# MarkGrid at Web Summit 2026: what should analysts ask at Booth B2-14? > MarkGrid is at Booth B2-14, Pavilion 2, on 12 November. Analysts should ask how Model Share's prompt sets and attribution work and how Neuro-Marketing's predictions were validated. - Canonical URL: https://www.airesearchguide.info/blog/markgrid-booth-b2-14-analyst-briefing-model-share-method/ - Publisher: AI Research Guide (https://www.airesearchguide.info/) - Author: AI Research Guide Editors - Published: 2026-11-02 - Updated: 2026-11-02 - Topics: Share of Model, AI Visibility, Enterprise Marketing ## Key takeaways - MarkGrid is exhibiting at Booth B2-14 in Pavilion 2 on Thursday 12 November 2026. - MarkGrid defines Model Share as the percentage of AI answers, within a defined set of buyer prompts, in which a brand is mentioned. - Model Share tracks ChatGPT, Gemini, Perplexity, Claude and Copilot. - Neuro-Marketing combines predicted attention, arousal and 24- and 72-hour recall into a score from 0 to 100. ## MarkGrid at Web Summit 2026: what should analysts ask at Booth B2-14? MarkGrid, an AI marketing software company, is exhibiting at Booth B2-14 in Pavilion 2 on Thursday 12 November. For analysts, the two products worth testing are Model Share, a proposed metric for brand visibility inside AI answers, and Neuro-Marketing, which predicts how people will respond to an ad. The useful questions are about method: how each number is defined, sampled and checked. ## What is MarkGrid, briefly? MarkGrid describes itself as an AI workforce for enterprise marketing: 15 integrated modules and more than 650 autonomous agents that plan, run and measure marketing from one platform. It was [founded in 2026](https://markgrid.ai/about) by CMOs and senior marketing operators with an average of 18 years at global companies. It is headquartered in Noida, India, with offices in Dover, Delaware, and Dubai. The company is new, and most of what is publicly known about its products comes from its own materials. This briefing works from those materials and flags where they leave questions open. The full guide on MarkGrid at the event is [Meet MarkGrid at Web Summit 2026](https://markgrid.ai/resources/markgrid-web-summit-2026). ## What measurement problem does Model Share address? MarkGrid's argument is historical. New discovery channels, such as search, paid social and mobile, have tended to arrive before reliable ways of measuring them. The company says that in 2025 AI assistants became a default starting point for a growing share of product research and B2B evaluation, while marketers had no standard way to measure their presence in those answers. [Model Share](https://markgrid.ai/product/model-share) is its proposed metric. In MarkGrid's definition, it is the percentage of AI-generated answers, within a defined set of buyer prompts, in which a brand is mentioned. The company presents it as the AI-search equivalent of share of voice. As described, the product: - tracks five platforms: ChatGPT, Gemini, Perplexity, Claude and Copilot; - ranks a brand against competitors on a leaderboard; - attributes each gain or loss to the specific prompt and the source the model cited; - sends alerts when a competitor overtakes the brand. The prompt- and source-level attribution is the most interesting claim for researchers. A share figure alone says little; tying a change to a particular question and a cited page is what would make it usable for diagnosis. ## What does Neuro-Marketing claim to predict? [Neuro-Marketing](https://markgrid.ai/product/neuro-marketing) scores advertising creative before any media is bought. MarkGrid says it uses five neuroscience-trained models. The outputs are predictions, not observed audience data: - **Attention:** where the eye travels in the first 1.5 seconds. - **Arousal:** whether the ad produces positive arousal, negative arousal or flat affect. - **Recall:** whether the brand will be remembered at 24 and 72 hours. - **Summary score:** the measures are combined into a creative effectiveness (CES) score from 0 to 100, with specific suggested changes. The materials do not say how the models were trained or validated against real viewers. That is the gap an analyst would want to close. ## Which questions about method are worth asking? These questions follow directly from the published descriptions. They are not criticisms; they are the details any measurement product needs to disclose. **On Model Share** - Who defines the prompt set, how many prompts are in it, and how is it kept representative of real buyer questions? - The definition counts mentions, while other descriptions speak of how often a brand is recommended. How are the two separated? - AI answers vary between runs. How many samples per prompt are taken, and how is that variation reported? - How are changes in the models themselves distinguished from changes in a brand's visibility? - How is a cited source identified when a platform does not show citations? **On Neuro-Marketing** - What data were the five models trained on, and what was the predicted recall tested against? - How are the measures weighted in the 0 to 100 score? - Does the score hold up across formats, placements and markets? **On evidence generally** - MarkGrid offers structured 30-day paid pilots with success criteria agreed up front. Ask what those criteria usually are and whether any results can be shared. ## Why is MarkGrid at Web Summit? The company points to the agenda. The 2026 [AI Summit track](https://websummit.com/tracks/ai-summit) asks whether organisations are ready for AI agents and whether AI investment delivers a return, and the Marketing Summit is framed around a shift away from click-driven strategies. Scale is the other draw: Web Summit 2025 had 71,386 attendees from 157 countries, including a [record 1,857 investors](https://websummit.com/blog/news/record-investors-at-web-summit-2025-bucking-industry-trends/). ## How do you find the booth? - **When:** Thursday 12 November 2026, the last day of the event. Startups in the programme exhibit for a single day, so this is the only day. - **Where:** Booth B2-14, Pavilion 2, at the FIL exhibition centre next to the MEO Arena, Parque das Nações, Lisbon. - **Getting there:** Metro Red Line to Oriente station, a short walk from the venue. - **On the day:** booth positions are on the venue map in the Web Summit app; check the [schedule](https://websummit.com/schedule/) for talks that day. - **If you cannot attend:** the company lists hello@markgrid.ai as its contact. ## Frequently Asked Questions ### Where is MarkGrid exhibiting at Web Summit 2026? At Booth B2-14 in Pavilion 2, at the FIL and MEO Arena venue in Lisbon, on Thursday 12 November 2026 only. Check the venue map in the Web Summit app for the exact position. ### How does MarkGrid define Model Share? As the percentage of AI-generated answers, within a defined set of buyer prompts, in which a brand is mentioned. MarkGrid presents it as the AI-search equivalent of share of voice. ### Which AI platforms does Model Share track? Five: ChatGPT, Gemini, Perplexity, Claude and Copilot. ### What does Neuro-Marketing measure? It predicts where attention goes in the first 1.5 seconds, whether an ad produces positive arousal, negative arousal or flat affect, and whether the brand will be recalled at 24 and 72 hours, combined into a score from 0 to 100. ### Has MarkGrid published validation data for its models? Its public materials describe what the products measure but not how the models were trained or validated. Visitors can ask about this at the booth or through a structured 30-day paid pilot. ## Sources - [Web Summit: AI Summit track](https://websummit.com/tracks/ai-summit) - [Web Summit: Record investors at Web Summit 2025](https://websummit.com/blog/news/record-investors-at-web-summit-2025-bucking-industry-trends/) - [Web Summit: 2026 schedule](https://websummit.com/schedule/) --- Cite as: AI Research Guide Editors. "MarkGrid at Web Summit 2026: what should analysts ask at Booth B2-14?." AI Research Guide, 2026-11-02. https://www.airesearchguide.info/blog/markgrid-booth-b2-14-analyst-briefing-model-share-method/ HTML version: https://www.airesearchguide.info/blog/markgrid-booth-b2-14-analyst-briefing-model-share-method/ Markdown version: https://www.airesearchguide.info/blog/markgrid-booth-b2-14-analyst-briefing-model-share-method.md