The document discusses the use of artificial intelligence in retail. It describes how retailers can use AI technologies like computer vision, recommender systems, and speech/text processing along with large amounts of customer data from online and physical interactions to improve the customer experience through personalized recommendations and context-aware interactions. The goal is to create a unified user profile across channels to optimize experiences using emerging interfaces like augmented reality, conversational assistants, and social advertising. Considerations are given to experimenting with AI-as-a-service models and using AI for disruptive innovation beyond just optimizing current processes.

1. Intelligenza Artificiale
nel Retail
a Fabio Gasparetti
AI Lab @ Dipartimento di Ingegneria
Università degli Studi Roma Tre

2. 22
Large amount of
Human-generated content
Posizione GPS, “mi piace”, precedenti
acquisiti, immagini sui social.
Infrastructures
Cloud computing, GPU-enabled
infrastructure
EU GDPR
personal information as
economic asset
AI-enabled frameworks
Recommender systems, Speech and Text
processing, Video and Image analysis
Oligopoly on data
Poche imprese possiedono
grandi quantità di dati sull’utente.
AI-based interpretation of content
by natural language processing, personality
traits, object recognition, etc.
AI & Big Data
La distribuzione tra intelligenza artificiale, e-commerce e abitudini di consumo - 12 aprile 2018Fabio Gasparetti

3. 3
Geotracking
• Self-scan terminals, RFID-tags, Bluetooth/Wifi
through smartphones, CCTV cameras [Oosterlinck],
IEEE 802.11mc
Use cases
.
AI-based
technologies
• Clustering of user behaviors,
• Computer vision applications
(e.g., people tracking, emotion recognition, action
recognition)
Hi-level
Goals
• Location Analytics
• Customization of offline promotions [Zhang,Arce-Urriza]
• Aggregated spatio-temporal analysis [Larson]
• Planogramming support [Geismar]
Attualmente gli ambienti virtuali, come i
marketplace online e i social networks, sono
impiegati per catturare dati sul
comportamento e le preferenze del cliente.
Gli ambienti fisici sono una
fonte altrettanto importante.
Context awareness: User behavior tracking
La distribuzione tra intelligenza artificiale, e-commerce e abitudini di consumo - 12 aprile 2018Fabio Gasparetti

4. 4
Desktop
e-Commerce
Collaborative Filtering,
Recommender Systems,
Pull paradigm
Mobile
e-Commerce
.
1st generation Chat bots
Voice User Interface, Single domain
rule-based bots [Van Hove]
Social Networks’
buy button
Social Filtering
Comparison
Shopping Engine
Information Extraction
Aspect-based mining
Start
Domanda in
input
La domanda è
“C’è la taglia X”
La domanda
è “Dov’è il mio
ordine?”
Interroga il db con
taglia X
NO
SI
NO
Stop
Non ho capito la
domanda
…
SI
2nd generation Chat bots
Context-aware, multiple-domain
self learn-able bots, conversational UI
Push paradigm
Augmented reality
& Visual Search
Computer Vision
Assistant-enabled commerce
Amazon
Dash button
User Interfaces: Multimodal & assisted interaction
La distribuzione tra intelligenza artificiale, e-commerce e abitudini di consumo - 12 aprile 2018Fabio Gasparetti

5. 5
AI & Retail Considerazioni
1
Unified backbone & Personalization
User-centric optimized and context-aware experience.
Estendere lo User profiling considerando dati strutturati e non strutturati da
più canali online e offline.
AI-enabled UI experience
Contemplare nuovi paradigmi di interazione (non mutuamente esclusivi).
Social advertising
I social networks non solo come fonte di profilazione ma anche per
influenzare l’acquirente.
Sperimentare AIaaS (AI as a service)
Valutare nuovi paradigmi e sistemi di raccomandazione sfruttando servizi
online AI-enabled [Janakiram].
Disruptive and Radical AI-based innovation
Pensare alla I.A. non solo per ottimizzare i processi attuali ma per creare
nuovi scenari, strategie, mercati.
2
3
4
5
La distribuzione tra intelligenza artificiale, e-commerce e abitudini di consumo - 12 aprile 2018Fabio Gasparetti

6. 6
.
• Sorensen, H. (2003) The Science of Shopping. Marketing Research 15(3), 30-35pp.
• Geismar, H. N., Dawande, M. , Murthi, B. P. and Sriskandarajah, C. (2015), Maximizing Revenue Through Two-Dimensional
Shelf-Space Allocation. Prod Oper Manag, 24: 1148-1163pp.
• Van Hove S. et al. (2018). How to (not) nudge customers? Methodological insights from a situated eye-tracking study on the
intrusiveness of a location-based shopping assistant in a supermarket. Etmaal van de Communicatiewetenschap.
• Oosterlinck D. et al. (2017) Bluetooth tracking of humans in an indoor environment: An application to shopping mall visits.
Applied Geography Vol 78, 2017, 55-65pp.
• Nazari A.K., Singh A. (2018) E-Commerce-Consumer Perceptive Model for Online Purchases. Progress in Advanced
Computing and Intelligent Engineering 409-415pp.
• Zhang J, Wedel M (2009) The effectiveness of customized promotions in online and offline stores. J Mark Res 46(2):
190–206pp.
• Arce-Urriza M. et al. (2016) The effect of price promotions on consumer shopping behavior across online and offline channels:
differences between frequent and non-frequent shoppers. Information Systems and e-Business Management, 2017, 15(1)
69–87pp.
• Larson J.S. et al. (2005) An exploratory look at supermarket shopping. International Journal of Research in Marketing 22(4)
395-414pp.
• Zeng F. et al. (2016) Omnichannel Couponing. Harvard Business Review. 2016, 22–23pp.
• Quadrana, M.; Cremonesi, P.; Jannach, D. (2018) Sequence-Aware Recommender Systems. arXiv:1802.08452
• Janakiram MSV (2018). The Rise Of Artificial Intelligence As A Service In The Public Cloud. Forbes.
• Khan M.M. et al. (2017) Cross Domain Recommender Systems: A Systematic Literature Review. ACM Comput. Surv. 50(3).
• Adomavicius and Tuzhilin (2015). Context-Aware Recommender Systems. Recommender Systems Handbook Authors.
• Guy I. (2015) Social Recommender Systems. Recommender Systems Handbook Authors 2015.
• Altexsoft (2017). Comparing Machine Learning as a Service: Amazon, Microsoft Azure, Google Cloud AI.
https://www.altexsoft.com/blog/datascience/comparing-machine-learning-as-a-service-amazon-microsoft-azure-google-cloud-
ai/
• Love, K. C. (2015) Social Media and the Evolution of Social Advertising Through Facebook, Twitter and Instagram.
La distribuzione tra intelligenza artificiale, e-commerce e abitudini di consumo - 12 aprile 2018Fabio Gasparetti
Riferimenti