Who is Boxalino?

Open

Boxalino is a Swiss AI company developing and operating the Winning Interactions real-time platform.

The Winning Interactions platform is connected to existing e-commerce systems, e-mail marketing systems, apps, and portals. This makes it possible to provide each visitor to an online store, portal, or app with the best possible customized user experience in real-time and in an automated way.

The automatic learnings of Artificial Intelligence can be combined with your own specifications. These are configured as guidelines and specifications, as well as supplemented by their own algorithms and predictive models.

Thus, companies of any size can get the most out of the possibilities of data science and artificial intelligence.

These optimizations achieve an additional 10% to 20%+ revenue percentage and improve the margin and competitiveness accordingly.

What is Boxalino’s offering?

Open

What problems does the Boxalino Platform solve?

A broken bridge between UX and AI

The world of User Experience (UX) saw a great number of improvements during the last few years, from Headless CMS to Micro-services oriented front-end architecture vastly overperforming their monolithic ancestors both in speed and efficiency.

The same can be said about Artificial Intelligence (AI), the rise of Deep Learning over the last decade changed the domain drastically making it much more commercially relevant than ever before. In addition, platforms like Google BigQuery and Tensorflow greatly reduced the costs and efforts to do data science, effectively democratizing the domain for SMBs (who were not able to afford such systems before these new platforms).

However, significant hurdles remain in the way of SMBs, and they are in the bridge between these two worlds. To transfer the data from the Real-Time UX behaviors requires a lot of efforts (most companies do it only a little bit with Web Analytics platforms which are not adapted for Data Scientist to work efficiently with data) and to load the results of data science into the real-time operational systems like the e-shop of the Newsletter systems requires a lot of custom work and for data scientists to explain what to do to Dev/Op (IT) engineers who do not always understand each other easily.

The main issue is caused by a lack of standards and platforms supporting the way to collect data from your operational systems and to upload predictive models and other types of data back into your operational systems.

The Data Science Dilemma

One of the most problematic issues with Data Science for SMBs the amount of time needed for Data engineering (mainly collecting, cleaning, and organizing data) which represents a scary 80% of the time for most Data Scientists.

Data Engineering Overload, Reporting without Actions, Big but Bad Data

As a result, data scientists tend to be overloaded with Data Engineering and pain to produce dashboards and reports for the management.

But a Dashboard is not an action, the real goal is to load the result of the data science work (if possible in a fully automated way) back to the production environment, but there, the IT resources have to become involved and they are also often overloaded.

As a result, you are likely to be swamped below a gigantic pile of heterogeneous data from different sources, in different formats, all of which are quite big (which makes them costly and hard to transfer) but are not what you need to actually bring a business impact that is measurable with an a/b tests so you know what you gain out of it.

The solution with Boxalino Data Science Ecosystem

Boxalino offers a complete Data Science Ecosystem.

This means that with Boxalino, all the platforms necessary to optimize your online success with data science and artificial intelligence will be automatically there.

In addition, Boxalino offers many automated and easy to configure Best Practices.

And once you get to a point where you want to invest more in data science, everything is there for you:

• Get basic training on the Boxalino Intelligence Admin so that non-technical people can start to configure and optimize Boxalino Best Practices step-by-step with A/B Testing.

• Benefit from the value of our Certified Data Science Partners who can configure many standardized Artificial Intelligence projects in the Boxalino platform (e.g.: Customer Clustering)

• Find and train a Winning Interactions Hero in your team (not necessarily a technical person)

• When you are ready, engage your own Data Science Analyst to get trained on Boxalino

A Data Science Ecosystem VS a Black Box Engine

Unlike other systems, Boxalino is not a recommendation engine working as a black box and claiming to have “great AI” which you can’t actually really see.

Boxalino is an open Data Science Ecosystem that empowers your team with many reports and administration interfaces so that you can truly understand and influence the effect Data Science can have on your business.

The Boxalino Winning Interactions Platform

In this “Lamp” diagram, we illustrate the entire Boxalino Winning Interactions Platform.

1. Boxalino Intelligence Admin, WPO and Narrative (Layout) API

Boxalino Intelligence Admin lets you confirm all Boxalino Widget & Page Optimizers (WPO) : WELCOME, SEARCH, UPSELL, …

Each WPO covers a specific function of your e-shop or other systems (e.g.: newsletter).

Learn more about the Boxalino WPOs…

Each WPO is delivered to your Website as a Narrative with a unified API: the Boxalino Narrative (Layout) API.

Learn more about the Boxalino Narratives.

