The modern data platform stack is maturing while AI shakes things up at the same time. Here is how LLMs touch each layer of the ten-layer stack — and what barely moves.

What is a modern data platform?

A modern data platform is the full path from raw data to useful insight. It pulls data in from databases, APIs, and streams. It shapes it into models you can use. It stores it in warehouses and lakehouses. It feeds dashboards, analysts, and ML models. And it governs the lot — who can see what, where it came from, and whether you can trust it.

Ten layers make up the stack:

  1. Ingestion — get data in (Kafka, Fivetran, Airbyte)
  2. Transform — shape and model it (dbt, Spark, Airflow)
  3. Storage — persist it (Snowflake, BigQuery, Iceberg, Delta)
  4. BI / visualisation — present it (Tableau, Looker, Power BI)
  5. Semantic — define metrics centrally (dbt metrics, Cube)
  6. Governance — control and document it (Unity Catalog, DataHub)
  7. ML / Analytics — model and analyse (MLflow, PyTorch, DuckDB)
  8. Security — protect it (masking, encryption, access control)
  9. Infra / Ops — run it (Terraform, Kubernetes)
  10. Observability — monitor it (Great Expectations, Monte Carlo)

Every large organisation runs some version of this. The tools vary — cloud-native, SaaS, open source — but the layers are universal.


Context: The Stack got quietly better

While everyone debated whether LLMs would replace data engineers, the modern data platform quietly built the foundation that makes AI useful:

  • Open table formats won. Iceberg became the strategic default. Portable data means AI tools aren’t locked to one vendor.
  • Catalogues became metadata APIs. Unity Catalog, Polaris, Dataplex — the context LLMs need to understand your data now lives in queryable systems.
  • Semantic layers matured. dbt metrics, Cube, Snowflake Semantic Views — governed definitions AI can trust.
  • Text-to-SQL went production. Snowflake Cortex Analyst, Databricks Genie, ThoughtSpot Sage — not demos, real features with 95% accuracy on governed datasets.
  • Warehouses became AI runtimes. Snowflake Cortex AISQL, Databricks AI Functions — call LLMs from SQL. The warehouse isn’t just storage anymore.
  • Local-first analytics emerged. DuckDB and Polars made laptop-scale prototyping viable. Teams can experiment with AI workflows before scaling.
  • BI-as-code emerged. Evidence, Rill, Lightdash — dashboards as code that LLMs can generate and version-control.
  • Lineage went portable. OpenLineage events flow across tools. AI can trace data from source to dashboard.

This foundation matters. LLMs don’t replace infrastructure — they query it. Better metadata means better AI.

The 10-Layer Stack: Before and After LLMs

How LLMs impact each layer of the modern data platform

Here is how each layer evolves:

LayerCurrent stateLLM evolutionImpact level
1. IngestionFivetran, Kafka, Airbyte, DebeziumAI agents generate connectors, handle schema driftSupplemented
2. Transformdbt, Spark, Airflow, Dynamic TablesText-to-SQL, auto-generated models, natural language pipelinesPartially replaced
3. StorageIceberg, Delta, Snowflake, BigQueryUnchanged — still the foundationNo change
4. BI / VizLooker, Tableau, Power BI, MetabaseChat-based analytics replaces dashboards for explorationDisrupted
5. Semanticdbt metrics, Cube, LookMLLLMs understand context natively; rigid definitions become optionalDisrupted
6. GovernanceUnity Catalog, DataHub, CollibraAuto-classify PII, generate documentation, explain lineageSupplemented
7. ML / AnalyticsMLflow, DuckDB, Polars, PyTorchAI-assisted feature engineering, pipeline orchestrationSupplemented
8. SecurityMasking, encryption, access controlAnomaly detection, policy generation — humans still own keysSupplemented
9. Infra / OpsTerraform, Kubernetes, monitoringLLMs help write IaC, don’t replace itNo change
10. ObservabilityGreat Expectations, Monte Carlo, SodaRoot cause analysis in natural language, auto-fix suggestionsSupplemented

Summary: 2 layers disrupted, 5 layers supplemented, 3 layers unchanged. Agents sit across those layers — there is a dedicated section after the layer-by-layer notes; they orchestrate tools and APIs and do not replace the foundation.


Layer-by-Layer: What Actually Changes

Layer 1: Ingestion — AI-Assisted, Not AI-Replaced

Current state: Choose between managed services (Fivetran, Zero-ETL) or build your own (Kafka, Debezium, Airbyte).

