September 29 – 30, 2025 | Geffen Hall, Columbia Business School, 645 W. 130th Street| In-Person & Livestream
9am-5pm on 9/29; 9am-12:30pm on 9/30
Call for Content is Open!
We are currently accepting academic submissions from MSI academic members for this event. Find out more here.
AI is transforming how organizations generate insights and forecast outcomes, but questions remain about reliability and the need for human oversight. On September 14-15 in NYC, industry and academic leaders gathered to discuss integrating AI into analytics workflows and maintaining transparency. CHECK THIS LINK FOR EVENT CONTENT.
AI is reshaping how marketing organizations generate insights, support decisions, and estimate uncertain outcomes. Yet many organizations are still determining when AI outputs are reliable enough to inform consequential decisions, and what forms of human oversight are necessary to preserve accountability.
This conference will examine how firms are incorporating AI into analytics and forecasting workflows, including how they evaluate its outputs, define appropriate human oversight, and maintain transparency and credibility.
The focus is not on AI hype or automation for its own sake. Instead, we will look at how organizations are using AI to improve core analytics and forecasting activities, including measurement, modeling, experimentation, and scenario analysis.
We will also consider a broader view of forecasting, such as traditional sales projections, product adoption, market response, innovation success, and strategic resource allocation.
Join us September 14–15 at Cornell SC Johnson College of Business on Roosevelt Island in New York City. Hosted by MSI, the ARF, and Cornell SC Johnson School of Business, this event brings together leaders from academia and industry to share new research, exchange ideas, and spark discussion.
Call for Content is now closed. Selected sessions will be determined this summer.
1) Validation and Performance Assessment
A central theme of the conference will be methods for validating AI-enabled analytics, including:
We will examine how leading organizations and thought leaders are building evidence to support (or challenge) AI-driven decisions, grounded in empirical research and real implementations.
2) Human + AI: Designing Productive Collaboration
As AI tools become embedded in modeling, forecasting, and decision support, firms must clarify the role of human expertise. Sessions will explore:
The goal is to move beyond automation toward thoughtful integration.
3) Transparency, Data Context, and Trust
As AI reshapes how data is generated, accessed, and interpreted, new transparency challenges emerge:
This discussion will prioritize practical experience over theory, focusing on how firms maintain credibility and accountability.
4) Evidence from the Field: Marketing Use Cases
Throughout the program, sessions will be grounded in ongoing empirical work and concrete use cases across marketing domains and verticals. We aim to highlight:
By grounding the conversation in evidence, the conference will help participants calibrate expectations and identify areas requiring further experimentation.
5) Emerging Research in AI and Marketing Analytics
Ongoing research continues to expand understanding of AI-enabled marketing analytics, decision systems, and data-driven marketing more broadly.
This work will help connect emerging research insights with evolving challenges in AI-enabled analytics and forecasting.
On September 29-30, the ARF and MSI co-produced the inaugural ANALYTICS & FORECASTING conference, exploring the evolving role of modeling in market research and forecasting, with a particular focus on the opportunities and limitations of synthetic data. Attendees engaged in critical discussions about the opportunities and limitations of modeling in market research applications and heard practical strategies and solutions from leading researchers and practitioners.
Imran Ahmed
Head of Measurement, Tech & Lifestyle
Google
Khaled Boughanmi
Assistant Professor of Marketing,
Cornell University, Johnson Graduate School of Management
Raymond R. Burke
Professor of Marketing, E.W. Kelley Chair of Business Administration,
Kelley School of Business Indiana University
Leabe Commisso
Senior Vice President of Strategic Growth,
Ipsos
Peter Danaher
Professor
Monash University
Brian Dennehy
Expert Partner,
Bain & Company
Harini Devulapalli
Senior Manager
Chewy
Ankit Dhawan
Founder & CEO,
BluePill AI
Jeffrey P. Dotson
Associate Professor, The Ohio State University, Fisher College of Business
Luiz G. Duarte, Ph.D.