2. Integrate Boxalino Narratives in your Website

Your IT can then integrate the Boxalino Narratives in your Website, which is becoming now an A/B Testing Front-end.

The Narratives can be integrated in any shop and on any (and all) pages to cover the most important pixel real-estate of your website (main part of the home page, product listing pages, search, brand pages, landing pages, PDP, basket, …)

Learn more about how to integrate Boxalino Narrative (in Magento, Shopware, or any other system).

3. Structured In-page tracking of the user behaviors (scrolling, clicks, …)

Boxalino collects every call to its API (both the structured request and its response) and makes it available in the Open Data Warehouse.

But that’s not all! Boxalino Tracker observes the scrolling and clicking behaviors of each visitor on each page and identifies (according to the screen of the device) what exactly appeared on the screen, for how long the customer looked at it, and what was clicked from where (exact KPI attribution).

Learn more about How Boxalino tracks structured data automatically

4. The Open Data Warehouse (CODW) on Google Cloud Platform (GCP)

All the data collected by:

• Data Exports from systems (products, customers, transactions, ratings, vouchers, bundles, etc.)
  Learn more about our Data Integration Structures

• Every request and response to our API

• Every event tracked by our tracker

Is automatically consolidated in a standardized and highly structured Data Warehouse hosted on the Google Cloud Platform (GCP) on a simply fantastic
Big Data technology from Google called BigQuery.

The Data Warehouse will be there for you, even if you don’t plan to use it at the beginning. But it will have collected months (and maybe years) of data for when you decide to start working with them!
Learn more about our Open Data Warehouse (CODW)

5. AI Automation (BAIA) & Analytics, Insight, Reports (B/C AIR)

Boxalino automates the generation of hundreds of pages of relevant and actionable reports automatically with Google Data Studio

Learn more about Analytics, Insight, Reports BAIR (use the standard reports) and CAIR (build your own reports on the basis of our reporting tables)

Boxalino also provides advanced standardized Artificial Intelligence Automation (BAIA) so you can benefit from the value of Machine Learning without having to develop it yourself (e.g.: Deep Collaborative Filtering, Customer Clustering, …)

6. Marketing Automation with the Data Science Lab (B/C DSL)

All the results of the automated reports (B/C AIR) and standardized Artificial Intelligence (BAIA) are loaded in another Data Warehouse, the LAB, which instead of collecting coming from your systems, collect all the data that should go back in the real-time platform of Boxalino (and other systems like your Newsletter) to be used by your marketers in an actionable way (improve recommendations, search relevance, personalization, communication journeys, etc.).

This is the second and very important piece of the puzzle, the standardization of deploying data science into your operational systems.

Learn more about the Data Science Lab BDSL (when done by Boxalino and its Certified Partners) and CDSL (when it is done by your team).

7. Integrate Boxalino with your own Data Science team

Finally, while everything will work well automatically with the standard system of Boxalino (the Lamp part of the illustration), Boxalino takes even more sense if (or maybe more when) you want to develop your own data science activities.

You can:

• integrate Boxalino Open Data Warehouse (CODW) with your Own Data Warehouse (consolidate the data with other sources like ERP, CRM, PIM, or your Online Ads)

• do your own Analytics in Google BigQuery (or other systems)

• do your own Reporting in Google Data Studio (or other systems)

• Integrate external AI providers or develop even your own Predictors using AI

• and load the results in the Data Science Lab so that the data goes automatically to your operational systems (website, newsletter, app, …)

And for when you are ready to get involved into Data Science

The different types of Data Scientists

As we explained in detail, in How to orient your Company towards Data Science? , there are 3 key profiles of Data Scientists:

1. Data Engineer

2. Data Science Analyst

3. Core Data Scientist

Each focusing on a different section of the Data Science Pyramid of Needs:

What Boxalino exactly means for a Data Science Analyst?

Boxalino provides the entire ecosystem to your Data Science Analysts. All the hurdles are removed!

Your Data Science Analysts can quickly and efficiently:

• Analyze data and reports fact from our BigQuery-based Open Data Warehouse (CODW)

• Find advanced patterns and causes with sophisticated Big Data Analytics based on the truly gigantic of highly structured behavioral data we collect from every scrolling of every visitor on every page

• Train AI predictive models easily in SQL using BigQuery ML

• Load these models and any other data to the Boxalino Real-Time platform to be used in your Website and other operational systems (Newsletter system, app, …)

• Configure A/B testing experiments to see the exact impact of each of the data science inputs on your business success

• Analyze the A/B testing experiment results with high granularity in the Big Data to identify exactly how the experiment was successful and for which segment of your visitors and/or customers