The products behind insight, creative testing and activation
One maintained map of the public portfolio, rather than a thin page for every .AI name.
The House-of-AI layers
Foundation
Global Data Platform, Plus.AI and the Data Ecosystem provide shared data and a conversational planning layer.
Insight.AI
Research.AI, Search.AI, Persona.AI, Touchpoint.AI and Behave.AI support research, visibility, audiences, journeys and behaviour.
Creative.AI
Pretest.AI evaluates video and static creative against a large benchmark set before campaign launch.
Activate.AI
Track.AI, NE.R.O. AI, Predict.AI, Total Video Integrator and Ecosystem.AI support data flow, targeting, modelling, planning and orchestration.
What the named products do
Mediaplus' Data & AI portfolio groups four core workflows: strategy, audience, media modelling and measurement. Research.AI generates synthetic profiles and conducts agent-led interviews; its product page says 100 profiles are the default and claims more than 90% accuracy compared with real samples. The same page says human and hybrid panels are future additions, so the current public offer should be understood as synthetic-panel research.
Search.AI builds statistically plausible digital twins from GWI behavioural data and simulates journeys across ChatGPT, Claude, Gemini and Perplexity. Its public page names outputs including mention rate, tonality, competitive positioning and content recommendations. These are repeated model observations, not a stable universal rank.
Pretest.AI evaluates video and static assets against more than 250,000 benchmark evaluations and 15 metrics. Mediaplus says results take 15–30 minutes, cover more than 20 markets and cost €1,500 for one asset. The page also says an unnamed specialist provider operates the underlying tool, with Mediaplus providing integration and interpretation.
Mediaplus-reported portfolio figures
How to evaluate a Mediaplus AI product
Match the evidence question to the product rather than treating the suite as one score.
Define the decision
Is the job market research, audience selection, creative validation, media planning, targeting or measurement?
Inspect the data
Ask whether inputs are observed people, synthetic profiles, first-party data, licensed panels, content corpora or model outputs.
Validate the measure
Request the comparison set, sample, error range and method behind accuracy, prediction or ROI claims.
Plan operational use
Document who interprets the output, approves action and checks drift after markets, models or creative change.
Public evidence gaps
The portfolio page describes capabilities more fully than validation methods.
- Sample and method behind the portfolio-level 18% ROI figure
- External validation for Research.AI's accuracy claim
- Named operator and model detail for Pretest.AI
- How Search.AI controls prompt, model, location and time variance
- Product-level retention, subprocessors, security and audit evidence
Mediaplus AI product questions
What is Plus.AI?+
Mediaplus presents Plus.AI as a conversational operating system for data-based media planning connected to its Global Data Platform and product modules.
Does Research.AI interview real people?+
The current public product describes synthetic profiles and agent-led interviews. Human and hybrid panels are described as future additions.
What does Search.AI measure?+
Mediaplus names mention rate, tonality, competitive positioning and content recommendations across major generative assistants using simulated audience journeys.
Are the Mediaplus performance figures independently audited?+
No independent methodology is published on the reviewed portfolio pages. The figures should remain attributed to Mediaplus.
SokosumiWhat the House of Communication learned about AI, as a product you can use on Monday
Sokosumi is the marketplace of named AI coworkers built by Serviceplan Group with NMKR. You brief a coworker like a colleague; the task shows on a shared board and comes back as a file: a PDF, a deck, a spreadsheet or a dashboard.