How LLMs change it:

  • Connector generation. LLMs draft CDC configs, Airbyte connectors, and Debezium pipelines in minutes. Choosing between paying Fivetran and generating code yourself is now a real trade-off.
  • Schema drift handling. AI agents detect schema changes and propose mapping updates.
  • Source discovery. Natural language queries like “find all customer data across our SaaS tools” become possible with proper metadata.

What stays the same: You still need Kafka for streaming. You still need private connectivity. Operational complexity doesn’t disappear — it shifts from “build the connector” to “maintain the generated code.”

The honest take: Managed ingestion isn’t dead. If you have 50 connectors and a small team, Fivetran still makes sense. But for straightforward sources with strong engineering teams, LLM-generated pipelines can shift the sums.

Layer 2: Transform — The Biggest Productivity Gain

Current state: dbt models, Spark jobs, Airflow DAGs, Dynamic Tables, declarative refresh.

How LLMs change it:

  • Text-to-SQL for ad-hoc queries. 80% of analytics queries can be generated correctly against well-documented schemas.
  • Auto-generated dbt models. Describe the transformation; get a working model with tests and documentation.
  • Natural language pipeline definitions. “Aggregate daily sales by region with a 3-day lookback for late data” compiles to microbatch config.

What stays the same: Complex business logic still needs human-authored, version-controlled SQL. dbt’s value isn’t the SQL — it’s testing, documentation, and lineage. AI-generated transforms still need governance.

The honest take: The analyst who wrote 50 dbt models now writes 10. The rest get generated on demand or become unnecessary because users self-serve via natural language. The data engineer role shifts from “write SQL” to “design systems and review AI-generated code.”

Layer 3: Storage — The Unchanging Foundation

Current state: Iceberg won the format war. Every vendor supports it. Catalogues (Unity Catalog, Polaris, Lake Formation) sit at the control plane.

How LLMs change it: They don’t.

You still need object storage. You still need Iceberg or Delta. You still need a catalogue for governance. LLMs query this layer — they don’t replace it.

What matters more now: The better your metadata, the better LLM-powered analytics works. Invest in:

  • Column descriptions
  • Certified metrics in your semantic layer
  • Lineage via OpenLineage
  • PII classification

The AI layer is only as good as the foundation it sits on.

Layer 4: BI / visualisation — the dashboard fades

Current state: Looker, Tableau, Power BI, Metabase, Superset. Self-serve exploration via drag-and-drop interfaces.

How LLMs change it:

The entire value proposition of BI tools was “explore data without writing SQL.” LLMs remove that barrier entirely.

  • Chat-based analytics. “Show me revenue by region for Q3, excluding returns, compared to last year” — answered in seconds.
  • Auto-generated charts. The LLM picks a sensible chart type for the question. Claude can render charts and diagrams inline. Snowflake Cortex and Databricks Genie do similar against your warehouse.
  • Conversational drill-down. “Why did APAC drop?” as a follow-up, not a new dashboard.
  • BI-as-code accelerates this. Evidence renders reports from SQL queries embedded in Markdown files. Rill defines metrics in YAML and auto-generates charts. Lightdash offers “dashboards as code” with full Git workflows — version control, pull requests, CI/CD for your charts.

What survives:

  • Embedded analytics in products
  • Operational dashboards with real-time refresh
  • Highly curated executive views
  • Regulated reports that need audit trails

What dies: The 47-chart dashboard that took three sprints to build and nobody looks at. Self-serve exploration layers. The BI developer role as currently defined.

The honest take: The landscape is splitting three ways:

  • Warehouse-native chat: Databricks Genie, Snowflake Cortex Analyst, ThoughtSpot Sage — text-to-SQL against your warehouse, no separate BI tool needed.
  • Incumbents adding AI: Tableau has Einstein Copilot. Power BI has Copilot. Looker integrates with Gemini. They’re bolting chat onto existing dashboards — useful, but not a full reset. Drag-and-drop stays the default.
  • BI-as-code: Evidence, Rill, Lightdash — dashboards as SQL/YAML/Markdown artifacts. Version-controlled, CI/CD-deployed, AI-writable. This is the sleeper category. When dashboards are code, LLMs can generate entire reporting layers.

The winners won’t only be the tools with the slickest chat. They’ll be the ones where AI can build, change, and maintain dashboards in code — not just answer questions on top of them.

This isn’t 5 years away. Snowflake Cortex claims 95% text-to-SQL accuracy on governed datasets. Kraken manages hundreds of Lightdash projects via CI/CD. The shift is happening now.

Layer 5: Semantic — The Sleeper Disruption

Current state: dbt Semantic Layer, Cube, LookML, MetricFlow. Weeks spent defining what “revenue” means — handling edge cases, currency conversions, fiscal calendars.

