From Intent to Trusted Data: Agentic Data Engineering is Now Generally Available

[CliveBearman]

Qlik’s Agentic Data Engineering is now generally available in Qlik Talend Cloud, bringing AI assistance to the entire data engineering workflow and not just for code generation. There are five new capabilities, from declarative data pipelines where coding agents can use YAML to build data ingestion, to conversational catalog and glossary, data quality, and data product agents, that help data teams move faster from intent to trusted, AI-ready data. Best of all, these new features do so without trading away the governance, lineage, and quality controls that make data worth staking a decision on, keeping your team’s expertise firmly in the loop.

[CliveBearman]

If you build data pipelines for integration or analytics for a living, then you already know that the hard part is not writing code. It is everything around deploying and maintaining that code: interpreting field meanings, testing optimum data flows, inspecting lineage, ensuring governance, and developing quality rules that keep dashboards honest. Data engineers expend a lot of effort when it's done properly. That's why I'm happy to announce the release of new data engineering capabilities to help you. Qlik’s Agentic Data Engineering features in Qlik Talend Cloud are Generally Available (GA) and focused on making many of those supporting tasks easier and more responsive to the ever increasing demands of the business.

We first shared the idea at Qlik Connect in April, but an agentic vision is just a slide deck until features actually ship. Our new capabilities bring AI assistance to the entire data engineering workflow, not just code generation, and as a result, your data team can move faster from intent to trusted data products.  All while leveraging your data domain knowledge and keeping your expertise firmly in the loop.

Note: Plenty of tools can autocomplete a transformation for you. The interesting question is no longer whether an agent can write a pipeline, but whether the data coming out at the other end is one you would stake a decision on. That is the bar we set, and it is the part most demos quietly skip over.

Why now?

The single biggest obstacle to AI value is rarely the models themselves. It is the gap between ambition and data readiness. Agents and analytics teams need timely data, consistent meaning, excellent quality, lineage, and policy controls, all at once. Miss any one of those, and the impressive insights your confidently predicted turn into a supremely wrong answer in production.

Agentic Data Engineering closes that gap by improving data engineering developer efficiency through AI assistance and automation, and it does so without asking you to trade away governance to get there. As AI shifts from assisting to acting, the data engineer’s role is shifting too from pipeline builder toward architect of intent, trust, and outcomes.

What’s in the release?

There are five headline capabilities. Each one is useful on its own, but they are designed to compound as a productivity force multiplier when used together.

1. Qlik’s Agentic Experience – Qlik’s natural language AI assistant is now available to data users.
2. Catalog and Business Glossary Agent - Automated, always-current documentation and discovery
3. Data Quality Agent - Natural-language data observability and quality rule creation.
4. Data Product Agent - Conversational lifecycle management for data products.
5. Declarative Pipelines - YAML-based pipelines that your coding agent of choice can read, generate, and update.

🔗 See the release notes [here]

Declarative Data Pipelines

Data pipeline building for everyone, to deliver trusted data that matters.

This is the one most engineers will reach for first. Declarative Pipelines let you define Qlik Talend Cloud data pipelines as YAML-based code that third-party AI coding agents, such as Anthropic Claude Code or GitHub Copilot, can read, generate, and update on your behalf. You prompt your coding agent of choice from the IDE of your choice. The agent pulls pipeline context from your GitHub repository, generates or modifies the pipeline, allowing you to commit the changes, and refresh your project.

The best thing is that Qlik now gives you the freedom to build pipelines visually, manually, or agentically with any third-party coding agent and any code editor. Agents fetch only the pipeline context they need when they need it, following best-practice, schema-enforced design integrated with GitHub. Competitors may offer similar natural-language pipeline generation, but many lack the integrated data quality and governance of the wider Qlik platform, which is exactly what turns generated pipelines into trusted data products for AI.

Note: Typically a new UI debuts when vendors rollout new features. However, we're meeting data engineers in the IDE and the repos they already use, rather than asking them to pivot to new tools. This is the right call, and the YAML-plus-LLM-plus-GitHub foundation means this fits into the CI/CD process you already trust.

Why it matters: The market has moved since the debut of ChatGPT in late 2024, and we are meeting the data engineer where you naturally want to work, in the agentic development environment of your choosing. We are not forcing you into our web UI but rather offering our data expertise and heritage through new agentic tools. Organizations using agentic pipeline development anecdotally report meaningful gains in developer productivity and flexibility, and by redirecting engineering effort from pipeline maintenance towards more strategic work adds extra value.

Watch the demo below to see a quick overview of declarative data pipelines in action:

Catalog and Business Glossary Agent

Manages the lifecycle of catalog and glossary entities so data assets are accurately defined, classified, and discoverable by both humans and AI systems.