Sr. Research Consultant,
Wortya
Mike Finnerty
US President,
Mutinex
Qiwei Han
Assistant Professor,
Nova School of Business and Economics, Portugal
Manish Hira
Data Analysis Manager, Capital One; Ph.D. Candidate,
Virginia Commonwealth University
Jianjun (JJ) Huang
Director, Global Data Science,
Kenvue
Raul Jurado
Principal and Planning Intelligence Lead,
Cloverpop
Lana Klein
Practice Lead,
Cloverpop
Alice Li
Associate Professor,
The Ohio State University,
Fisher School of Business
Jia Li
Associate Professor,
Wake Forest University
Scott McDonald, Ph.D.
CEO & President
ARF
Wendy Moe
Dean’s Professor of Marketing,
University of Maryland’s Robert H. Smith School of Business
Hayden Ostendorf
Practice Senior Manager,
Bain & Company
Adithya Pattabhiramaiah
Associate Professor of Marketing, Faculty Director of the Center for AI in Business,
Georgia Tech University
Audrey Poinsot
Research Lead,
Ekimetrics
Stefano Puntoni
Sebastian S. Kresge Professor of Marketing,
The Wharton School
Manoj Thomas
Sabanci Professor of Marketing & Management and Associate Dean of NYC Initiatives,
Cornell University
Rogier Verhulst
CEO
Kwantumlabs.ai
Marco Vriens
Founder
Kwantumlabs.ai
Shane Wang
Professor of Marketing, Virginia Tech University Pamplin College of Business
Shane Wang
Professor of Marketing, Virginia Tech University Pamplin College of Business
Isamar Troncoso
Assistant Professor of Business Administration, Harvard Business School
Shane Wang
Professor of Marketing at Pamplin College of Business,
Virginia Tech
Xufeng Wang, Ph.D.
Partner and Chief Data Scientist
Ipsos U.S. Data Labs
Helen Wolf, Ph.D.
Senior Director, Global Consumer Experience Insights
Colgate-Palmolive Company
Helen Wolf, Ph.D.
Senior Director, Global Consumer Experience Insights
Colgate-Palmolive Company
EVP Global Media, CPG and Retail, Circana
Head of Marketing Insights, LG Ad Solutions
E.W. Kelley Professor of Business Administration Indiana University, Kelley School of Business
Analytics Manager, Arc Worldwide
News Editor, Ad Age
eCommerce/omnichannel Marketing Director, PepsiCo
Research & Insights Manager, North America & Global Functions, TikTok
Professor of Marketing, Kellogg School of Management, Northwestern University
Managing Director, Commerce at GroupM's Mindshare
Founder and CEO,
Data POEM
Head of Insights & Analytics, Walgreens Advertising Group
Global Head of Virtual Experiences, Ipsos
Research Director, MSI
Chief Knowledge Officer, MSI
Chief Research Officer, ARF
| 8:00-9:00am | Registration & Breakfast |
| 9:00 – 9:15am | Opening Remarks Keith Smith – Managing Director, MSI |
| Emerging Research in AI and Marketing Analytics | |
| 9:15 – 9:45am | From Prompt to Product: Using AI to Guide Aesthetic Design Decisions This session presents a research-based framework for structuring AI-assisted product design. While image generation tools enable rapid exploration of design alternatives, they offer no guidance on which aesthetic directions are appropriate for a given product category or which will resonate with consumers. The framework addresses this gap by identifying the design dimensions that can be manipulated within a product category, distinguishing them from dimensions locked in by category conventions, and isolating those that independently drive consumer choice. Results from studies across different product categories demonstrate how the approach informs design exploration, testing, and differentiation strategy. Jeffrey P. Dotson – Associate Professor, The Ohio State University, Fisher College of Business |
| 9:45 – 10:15am | The AI Value Gap: What Marketing Leaders Do Differently |
| 10:15 – 10:45am | Discovering Textual Drivers of Marketing Outcomes from LLM Internals: A Sparse Autoencoder and Multi-Agent Pipeline Firms sit on enormous volumes of customer text: reviews, support tickets, chats, social media posts. There is so much text that they need machine learning to make sense of it. But those techniques can only answer the handful of hypotheses the analyst thinks to test. This session shows how the internals of large language models can be used as a reusable measurement tool to test thousands of hypotheses analysts never considered. Shane Wang – Professor of Marketing, Pamplin College of Business, Virginia Tech University |
