- 01 Insight.AI Serviceplan · Mediaplus Audience, behaviour and journey data become research and direction.
- 02 Creative.AI Serviceplan Generative systems and production workflows develop and adapt content.
- 03 Activate.AI Mediaplus Media is planned, activated and optimised continuously.
- 04 Agentic.AI Plan.Net Specialised agents take bounded work and feed results back into the data layer.
All four layers read from and write to the Global Data Platform.
The physical House of Communication in Munich. Serviceplan describes the House of AI as its digital twin.
The public architecture
Global Data Platform
The common data foundation is designed to make current, market-level information usable across insight, creative and media workflows.
Insight.AI
Research, audience understanding, personas and behavioural intelligence turn data into direction for human teams.
Creative.AI
Generative and production systems help teams develop, test, personalise and adapt creative work without removing human craft.
Activate.AI
Media planning, activation and optimisation use the shared intelligence layer to improve execution continuously.
Agentic.AI
Specialised agents perform bounded tasks, connect tools and hand work between systems or people.
Human expertise
Serviceplan calls the principle augmented intelligence: technology expands human capability rather than replacing judgment.
A digital twin, not a tool catalogue
Serviceplan calls the House of AI the digital twin of the House of Communication. That description matters: the model mirrors how the group already combines creative, media, data and technology disciplines. The goal is to connect those disciplines through one operating model rather than to place an AI feature in each agency independently.
The official page describes three connected suites — Insight.AI, Creative.AI and Activate.AI — above a fully compliant Global Data Platform. Agentic.AI sits across the system as a layer of specialised agents. Mediaplus' Plus.AI adds a conversational intelligence layer for media work. See Serviceplan Group's architecture.
What is public and what is not
The names and high-level structure are public. Serviceplan also documents individual products and units elsewhere, including Plus.AI, Behave.AI, Agentic Services, HealthContent.AI and Sokosumi. Detailed model choices, client data architecture, governance controls and performance by client are not published as one technical specification. A useful evaluation should therefore ask for the exact data sources, approval points, hosting, measurement plan and human responsibility for the proposed use case.
House of AI questions
Is House of AI a software product?+
Serviceplan presents it as an integrated operating system and portfolio, not as one off-the-shelf application. Parts of it are agency services, internal workflows, partner technologies or products such as Sokosumi.
What is the difference between Insight.AI and Plus.AI?+
Insight.AI is one suite in the House of AI model. Plus.AI is Mediaplus' media-planning operating system and conversational intelligence layer, with seven named modules including Research.AI, Persona.AI and Behave.AI.
What does Agentic.AI mean here?+
It is the layer for specialised AI agents that perform defined tasks and can be orchestrated into client-specific systems. Plan.Net's Agentic Services and Sokosumi are two public expressions of that layer.
Does Serviceplan publish an AI ethics code?+
No standalone public AI ethics code was found in the reviewed sources. The group does make compliance, GDPR, EU AI Act and human-augmentation claims for parts of the portfolio; those should be verified for each engagement.
Primary sources cited on this page
Reviewed 26 Aug 2026. Claims remain attributed to the organisations that published them.