How LLMs change it:

The semantic layer exists because machines couldn’t understand context. You had to spell out every business rule explicitly. LLMs understand context natively.

  • Contextual inference. The LLM reads your schema, query history, and Slack messages about edge cases. It infers what “active user” means in your company.
  • On-demand definitions. The 80% of ad-hoc questions that require “well, it depends on how you define X” — the LLM handles those without a predefined model.

What stays:

  • Critical metrics still need governance
  • Regulated reporting needs explicit, auditable definitions
  • The semantic layer becomes simpler, not unnecessary

The honest take: Tools like Cube and dbt’s semantic layer become optional for most queries. You’ll still want them for official metrics. But the tax of maintaining elaborate semantic models drops dramatically.

Layer 6: Governance — AI-Powered, Human-Controlled

Current state: Unity Catalog, DataHub, OpenMetadata, Collibra, Atlan. Manual cataloguing, PII tagging, documentation.

How LLMs change it:

  • Auto-classification. PII detection across every table — Unity Catalog already does this for 12+ types.
  • Generated documentation. Column descriptions, table summaries, usage examples — all inferred from schema and query patterns.
  • Simple-English lineage. “Where does this revenue number come from?” answered in natural language, not cryptic DAGs.
  • Policy suggestions. “This table contains PII — recommend access controls” based on data sensitivity.

What stays the same: The catalogue still matters — arguably more than ever. The data steward role shifts from “manually tag 10,000 columns” to “review and approve AI-generated classifications.”

The honest take: Governance becomes easier, not optional. LLMs make the case for proper metadata stronger — the better your catalogue, the more useful AI becomes.

Layer 7: ML / Analytics — Copilots, Not Replacements

Current state: MLflow, feature stores, DuckDB, Polars, PyTorch, Feast, Evidently.

How LLMs change it:

  • Feature engineering. Describe the prediction problem; get candidate features from your data catalogue.
  • Pipeline orchestration. Natural language instructions that compile to Airflow DAGs.
  • Model explanation. “Why did the model reject this loan?” in simple English, not SHAP values.
  • Experiment navigation. “Show me models trained last month that beat baseline on precision” — no MLflow query syntax needed.

What stays the same: PyTorch doesn’t go away. You still need MLflow, feature stores, and proper MLOps. The PhD who hand-crafted features now supervises an agent that proposes them.

Layer 8: Security — Smarter Detection, Human Control

Current state: Column masking, row filters, dynamic masking, encryption, access control.

How LLMs change it:

  • Anomaly detection. Spot unusual access patterns, flag suspicious queries.
  • Policy generation. Natural language requirements → access policies.
  • Explanation. “Why was this request denied?” in simple English.

What stays the same: Humans own the keys. “The AI gave me access” is not an acceptable audit trail. Security decisions require human oversight.

Layer 9: Infra / Ops — Better Tooling, Same Foundation

Current state: Terraform, Kubernetes, Helm, ArgoCD, monitoring, cost management.

How LLMs change it:

LLMs help you write Terraform faster. They help you debug Kubernetes issues. They’re excellent pair programmers for infrastructure code.

What stays the same: Kubernetes doesn’t go away. CI/CD, monitoring, cost management — all still needed. The pipes and plumbing remain.

Layer 10: Observability — root cause in simple English

Current state: Great Expectations, Monte Carlo, Soda, Elementary, Lakehouse Monitoring.

How LLMs change it:

  • Root cause analysis. “Why did the pipeline fail at 3am?” gets a real answer, not a stack trace.
  • Anomaly explanation. “Data volume dropped 40% — likely due to upstream source outage at 2:47am.”
  • Auto-fix suggestions. “Schema changed — here’s the migration script.”

What stays the same: You need to detect the problem before you can explain it. Proper monitoring comes first.

Agents: where they fit

Agents are not an eleventh layer — they cut across the ten. A single-turn chat answer is one pattern; an agent is closer to a bounded loop: plan → call tools (SQL, APIs, tickets, orchestrator triggers) → observe results → stop or escalate. The warehouse, formats, and orchestration underneath stay the same; the agent is the control surface that coordinates them.

Where they show up in practice:

  • Ingestion and integration — scaffolding connectors, proposing mappings after schema drift, or driving repeatable “onboard this source” runbooks (humans still own production cutover).
  • Transforms and analytics — multi-step workflows: clarify the question, run governed SQL, check row counts, produce a chart or narrative — not only one-shot text-to-SQL.
  • Governance and catalogues — draft descriptions, suggest PII tags, queue human review (the “agentic catalog” pattern vendors are shipping).
  • Observability — triage incidents, correlate failures, suggest fixes; execution and alerting stay in your existing tools.
  • Orchestration — agents as a layer above Airflow/Dagster/Prefect: deciding when to trigger jobs or backfills from events and policies, not replacing the scheduler.

