Travel companies consider personalization as a major lever in their quest to improve its revenue and the customers travel experience. However, in the time of pandemic and afterwards, travel will not be the same. A customer needs to feel safe at all times wherever they plan to travel. They would like to be notified in advance if the destination is a healthy to travel or not. At the same time, a customer like to be pampered with the best offers and recommendations so that he or she is encouraged to travel more often. A sustainable travel ecosystem is possible when the local businesses can also join hands with the travel companies to provide more offers and discounts. To regain customer trust will be a key goal in the personalization journey and alma is designed to solve just that!

Target Customer & Market

Any Travel company and in destination travel businesses.

What it does

alma (Soul or spirit in Spanish) is an AI powered personalization solution built specifically for travel. It can help the travel companies and their partners in considerably improving the ancillary revenue and at the same time improve customers travel experience. alma has two components. • Local alma is a mobile plug in with AI capability which can be attached to any travel company’s mobile app. A mobile app knows you better than anyone else which include your purpose of travel, personal preferences, health data, social data and much more. Hence, Local alma mines this knowledge and creates a Specialized Travel Experience Quotient (sTEQ) from it. • Space alma is a AI based distributed cloud solution in which both travel companies and in destination business partners can be on boarded to provide personalized contents and offers. Space alma will be using industry available data attributes (flight, product) along with online demand context to create a generalized Travel Experience Quotient (gTEQ).

How it works

Alma simplifies personalization by abstracting the complexities of the AI and ML algorithms using a concept called “Travel Experience Quotient” (TEQ). TEQ is a combination of sTEQ) and gTEQ. For more details, refer to the submission documents including architecture.

The TEQ can then empower the travel companies and its partners to make tailor made offers to each customer at the right touch points along the customer journey. This unlocks the potential of further merchandising opportunities.

Figure 1: the concept

Concept Image

Figure 2: alma architecture

Concept Image

How we built it

We have created a mobile client with Local ALMA plugged to get the personalized feeds using flutter. Flutter Web is used to create the space ALMA admin console. As of now, the inventory is based on a firebase store. Machine learning and AI solutions which provide the TEQ values (both specialized and generic ) were built in python and integrated to FWs Django and ML kit. Space ALMA APIs including the ML endpoints are Rest based. Space ALMA consoles are built with angular 8, and the Simplified TEQ APIs are deployed as AWS lambdas.

We started with ML solutioning, we were able to create the TEQ based on micro-segmentation on the attributes and bayesian classification. The demand probability of a person will be identified in the Local Space and which will then combine in the Cloud Space to curate the dynamic offers.

Challenges we ran into

The scope is very huge and we struggled to focus on the use cases.

Accomplishments that We're proud of

TEQ - the concept is very energizing for the team, and the machine learning solution finally provides meaningful values that can be used for pricing, offer curation, bundling, and many more. Within a short span of time, we were able to make a working model.

Alt text

What we learned

All of us are new to flutter and flutter web, Somehow we cracked it.

What's next for Alma

On a Roadmap perspective, user role-based screens and workflows are not added for retail window and admin screens. As of now, the TEQ is working based on ML algorithms based on available data. The concept of reinforcement learning is on the road map to create offers and recommendations and explore to see the market reaction. NDC enabled API gateway is not yet created. Corporate policy upload options will be a nice feature to add. Converting business to leisure and mix and match to create bleisure will be interesting for Alma. Web plugins are not yet developed.

Links to Working suite :

There is a Github repo [link]

where you can find the admin panel and mobile client codebase The Mobile version is not uploaded to any store. So if you want to try out, take a pull and build it. Admin panel is live and you can find the link here.

[link] password : 123456

Machine Learning and AI solutions are exposed as Rest APIs, you can try out at using below Links .

Simplified TEQ APIs for testing :

Covid Status EQ [link] sample RQ :

 "country_code" : "IN".        
sample RS :

    "statusCode": 200,
    "body": "{\"zone\": \" red\", \"teq\": 0.67, \"international\": \"closed\", \"domestic\": \"partially\", \"interstate\": \"open\"}"

can be country code like IN, CH, CA,US,etc... 

Sample RQ:

 "FlightAttributes": [

 "dateOfTravel": "2020-6-14",
 "origin": "AYT",
 "destination": "COK",
 "fareClass": "Y",
 "carrier": "AI"


 "PassengerAttributes": [
 "posCountry": "AL",
 "pnrType": "NORMAL",
 "age": "30-40",
 "currencyCode": "EUR",
 "fareType": "BASE",
 "countryOfOrigin": "IN"


Sample RS : 

    "statusCode": 200,
    "body": "{\"TEQ\": 0.61, \"AlmaOffer\": \"ALMAXYUU13063\", \"Air\": {\"flight_number\": \"AI 917\", \"flight_date\": \"15-Sep-2020\", \"origin\": \"AYT\", \"destination\": \"COK\", \"fare\": 10370, \"fareClass\": \"Y\"}, \"Hotel\": {\"property_code\": \"HIL30\", \"property_name\": \"HILTON\", \"price\": 13703, \"no_of_days\": \"5\", \"offer\": \"30% off 3 Day Stay\"}, \"Car\": {\"provider\": \" OLA\", \"offer\": \"OLA-2 days\", \"cost\": 1350}, \"Other\": {\"type\": \"GUIDEDTOUR\", \"provider\": \"180Degrees\", \"offers\": \"30% off on drinks, pickup and drop\", \"cost\": 1359}}"

+ 18 more
Share this project: