CONTEXT LAYER

The context that gives your numbers meaning.

Your AI agent answers “What was the ROAS in March?” correctly. When asked “Why did the ROAS drop?,” it speculates about rising click prices and your competition. Your senior analyst, however, knows that the consent banner was updated in March, resulting in a 20% loss of attribution data. The Semantic Layer defines what your metrics mean. The Context Layer provides the “why”: events, decisions, definition history, and rules that an agent must know in order to act reliably.

We design and build your Context Layer, which turns accurate numbers into sound decisions.

Request initial consultation

21→95%

Response accuracy of the same AI model with and without structured business context (Anthropic).

>20%

Accuracy through plain-language ontology (a machine-readable vocabulary for key metrics), with 39% fewer tool calls (Snowflake experiment).

42%

of companies discontinue the majority of their AI initiatives before they go live, up from 17% the previous year (S&P Global Market Intelligence).

20+

years of our experience as boutique consultancy in marketing analytics and data minimize your risk.

Tobias Lanzl
Manager AI-ready Data Stack & AI Expert

+49 89 219 099 021

Book your free 30-minute initial consultation today.

Contact

Two needs. The same foundation.

Who benefits from the Context Layer?

The Context Layer addresses different challenges depending on whether your focus is on reliable data or better business results.

CTO | Head of Data | IT

Transparent AI Responses Instead of a Black Box

Agents that only know the Semantic Layer provide the correct number but the wrong explanation. The Context Layer makes every response auditable: Which definition was used, which events were taken into account, which source was decisive, and which access rules applied. AI responses become verifiable rather than merely plausible.

Auditability Governance Provenance AI Reliability

Use Cases for IT

CMO | Marketing | BI

Immediate answers to “why” questions. No time-consuming research required.

“Why did the ROAS drop in March?” Until now, answering this question took hours: searching through Slack, asking colleagues, and hoping someone remembered the tracking change. With Context Layer, the agent knows the campaign history, promotions, changes, and seasonality—and provides the explanation right away.

Why Analytics Quick Insights Knowledge Stays In-House

Use Cases for Marketing

Context Layer Event & Webinar

Your AI agent answers every question with confidence, but how do you know if that answer is backed by evidence or just a guess? At our hybrid Raw & Roasted event on September 17, 2026, Tobias Lanzl will show concrete examples of how the Semantic Layer and Context Layer work together so AI agents understand your data correctly and answer reliably. Afterward, there’s plenty of time for your questions and a real exchange of ideas.

Save your Seat

CONTEXT IN THE OVERALL SYSTEM

One of five layers: exploring the “why” behind the numbers

The Context Layer is the fourth of five layers in our Hopmann AI Foundation: the foundation for reliable AI decisions in marketing. It builds on a stable Semantic Layer and provides exactly the knowledge a metric needs at any given moment: events, decisions, and rules. The following overview shows where it fits into this system.

Marketing Measurement Strategy

Strategy and structure for governance and comparison

Data Quality & Governance

The foundation you can’t skip

Semantic Layer

One consistent definition for everyone

Context Layer

Business context for the right conclusions

You are here

Agentic Marketing Intelligence

Insights for better decisions

Details on the Hopmann AI Foundation

Use Cases for IT

How the Context Layer Helps in IT & BI

EXPLAINABILITY

“Why” questions become answerable rather than speculative

TODAY
The data warehouse stores numbers and statuses, but often not the events behind them. If someone asks the AI agent why a metric has changed, the agent can only offer a plausible guess. A migration, an outage, or a promotion simply don’t show up in the data.

WITH CONTEXT LAYER
Relevant business events such as migrations, promotions, or system changes are recorded in a structured manner within the data stack. When asked “why” questions, the agent first checks which documented events took place during the relevant time period. Only then does it interpret the trend.

DEFINITION HISTORY

Changes to your metrics remain traceable

TODAY
The churn definition was changed in May. The code shows what changed, but not necessarily why. As a result, comparisons across this point in time can be misleading, and the AI agent may not recognize this break in logic.

WITH CONTEXT LAYER
A versioned changelog documents changes to metrics and their definitions. When making temporal comparisons, the agent automatically detects such breaks, flags them, and can better contextualize the results based on both definitions.

TRUST & OPERATIONS

Answers Become Traceable Instead of a Black Box

TODAY
The agent provides a number, but it often remains unclear how that number was arrived at. It is difficult to determine which definition was used, which source was referenced, and how current the data is. As a result, errors are often not noticed until much later.

WITH CONTEXT LAYER
Every answer is linked to its context: the definition and version used, the data source, how current the data is, and relevant rules. Verified queries can also be used as tests to identify quality issues early on.

Marketing Use Cases

How the Semantic Layer Helps in Marketing

Time Savings

Understand ROAS drops without days of research

TODAY
ROAS is dropping. Then the search for the cause begins: Slack threads, meetings, follow-ups, and guesswork. Eventually, someone remembers that the consent banner was updated during the same period.

WITH CONTEXT LAYER
“Why did ROAS drop in March?” The agent finds the documented consent banner update and realizes that it’s not necessarily that the campaigns have gotten worse, but that the way they’re measured has changed.

KNOWLEDGE RETENTION

Campaign knowledge stays within the company

TODAY
Why some brand campaigns are excluded from the attribution model is known primarily by the person who made that decision. When that person leaves the company, part of the context often disappears as well.

WITH CONTEXT LAYER
Important decisions are documented along with the rationale and rejected alternatives. This allows teams and AI to understand at any time why a particular decision was made. This simplifies onboarding and prevents knowledge from remaining tied to specific individuals.

THE RIGHT CONCLUSIONS

Recommendations Are Made in the Business Context

TODAY
The AI agent recommends shifting more budget to the channel with the best CPA. What the agent doesn’t know: The Christmas campaign for this channel is already fully planned, and additional budget will add little value.

WITH CONTEXT LAYER
Goals, campaign logic, seasonality, and market conditions are factored into the recommendation. The AI agent evaluates not only metrics and KPIs, but also the business context in which they arise.

Brief insight: The distinction

How the Context Layer effectively complements the Semantic Layer

The Semantic Layer is your company’s dictionary: It provides a definitive definition of terms such as “revenue,” “churn,” and “conversion.” This is the foundation. To ensure reliable answers, the Context Layer comes into play: It ensures that an agent draws the correct conclusion from the right number.

The Semantic Layer consistently describes the past in a way that’s the same for everyone: What does this metric mean?

The Context Layer goes further. It supplements events, decisions, definition history, playbooks, permissions, and origin—tailored to a specific task at a given moment—and answers the question: What does an agent need to know to act correctly here and now?

Important to know: You can’t buy a Context Layer, only host it. The content, that is, your team’s business knowledge, only exists once it’s written down. We have a process model for this.

Deep Dive: Semantic Layer vs. Context Layer

Roadmap & Milestones for Implementation

Our 5-Phase Approach for Your Context Layer

The Context Layer is built on a stable Semantic Layer. We build it step by step and integrate knowledge capture into your existing workflows so that it remains up-to-date rather than becoming outdated.

Checking the foundation

Objective: Validate the semantic layer, data quality, and the most important business questions.

This ensures that the context layer is built on a reliable data foundation.

Capture Events & History

Goal: Build a business events and metrics changelog and maintain a record of known events from the past twelve months.

This reveals what has changed and when, as well as which events have influenced key metrics.

Document decision-making knowledge

Goal: Create decision records, caveats, and playbooks for key metrics and recurring questions.

This preserves relevant knowledge and ensures it remains traceable for both humans and AI at all times.

Integrate and Verify Agents

Goal: Integrate the Context Layer via MCP and build verified queries for key business questions.

This generates AI responses that are traceable and can be verified during ongoing operations.

Embed context in the workflow

Goal: Integrate PR templates, incident runbooks, and feedback loops into existing processes.

This ensures that new context is captured where it arises and remains up-to-date at all times.

The Result

A context layer that people and agents work together to keep up to date. Answers your company can trust because they are backed by evidence.

HOPMANN: STRATEGY, DATA, AND AI WITH A HOLISTIC APPROACH

Your partner for reliable context in Analytics- and AI systems

An AI pilot only creates real value when it works reliably in everyday use and is adopted by teams. We lay the groundwork so that your agents can provide answers that are transparent and enable sound decisions.