What stays non-negotiable: tool allowlists, session boundaries, audit trails, and human approval for anything sensitive. “The agent did it” is not a governance strategy.

Summary: chat lowers the floor for ad-hoc questions; agents raise the ceiling for repeatable, multi-step data work — still grounded in the same storage, semantics, and catalogues as before.

Build vs buy: what shifts

LLMs change what it costs to build things in-house versus pay a vendor:

Before LLMsAfter LLMs
Building connectors took weeksLLMs draft configs in hours
Custom pipelines needed specialistsNatural language → working code
Managed services saved timeGenerated code can be cheaper to run
Documentation was expensiveOften auto-generated from schema and usage

The question isn’t only “which tool is best”. It’s what you’re paying for — and whether that still matches your team and sources.

For small teams with many connectors, managed services often still win. For strong engineering teams with simple sources, LLM-assisted code can tip the balance.


What Doesn’t Change

SQL still wins. Spark still scales. Airflow still runs (everyone still complains). The warehouse didn’t die — it learned to talk to lakes. Good modelling beats clever tooling. Security stays human.

What This Means

The stack is consolidating around fewer, more capable platforms:

  • Warehouses absorbing everything — semantic layers, AI features, governance moving inside Snowflake/Databricks
  • Chat with data going mainstream — Cortex Analyst, Databricks Genie, ThoughtSpot Sage, Claude rendering charts inline. The BI tool becomes a conversation.
  • BI-as-code enabling AI-generated dashboards — Evidence, Rill, Lightdash turn dashboards into code artifacts that LLMs can write and maintain
  • Orchestrators becoming control planes — Dagster and Airflow 3.0 pulling in catalogue, observability, cost management
  • Catalogues going agentic — auto-documentation, auto-classification, auto-policy (Alation, Atlan, Informatica)
  • Agents on top of the stack — planners with tools (SQL, orchestrator, tickets) for multi-step workflows; governance and allowlists matter more, not less
  • Observability extending to AI — monitoring model inputs AND outputs, not just data pipelines
  • Open formats winning — Iceberg as strategic default, OSI for semantic layer interoperability

The common thread: AI reduces manual work at every layer. Data engineers shift from building pipelines to governing AI-assisted pipelines.


What To Do Now

This article covers a lot. Here is how to act on it:

1. Audit your metadata readiness. LLM-powered analytics only works if the AI can understand your data. Check: Do your tables have column descriptions? Is PII classified? Can you trace lineage from source to dashboard? If not, start there — everything else depends on this foundation.

2. Run a chat-based analytics pilot. Pick one well-documented dataset and test Databricks Genie, Snowflake Cortex, or an open-source alternative against it. You’ll learn more in a week than from reading vendor whitepapers. The goal isn’t to replace BI — it’s to understand what works and what breaks.

3. Evaluate BI-as-code for your next dashboard. Before building another Tableau dashboard, try Evidence, Rill, or Lightdash. If your team can write SQL, they can write dashboards as code. The payoff: version control, CI/CD, and dashboards that LLMs can generate and modify.

4. Revisit your ingestion costs. That Fivetran contract made sense when building connectors took weeks. With LLM-assisted development, generating an Airbyte connector or Debezium config takes hours. Run the numbers again — especially for straightforward sources where you have engineering capacity.

5. Watch the orchestrator space. Airflow 3.0 shipped data assets. Dagster is pulling in catalogue and observability features. Your orchestrator is becoming your control plane. If you’re starting fresh or planning a migration, check whether your current tool will keep up.

6. Treat governance as AI infrastructure. Every investment in catalogues, semantic layers, and lineage pays off again when AI queries your data. The better your metadata, the better AI works. Governance isn’t overhead anymore — it’s the foundation that makes everything else possible.

7. If you pilot agents, scope them tightly. Pick one workflow (e.g. “monthly metrics pack” or “incident triage”) with explicit tools, a human-in-the-loop step, and logging. Agents amplify mistakes as well as speed — metadata and policies are your guardrails.

Closing Remarks

The modern data platform isn’t being replaced. It’s getting a new interface — and that interface speaks English. Agents add a second pattern: not only “ask once,” but “run a short, tool-backed workflow” with the same foundations underneath.

The biggest shift isn’t technical — it’s who can work with data. Natural language drops the barrier to zero. Data teams shift from “answer questions” to “build foundations that let anyone answer questions” — and, where it pays off, governed agent workflows on top.