I think the Catalog and Business Glossary Agent is going to be one of the most used agents in a data engineer's arsenal. You can ask it to generate a comprehensive glossary for a business domain, or derive one directly from existing data products, datasets, or Qlik application. It auto-generates contextually accurate term definitions, organizes them into category hierarchies, and links them back to the catalog objects they describe. You can search and explore the whole Qlik catalog through the same conversation, and the agent handles the full lifecycle from creation through to deprecation.

The difference from the usual glossary tool is where the definitions come from. They are derived from business domain context, data product structure, field metadata, and app measures and dimensions, rather than hand authored. The terms are automatically linked to the assets they describe and therefore your "documentation" always stays in sync with the data itself instead of rotting from the moment someone renames a column.

Note: Virtually every analytics or data team has a glossary that was lovingly written once and never touched again. Documentation that regenerates from the data, and links itself back to that data, is the only kind that survives contact with a real backlog.

Why it matters: The catalog and glossary agent removes the manual burden of documentation, classification, and maintenance, and it provides other AI agents the correct, governed metadata context they need to behave. The business payoff is faster time-to-insight through always-current terminology.

Watch the demo below to see the Catalog and Glossary Agent in action:

Data Quality Agent

Helps anyone explore the data quality of an asset, creates and monitor data quality rules in natural language, so decisions rest on data they can trust.

Describe what good data looks like, and the agent does the rest. It finds the target dataset, checks existing fields and rules to avoid redundancy, generates the rule expressions, and lets you refine them iteratively before anything is written. Rules are only created and bound to fields once you approve them. For a fuller pass, the guided assessment walks an entire dataset field by field, surfacing AI-generated suggestions for you to accept, refine, or skip, then summarizes what it found.

You can also ask for quality status at any time and get trust scores, valid and invalid counts, applied rules, and computation freshness for any dataset. Rules come from natural-language descriptions rather than hand-written expression syntax, the guided workflow makes sure no field is quietly missed, and status, including staleness indicators and disabled rules, is always available through conversation without hunting through separate UI views.

Note: The quiet win here is Qlik Trust Score™. This is a metric that evaluates the reliability and quality of datasets or data products, then provides a single, intuitive score for data trust and AI readiness.

Why it matters: The Data Quality Agent removes the technical barrier to quality governance. Anyone who can describe good data can now create and apply rules, which widen participation well beyond specialist roles and gives you faster, more consistent coverage and always-current visibility into quality. Watch the demo below to see the Data Quality Agent in action:

Data Product Agent

Helps teams create and govern trusted, AI-ready data products that are easy to maintain and consume.

This agent manages the full data product lifecycle through conversation: creation, activation, deactivation, and deletion, each with confirmation steps, pre-activation validation, and dependency awareness. For creation it checks products with similar names, collects required fields, suggests optional metadata, and presents a full summary before committing anything. For activation it runs pre-activation checks on required fields, dataset inclusion, and owner assignment, then resolves gaps through dialogue. For deactivation it spells out downstream impact before asking for explicit confirmation, and for deletion it surfaces every known dependency and requires unambiguous sign-off.

Crucially, no operation executes without your explicit confirmation. Missing information is gathered through conversation rather than thrown back as a wall of validation errors, duplicates are caught before creation, downstream impact is always communicated before destructive actions, and failures are explained in plain language with suggested next steps.

Note: “Nothing happens without explicit confirmation” should be table stakes for any agent that can delete things. It is genuinely reassuring to see destructive actions gated behind dependency awareness rather than an optimistic shrug.

Why it matters: Data Product lifecycle management means coordinating metadata, ownership, datasets, and consumers. The agent orchestrates that complexity behind a conversational interface, so you get faster creation and activation, a much lower risk of misconfiguration, duplication, or accidental deletion, and governance enforced at every step without extra process overhead.

Watch the demo below to see Data Product Agent in action:

How these capabilities fit together

Used individually, each agent saves time. Used together, they form a loop. The Catalog and Glossary Agent gives every asset consistent meaning. The Data Quality Agent ensures that meaning is backed by data you can trust, with a freshness signal attached. The Data Product Agent packages it all into governed products with their dependencies understood. Declarative Pipelines feed the whole thing from the IDE and code repository your team already maintains. The result is a path from intent to trusted, AI-ready data product that keeps human judgment in the loop at every decision point.

Note: Intent replaces instructions, and intelligence replaces toil. The teams that lean into this will not only ship data faster, but they will also change how their organizations think, decide, and act. The future of data engineering has arrived, and it is agentic!

Availability and getting started

All Agentic Data Engineering capabilities are GA today and available to current Qlik Cloud users. Some of the Agents require an additional cost depending on subscription tier, so check with your Qlik administrator. If you want to go deeper, then I’d start with the release notes, then navigate to the new feature documentation. Alternatively contact your account rep or CSM for a personalized demonstration.

🔗Release notes: [here]

🔗Watch the launch webinar: [here]

🔗Book a guided demo with a Qlik Solutions Engineer: [here]