Many companies have data, dashboards, and a solid analytics infrastructure. What’s often missing is a shared business logic that both people and AI can understand and use. That’s exactly where we come in.

Context in Your Team’s Workflow

The knowledge needed for your Context Layer already exists within your company. We provide the methodology and processes to capture this knowledge in a structured way, make it usable, and keep it up to date on an ongoing basis.

The difference: A Context Layer that grows with your company, rather than documentation that will be outdated by tomorrow.

20 Years of Experience

For two decades, we’ve been supporting corporations and mid-sized companies with challenging data and transformation projects. Personal, experienced, and on equal footing. TISAX-audited*.

The difference: Proven methods and international best practices instead of unnecessary learning curves.

Semantic and Context Layers from a Single Source

The Context Layer builds on the Semantic Layer. With us, both stem from the same approach, featuring a consistent architecture from the very beginning.

The difference: A consistent architecture instead of two separate projects.

A Holistic View

For us, the Context Layer never stands alone. We always consider the interplay of data, processes, technology, and usage, and implement individual building blocks where they make the most sense.

The difference: We always keep the architecture, the target vision, and future scalability in mind.

Our Offer for your Context Layer

We offer a variety of entry-level formats and project models tailored to your current situation and goals. The best way to determine the most appropriate approach is to discuss it together during a brief initial consultation.

From €4,900 for 2 days with 2–3 Hopmann experts

Strategic Planning Workshop

A collaborative in-house workshop with your team to develop your data architecture strategy across all five layers: from Marketing Measurement Strategy through Data Quality & Governance and the Semantic Layer to the Context Layer and Agentic Marketing Intelligence.

Outcome: a prioritized implementation plan

Initial consultation

Upon request

Concept, Implementation & Support

Complete design, implementation, and ongoing development of your semantic layer using our workflow model: from business model design to production-ready AI integration.

Customized proposal based on your objectives, data availability, and complexity.

Initial consultation

Does your AI provide reliable answers, or just guesses? Book a free initial consultation now!

Tobias Lanzl
Manager AI-ready Data Stack & AI Expert

Checkbox *

I have read the privacy policy and agree to the storage and processing of my personal data by Hopmann Marketing Analytics GmbH. Your contact information is required to contact you regarding your request. You can unsubscribe from our notifications at any time.

Frequent questions on the AI Foundation

FAQ on the Context Layer

What is a context layer?

A Context Layer is the infrastructure that provides AI agents with the business knowledge that isn’t stored in any table: documented business events, the history and rationale behind key metric definitions, operational playbooks, access rules, and the source and recency of each response. It ensures that agents not only perform calculations correctly but also draw the right conclusions in the right business context.

Can you buy a Context Layer?

No. Tools can host and deliver the Context Layer, but the content consists of documented business knowledge that is created only when your team writes it down. Setting it up is therefore primarily a workflow issue: knowledge capture becomes an integral part of existing processes, not a one-time documentation effort.

Why isn’t a better AI model enough?

Because context has a greater impact than model size. Anthropic has shown that the same model, when provided with structured business context, answers 95% of analytics questions correctly, compared to just 21% without it. Documented knowledge beats a larger model, and it takes an afternoon instead of a procurement cycle.

How does the Context Layer differ from the Semantic Layer?

The Semantic Layer is the dictionary: It provides a definitive definition of what each metric means. The Context Layer is the rationale: why the metric was defined that way, what events explain its fluctuations, when it no longer applies, and what the business actually means by a particular question. The two go hand in hand; the Semantic Layer is a component of the Context Layer, but it does not replace it.

What components make up a Context Layer?

Typical components include a joinable business events table in the warehouse, a versioned metrics changelog, short decision records in the repository, operational playbooks and guardrails, identity resolution across system boundaries, provenance and freshness signals, as well as verified queries serving as grounding and a test suite. The entire system is connected via a small number of clearly defined MCP servers (Model Context Protocol server).

How does Hopmann support the design and implementation of a Context Layer?

We guide you through the entire process: validating the data foundation and the Semantic Layer, building the event history and definition changelog, facilitating the documentation of decision-making knowledge with the business departments, integrating your AI agents, and embedding knowledge capture into your workflows to ensure the layer remains up to date.