| 10:45 – 11:15am | Morning Break |
| 11:15 – 11:45am | Panel Discussion: AI and the Future of Marketing Research What happens when AI becomes not just a tool for analyzing marketing research but a participant in it? As organizations experiment with synthetic respondents and AI-generated consumers, longstanding assumptions about research design and validation are being challenged. This panel brings together industry leaders and academic researchers to explore where AI is delivering meaningful value today, where expectations may be outpacing reality, and how marketing research is likely to evolve in the coming years. Raymond Burke – Professor of Marketing, Kelley School of Business, Indiana University Leabe Commisso – Senior Vice President of Strategic Growth, Ipsos Ankit Dhawan – Founder & CEO, BluePill AI Stefano Puntoni – Professor of Marketing, The Wharton School Moderator: Scott McDonald, Ph.D. – President & CEO, ARF |
| 11:45am – 12:15pm | Collaborative Intelligence: Reconstructing the Invisible Consumer From Fragmented Data Firms routinely use large-scale consumer surveys to support segmentation, targeting, and product strategy, but these surveys are often modular by design and fragmented in use, making it difficult to represent and interpret some consumers. The framework supports accurate prediction, outperforming off-the-shelf GPT and linear multi-task baselines, while recovering interpretable heterogeneity that fragmentation can leave unseen. Alice Li – Associate Professor, Fisher School of Business, The Ohio State University |
| 12:15 – 12:45pm | Synthetic Conjoint Methods for B2B Business Decisions In B2B software, where buying cycles run 9 to 12 months, and most prospects are locked into multi-year contracts with incumbents, forecasting adoption is one of the hardest jobs in product strategy. Learn about the results of a human conjoint with five synthetic conjoint results. Two yielded decision-grade results. This was achieved this by leveraging data from existing sales calls and implementing specific decision instructions. The improved method showed only a small (1.5%) difference in the share of the none-option (used for adoption decisions). Rogier Verhulst – CEO Kwantumlabs.ai Marco Vriens – Founder Kwantumlabs.ai |
| 12:45 – 1:45pm | Lunch |
| Validating AI for Marketing Analytics | |
| 1:45 – 2:15pm | When AI Incentives Help and Hurt |
| 2:15 – 2:45pm | Validating Enterprise AI Retrieval Systems for Structured Analytics Data Can AI accurately answer questions across millions of rows of enterprise data? This session explores the challenges of applying RAG to structured datasets and compares several retrieval strategies through real-world evaluations. Learn which approaches perform best, where they fail, and how to validate AI before deploying it to production. Harini Devulapalli – Senior Manager, Chewy |
| 2:45 – 3:15pm | From Reviews to Actionable Insights: An LLM-Based Approach for Attribute and Feature Extraction Firms have access to vast amounts of customer review data but often lack a scalable way to turn it into actionable decisions. This research introduces an LLM-based framework that distinguishes broad perceptual attributes from specific, managerially actionable product and service features. Applied to 20,000 Starbucks Yelp reviews, the approach yields results that closely align with human coding and strongly predict customer ratings, while processing reviews in seconds rather than minutes. The resulting structured data support dashboards that track sentiment across stores and over time, identify customer “joy points” and “pain points,” and prioritize high-impact improvements. Khaled Boughanmi – Assistant Professor of Marketing, Cornell University, Johnson Graduate School of Management |
| 3:15 – 3 :45pm | Afternoon Break |
| 3:45 – 4:15pm | Why Your AI Persona is Lying to You: The Danger of Data Overload in Synthetic Populations Amid growing reliance on Artificial Intelligence for market research and decision-making, the authors explore the limits of the technology by reviewing the conclusions of their multi-stage study comparing actual respondents to synthetic panels. While the literature thus far has focused on the ability to “sound human,” the study reveals that actually mimicking consumer behavior requires more than simply adding more data to the model. Luiz G. Duarte, Ph.D. – Sr. Research Consultant, Wortya |
