not mean specialized databases disappear. Many of them will continue to win in narrow use cases.

But the center of gravity is clearly shifting.

As AI agents become first-class users of data systems, platforms that remove boundaries matter more than platforms that optimize for one workload in isolation.

This was a sharp, honest conversation on where databases are heading in the AI era.

#data #ai #awsreinvent #qlik #agenticai #sovereigncloud #aws #theravitshow

Big companies do not switch databases unless something is broken. At AWS re: Invent I spoke to Max Liu, Co-Founder and CEO of TiDB, powered by PingCAP, and Bo Liu, Head of Online Infrastructure at Pinterest, to talk about why teams like Pinterest are moving core systems back to SQL with TiDB.

This was a deep, honest conversation about scale, reliability, and where databases are headed next.

We covered:

- Why Pinterest finally hit a wall with NoSQL after years on HBase

- What really breaks first at massive scale: consistency, operations, or developer velocity

- How Pinterest migrated critical graph systems to TiDB with no downtime and five nines consistency

- The tension CIOs face between rock-solid stability and fast-moving AI teams

- Whether embeddings and AI workloads should live inside the same database as transactions

- How agentic workloads are changing what enterprises expect from a database

This was not about hype. It was about real systems, real migrations, and real trade-offs.

#data #ai #awsreinvent #agents #agenticai #aws #database#enterprise AWS Partners #pingcap #theravitshow

Most cloud conversations ignore one simple truth. AI only moves as fast as your connectivity. At AWS re:Invent, I spoke to Ed Baichtal, Principal Engineer at Equinix, to talk about what actually sits underneath modern cloud and AI architectures!!!!

A few takeaways from the conversation

- Equinix operates more than 270 data centers worldwide and acts as the physical home for servers, networks, and cloud on ramps

- Equinix is the largest AWS on-ramp provider, with native AWS connectivity built directly into its data centers

- The Equinix Fabric Cloud Router is now available on the AWS Marketplace, making multi-cloud connectivity much simpler

- Customers can virtually connect up to 25 Gbps to AWS across different regions without deploying physical hardware

- AI workloads are pushing serious demand for higher throughput and more reliable connectivity between clouds

Over 3,000 customers are already connected through the Equinix Fabric, which means new connections can happen virtually

One customer moved massive volumes of data to AWS in under 30 days using the Fabric Cloud Router

Looking ahead, Equinix plans to introduce Fabric Intelligence in Q1 2026 to make multi-cloud and NeoCloud connectivity even smarter

If you are building AI or multi-cloud systems, this layer matters more than most people realize.

#data #ai #awsreinvent #aws #agents #awspartners #agenticai #theravitshow