### Autonomous Data Products

### Our Product

May 7, 2025

# The 10 most-asked questions from the Nextdata OS launch (and how we answered them)

By:

Cornelius Willis

Two weeks ago, we launched the Nextdata Operating System—and the technical community showed up. At the event and in follow-ups, we fielded hundreds of questions from architects, engineers, and data leaders eager to dig deeper into how the platform works, where it fits, and how it supports real-world data product development.

So we sat down with our Head of Product, Jonathon Morgan, to go through the **10 most frequently asked technical questions**. Here’s what we heard—and how we answered.

### 1\. **How do you handle access control, security, and governance?**

We integrate with your existing identity and access management systems, not replace them. Nextdata OS supports **RBAC, ABAC**, and plugs into enterprise systems like **SSO and SCIM**. Governance can be enforced locally (per data product) and globally (across domains) using **policy-as-code**, supporting things like encryption, masking, and semantic enforcement.

_“We’re not asking you to replace what you have. We’re extending what you already trust.”_

### 2\. **How does Nextdata OS integrate with my existing data ecosystem?**

Nextdata OS is designed as **middleware**, not a monolith. It integrates via APIs with catalogs (like Collibra), data quality tools, ETL platforms, and more. It’s built to enhance—not disrupt—your current data stack.

_“We hesitate to use the word ‘migration’—this is more like an upgrade-in-place.”_

### 3\. **Where does Nextdata OS fit in my architecture, and how is it deployed?**

Think of it as a **logical layer above your storage and compute**. It runs in your environment as a Kubernetes application and orchestrates your existing infrastructure to encapsulate and manage autonomous data products.

_“We let you build on your existing investments.”_

### 4\. **How do you manage semantic alignment and resolve conflicting definitions?**

We use **generative AI** to infer and generate semantic models, but also support explicit DSL-based definitions (in YAML, Python, SQL, etc.). Semantic contracts act as an interface between data producers and consumers. Global policies can enforce consistent semantics across domains.

_“We combine local autonomy with global standardization—without central teams becoming a bottleneck.”_

### 5\. **How do you ensure data quality, lineage, and observability?**

Nextdata OS uses **data contracts** composed of “promises” (what a product guarantees downstream) and “expectations” (what it requires upstream). These are validated at runtime, enabling **loose coupling** and robust lineage without fragile dependencies.

_“If a change breaks an expectation, it’s contained. Bad data doesn’t flow downstream.”_

### 6\. **Is it usable by non-technical users?**

Yes. The platform is designed for **multi-persona collaboration**—from business users to domain experts to developers. A visual interface supports search, discovery, and trust. DSLs like YAML and Python make development approachable for non-data-engineers.

_“We’re lowering the bar so subject matter experts can create data products, not just data engineers.”_

### 7\. **Can it scale across a complex enterprise with multiple domains?**

Absolutely. That’s what it was built for. Nextdata OS is designed to support **federated architectures**, complex governance, and heterogeneous stacks—common in large enterprises and M&A-heavy organizations.

_“It’s not just compatible with complexity—it is designed to simplify it.”_

### 8\. **How does the Nexty AI Assistant work?**

Nexty is our **multi-agent AI tool** that ingests legacy pipelines, transformation logic, and documentation—then generates deployable data products and domain designs. It uses models like **Gemini and others**, cutting transformation time from **months to hours**.

_“Nexty used to be a team of human experts. Now it’s an AI assistant that does 80% of the work.”_

### 9\. **Can I extend the platform and build my own integrations?**

Yes. We support a **driver architecture** that abstracts infrastructure complexity. These drivers (for Snowflake, Redshift, S3, etc.) will be **open sourced** so customers and partners can build their own integrations.

_“The 80% use case is supported out of the box, and the long tail of heterogeneous storage and data types can be supported.”_

### 10\. **How is the product priced?**

Nextdata OS is offered as an **annual or multi-year license**, priced based on the number of data products under management. It is typically deployed into the customer’s environment for full control and integration.

_“Simple model, no per-query surprises.”_

**Want to go deeper?**

If any of these questions are top-of-mind for your team, let’s talk. Reach out to [sales@nextdata.com](/content/our-pov/10-most-asked-questions-from-the-nextdata-os-launch-and-how-we-answered-them#/index.html) and we’ll set up a technical session tailored to your needs.