| 4:15 – 4:45pm | Extracting Consumer Insight from Text: A Large Language Model Approach to Emotion and Evaluation Measurement This study introduces the Linguistic eXtractor (LX), a fine-tuned large language model trained on consumer-authored text that has also been labeled with consumers’ self-reported ratings of 16 consumption-related emotions and four evaluation constructs: trust, commitment, recommendation, and sentiment. LX consistently outperforms leading models, including GPT-4 Turbo, RoBERTa, and DeepSeek, achieving 81% macro-F1 accuracy on open-ended survey responses and over 95% accuracy on third-party–annotated Amazon and Yelp reviews. Peter Danaher – Professor of Marketing and Econometrics, Monash University |
| 4:45 – 5:15pm | Closing the Innovation Forecast Gap: An AI/ML Approach to New Product Forecasting and Incrementality at Kenvue Forecasting new product demand remains a persistent challenge because of historical sales data are often unavailable, and traditional methods frequently rely on surveys, analogs, or subjective judgment. This presentation introduces an AI/ML-based forecasting framework that unifies demand prediction and incrementality measurement within a decision-intelligence platform. The solution integrates syndicated POS data, historical launches, execution variables, and AI-enriched product attributes to forecast Year 1 and Year 2 performance while estimating cannibalization and net-new growth. Beyond improved accuracy, the approach enables scalable concept evaluation, scenario simulation, and execution guardrails, allowing organizations to make faster, more confident innovation decisions while reducing dependence on costly concept-testing studies. JJ Huang – Director, Global Data Science, Kenvue Raul Jurado – Principal and Planning Intelligence Lead, Cloverpop Lana Klein – Practice Lead, Cloverpop |
| 5:15 – 5:30pm | Closing Remarks Scott McDonald, Ph.D. – President & CEO, ARF |
| 5:30 – 6:45pm | Cocktail Reception |
| 8:00-9:00am | Registration & Breakfast |
| 9:00-9:15am | Opening Remarks Scott McDonald, Ph.D. – President & CEO, ARF Manoj Thomas – Sabanci Professor of Marketing & Management, Associate Dean of NYC Initiatives, SC Johnson, Cornell SC Johnson College of Business |
| Human + AI: Designing Productive Collaboration | |
| 9:15 – 9:45am | GEO in the Loop: A Framework for Integrating Generative Engine Optimization Into Marketing Analytics Workflows Generative AI tools such as ChatGPT, Perplexity, Claude, Gemini, and Google AI Overviews are now a de facto customer acquisition channel. In Q1 2026, AI-driven channels generated 876,000 credit card acquisitions in the U.S. alone, nearly doubling AI’s share of card issuance in a single quarter. Yet most marketing analytics teams lack a validated framework to measure, audit, or govern their brand’s presence in these outputs. This session introduces a practitioner-oriented GEO measurement framework built on four new KPIs – Answer Share, Citation Inclusion Rate, Factual Accuracy Score, and Brand Framing Sentiment, paired with a structured human-in-the-loop (HITL) audit workflow and governance model. Drawing on early field deployments in financial services, we share design principles, organizational friction patterns, and preliminary audit findings, including three error types (competitive misattribution, outdated product information, and brand framing inconsistencies) that automated scoring cannot detect. Manish Hira – Data Analysis Manager, Capital One; Ph.D. Scholar, Virginia Commonwealth University |
| 9:45 – 10:15am | Who Chops the Onions? The Hidden Ingredients in Every Profitable AI Agent AI agents are moving into business processes much faster than firms are learning how to evaluate them. This session traces where agentification stands today and walks through its six stages, from deciding whether agents belong in a workflow at all to measuring what they actually returned. At each stage, anecdotes from the field are paired with the best available academic and industry evidence to show where AI initiatives most often succeed, stall, or leak value. The central question is: where in your own process does the next dollar of return come from? Dr. Adithya (Adi) Pattabhiramaiah – The Sharon A. and David B. Pearce Associate Professor of Marketing and Faculty Director, Center for AI in Business at the Scheller College of Business, Georgia Institute of Technology |