The AI layer amplifies whatever foundation you’ve built. Better metadata, cleaner catalogues, governed metrics — these amplify everything you build on top. Weak foundations show up faster.

Appendix: Watch list

Watch List: Emerging Players and Features

The stack isn’t static. Here is what is emerging across each layer — tools, features, and trends worth tracking in 2025–2026.

Storage and table formats

Apache Paimon is the streaming-first table format to watch. Built from Flink Table Store, it uses LSM-tree architecture optimised for high-velocity writes with minute-level freshness — versus Iceberg’s hour-level recommended cadence. If you’re building streaming lakehouses, Paimon fills the gap between pure streaming (Kafka) and batch-optimised formats (Iceberg).

Apache Fluss (incubating) takes it further: sub-second columnar streaming storage that tiers automatically into Iceberg or Paimon. Think of it as the “hot layer” for real-time data before it lands in your lakehouse.

DuckLake (2025) rethinks metadata entirely — storing it in SQL databases instead of JSON/Avro manifests. Early days, but watch for simplicity gains.

Iceberg v3 remains the strategic default. Every major vendor supports it now. The REST Catalog spec solved interoperability. Polaris Catalog (donated to Apache by Snowflake) is emerging as the open governance layer.

Semantic layer

The semantic layer war heated up in 2025:

  • dbt MetricFlow reached production maturity (v0.209+), now the default for multi-cloud shops
  • Snowflake Semantic Views went GA with AI-native integration to Cortex
  • Databricks Metric Views hit GA, tightly coupled with Unity Catalog
  • Open Semantic Interchange (OSI) initiative launched — dbt Labs, Snowflake, Salesforce collaborating on vendor-neutral YAML specs

The pattern: semantic layers are moving from middleware to platform-native. Expect warehouse vendors to absorb this layer entirely.

Text-to-SQL and AI-BI

The native AI analytics features shipping now:

  • Snowflake Cortex Analyst — 95% text-to-SQL accuracy on verified repositories, plus Cortex AISQL for multimodal queries (text, images, audio in SQL)
  • Databricks Genie — no-code natural language queries with Unity Catalog integration
  • Snowflake Intelligence (preview August 2025) — agentic AI for action-oriented analytics, not just answers
  • AtScale integrating with both platforms, adding semantic layer to AI queries

These aren’t pilots anymore. They’re production features competing for how users interact with data.

Data observability

Monte Carlo launched Agent Observability (2025) — unifying data and AI observability to monitor both inputs and model outputs. Their Observability Agents can triage incidents and suggest fixes (though humans still execute). Also added unstructured data monitoring for GenAI pipelines.

Emerging challengers: Sifflet, Metaplane, Sparvi offering lighter-weight alternatives. Elementary remains the dbt-native choice. Acceldata pushing enterprise hybrid environments.

The trend: observability is extending to AI outputs, not just data inputs. Expect “LLM-as-judge” monitors and output quality scoring to become standard.

Data governance and catalogues

AI is automating the grunt work:

  • Alation Agentic Platform (2025) — AI agents for auto-documentation and policy enforcement, plus AI Agent SDK supporting Anthropic MCP
  • Atlan — active metadata platform, Forrester Leader, 90%+ adoption in 90 days
  • Informatica Fall 2025 — AI Agent Hub with autonomous data management
  • Ataccama — claims 83% faster AI-ready data through autonomous curation agents

The shift: catalogues are becoming agentic. Instead of humans documenting data, agents propose descriptions, classify PII, and suggest policies. Humans review and approve.

Orchestration

Airflow 3.0 (April 2025) was the biggest update in its history:

  • DAG versioning (most-requested feature)
  • Data Assets (finally catching up to Dagster’s asset-centric model)
  • Multi-language Task SDKs
  • Event-driven scheduling

Dagster continues gaining ground with asset-aware orchestration, built-in cataloguing, and 2x developer productivity claims vs. Airflow. The Components framework (GA October 2025) makes modular pipelines easier.

Prefect remains the developer-friendly option with the simplest Python API.

Kestra emerging for YAML-first, GitOps-style orchestration.

The trend: orchestrators are absorbing catalogue, observability, and cost management features. The “orchestration platform” is becoming the “data control plane.”

AI-native infrastructure startups

Worth watching in the broader ecosystem:

  • Syncari — agentic MDM (Master Data Management) for AI-ready data, 50x faster than traditional MDM
  • Legion — enterprise AI infrastructure for data transformation and model orchestration
  • Reducto — unstructured data analytics (PDFs, images, scans)

References

Semantic layer and text-to-SQL

BI-as-code

Table formats and storage

Data observability

Data governance and catalogues

Orchestration

AI and data infrastructure