What is LloydsFarmacia
ADMENTA Italia S.p.A. is the Italian holding of the German society McKesson Europe, European leader of intermediate and retail pharmaceutical distribution. ADMENTA Italia directly manages 168 shops in 26 Italian municipalities..
LloydsFarmacia has embraced the current technological challenge, integrating a chatbot in their online shop.
The cooperation between our two companies allowed the development of a competitive, omnichannel product through a multistep project, that involved an initial assessment, technical and conversational setup, attention to the legal aspects, go-live and evolution phases.
The chatbot for an omnichannel customer service
It is known that an accurate assessment is the added value that makes a product vertical: a successful chatbot doesn’t need to know everything in general terms, but it must be able to carry out a specific task that, in Dory’s case, is customer assistance.
This is why our domain specialists have analysed the needs of the people who purchase on the multichannel e-commerce website of LloydsFarmacia and have underlined a few critical issues the users contact the customer service about, like orders, shipping, fidelity card, payments, newsletters, pharmacies and pickup points.
The second phase in the development of the conversational interface Dory required us to implement conversational paths and connectors. Conversational paths lead users to the solution of their problem, while connectors integrate the chatbot with the platforms used by LloydsFarmacia to manage orders (Hybris), shipping (Qaplà), fidelity card and its points balance (ICteam) and tickets opened by the users (Zendesk).
The delicate medical and pharmaceutical field, moreover, required extreme attention in order to respect the consumer privacy rules. We worked from a GDPR perspective, implementing tools and methods to minimize risk, like pseudo-anonymisation, that is data encryption both for third parties and the Heres team, and the use of Stopwords, sensitive words we don’t want to be dealt with in chat.
Analytics: a report of Dory’s first few months of life
It is already possible to obtain statistics concerning the most frequent questions asked to the chatbot from the interactions during its first months of life. For example, it seems clear that a great number of users are interested in products sold by LloydsFarmacia, even if we don’t let the chatbot offer its advice on this specific field: only a pharmacist can!
The first three customer service topics by number of sessions were “Store Locator” (including locations, opening times and pharmacy services), “Loyalty” (including the subcategories CartaClub, points balance, newsletter and promotions) and “Orders” (modification, cancellation, order details and much more).
Leaving the graphs aside, let’s see a concrete example to get an idea of what kind of requests the automatic assistant of LloydsFarmacia was useful for!
Evolution: Dory’s constant learning process
Among Dory’s strengths there is the chance to exploit the users’ questions to monitor potential problems of the website and to analyse the most requested products and services. The virtual assistant gathers reports 24/7, scanning an amount of useful information that is far superior than a live chat.
Dory uses what it extracts from its acquired knowledge in evolutionary phases, to progressively improve the performance of the conversational software. Interacting with humans, the bot learns to understand more complex requests and to reply to questions it didn’t know the answer to before.
After 3 months, Dory can autonomously solve 1 ticket out of 4: a fact that is destined to improve, because Dory’s artificial intelligence isn’t based on a prior knowledge, but on the capability of learning from conversation thanks to the synergy between machine learning and human learning.