| 10:15 – 10:45am | Operationalizing Causal Machine Learning to Optimize Online Ads Parameterization Using Brand Lift Studies Data In this session, we present work conducted with a multinational food and beverage corporation to optimize online ad parameterization using Brand Lift Study data. Our primary questions were: what are the causal drivers of brand lift on a major online video platform? And how do these effects vary across campaign levers and cost metrics? Our goal was to build a scalable, auditable analytical system. This system produces calibrated causal effect estimates for key campaign dimensions (format, duration, device, frequency, and audience) suitable for operational decision-making within a global marketing organization. We will first introduce Causality ETE, an end-to-end causal AI framework designed for applied marketing use cases. We will then report results from its large-scale deployment for the use case described above. This work makes two contributions: a structured methodology for integrating human expertise and causal AI in marketing analytics, and a reusable, auditable artifact implementing this methodology in a production environment. Audrey Poinsot – Research Lead, Ekimetrics |
| 10:45am – 11:10am | Morning Break |
| 11:10 – 11:40am | When Search Stops Sending Traffic: The Impact of Google AI Overviews on Organic Traffic Monetization in E-Commerce Google’s AI Overviews are changing not only how much traffic search engines send to e-commerce websites but also the commercial composition of that traffic. Using a large-scale panel of e-commerce domains worldwide, we examine the staggered international rollout of AI Overviews and its effects on organic sessions, transactions, revenue, and monetization quality. We find that AI Overviews reduce Google organic traffic, while completed transactions and total revenue remain largely unchanged. At the same time, conversion rates and revenue per session increase, whereas average order value does not change materially. These results suggest that AI Overviews act as a filtering mechanism: they absorb lower-intent information-seeking visits within the search interface while preserving more commercially qualified traffic. Search Console evidence further links this pattern to changes in search-result visibility, particularly for informational queries. Qiwei Han – Assistant Professor, Nova School of Business and Economics, Portugal |
| 11:40 am – 12:10pm | Foundational Relational MMM: Zero-Shot Generalization and Row-Level Attribution Learn about a relational deep-learning architecture for marketing-mix modeling that operates directly on raw heterogeneous database tables, with the aim of learning marketing dynamics from the data rather than imposing them through priors and bespoke feature engineering. The contribution is methodological: a schema-agnostic forecaster that supports row-level attribution by construction and is compatible with foundational pre-training across heterogeneous datasets. Get a report on intermediate evidence on a benchmark against a Bayesian MMM, on zero-shot generalization to entities held out at training, and on a quantitative feature-importance decomposition of the trained function. Mike Finnerty – U.S. President, Mutinex |
| 12:10 – 12:15pm | Closing Remarks Scott McDonald, Ph.D. – President & CEO, ARF |
September 14–15 at Cornell SC Johnson College of Business on Roosevelt Island, in New York City.
Hotels near the Cornell SC Johnson College of Business.
This is a member only event and included as part of the membership benefits. Explore what membership includes here.
Analytics & Forecasting 2025: Past Event Highlights
Privacy, Trust & AI: How U.S. Consumers Are Rewriting the Rules of Advertising
To further, through research, the scientific
practice of advertising and marketing
© the Advertising Research Foundation. All rights reserved. – Privacy Policy
