#	Pagina
attuale pagina	/open-h2020/projects/216761/results.html

Report

Teaser, summary, work performed and final results

Periodic Reporting for period 1 - Grocerybot (Grocerybot â€“ AI-supported text dialog agent for communication with grocery stores in European languages)

Teaser

Summary

Loglab.pl is developing, using NLP techniques, a text dialog agent in Polish (chatbot) for communication with grocery stores. The company is finalizing a commercial version of the product which it will offer to Polish retail chains. Such agents will make e-grocery processes (e.g. the cumbersome process of filling the e-cart) and obtaining grocery product information much easier for customers who are increasingly diet/component/bio-food conscious. Creating high usability by consumers such text agents will become an important channel for marketing, customer service and sales with good perception by users and high opening and click-through rates for targeted messages sent through this channel. Loglabâ€™s project is a feasibility analysis including research of the market for chatbots, especially in the retail and grocery domain in several European languages (English, French, German and Spanish) and developement of a business plan for potential scaling of the product to EU markets. The objective is to find if such scaling is feasible, what would be the optimum path and resources needed.

Work performed

Loglab performed the study by means of desk reseach, expert and prospective client interviews and analysis based on experience from the on-going development of the Polish product. We found the European market for AI powered chatbots to be in very early stages in terms of commercial implementations â€“ but with very high potential. The size of the market for chatbots in the 4 languages in Europe is difficult to estimate, however, using data pertaining to related markets we believe that it was in 2016 of the order of Ä˜ur 100M and is growing with CAGR of 20-40%. The analysis indicated strong future potential for AI/NLP powered chatbots in the retail domain but has also shown that currently interest of retailers in AI chatbots is limited. Retailers are in this respect behind the tech or banking sectors. AI projects require a learning phase to train appropriate NLU neural networks and to develop the dialog managers and NLG modules. Commercial units of European retailers are concentrating on tools which are ready to implement and ensure high level of UX 100% of the time â€“ conditions which are still very dfficult to achieve in AI powered chatbots. We found that a suitable go-to market strategy for Loglab is to approach EU retailers (and CPG companies) through their accelerator programs using as reference our Polish language Grocerybot.
The analysis concentrated on text chatbots, available in the web services of companies and particularly on chatbots daveloped for the Facebook Messenger platform. Loglab sees also future potential for AI based text agents in the grocery domain on the largest voice platforms, Amazon Alexa and Google Assistant.
As the result of the project we decided to continue the project of scaling the Grocerybot to EU languages and markets.

Final results

Loglab will continue efforts to build relations with EU retailers, chatbot/voice platforms and CPG companies and to scale our product. Interactions with consumers during development of the chatbot are promising and we see potential for developing a tool widely appreciated and used by consumers shopping for groceries both online and off-line.