What is Augmented Analytics

In July 2017, the concept of Augmented Analytics was first introduced in a research published for Gartner. The topic immediately aroused interest and immediately became an impact trend in the field of Business Intelligence, so much so that just one year later Gartner published the report “Hype Cycle for Analytics and Business Intelligence“.

In its first definition, the concept of Augmented Analytics is described as a new data analysis approach that exploits Machine Learning and natural language (NLG) technologies, in order to automatically identify the most relevant results and independently suggest the concrete actions to be taken. Not bad as a goal …

A new approach

In fact, one of the main current problems is the fact that data tend to become increasingly numerous and increasingly complex. Often it is not possible to integrate them quickly also because of their extreme complexity. The risk of losing information becomes therefore very high, and precisely this complexity does not allow us to explore all the possible opportunities offered by the information available to us. It also increases the likelihood of having as a result just what we “looked for” and not what we “could find”.

We add the fact that, to date, all data extraction and transformation activities are still completely manual (“Artisanal” area of the figure shown below), considerably increasing the possibility of error. Furthermore, finding a data scientist today, a relatively recent profession that requires skills in information technology, statistics and economics, and for which there are still few studies, is really very difficult.

It therefore seems natural that the introduction of analytical tools capable of interacting with the human being in its natural language and autonomously identifying the most significant data without the mediation of analysts and without human intervention in general, becomes a fundamental key for the future of Business Intelligence.

The traditional approach

He is currently approaching traditional projects involving various figures such as analysts, computer scientists, data scientists and managers, involved in the usual decision-making activities that will then lead to the final request. This is followed by various data extraction and cleaning activities performed by computer scientists or technicians with specific skills. We will try to put together these data by linking them together and only then transforming them into information that is more synthetic and “decisive”.

As we can easily imagine the costs and times of such an activity are absolutely remarkable.

The Augmented Analytics approach

With this new approach we try to centralize all the procedures in a single solution, from data collection to analysis processing, to monitoring results. Machine Learning and Artificial Intelligence algorithms are used to automate the procedures making data analysis easier. Billions of data combinations are automatically analyzed, automatically finding correlations and identifying any predictive scenarios.

The future of Augmented Analytics

According to the slide (source Oracle Corporation), there are several levels:

Level 0 – Artisanal: everything is handmade, as in the classic approach. The data model and reporting are the responsibility of IT.

Level 1 – Self Service: data management is still largely manual, but human interaction with data will be done with the natural language (Natural Language Query). Visualizations and graphs will be suggested based on the data we are querying.

Level 2 – Deeper Insights: the first phases of advanced data management are displayed (recommended sources, joins, crowdsourcing suggestions, intelligent cataloging) and increased navigation helps to discover information that would otherwise require a strong human effort.

Level 3 – Data Foundation: data management is increased, corrections and enrichments are automatically identified. New views and new data sets are added.

Level 4 – Collective Intelligence: the system learns metrics and KPIs that alert you when they require your attention. You have both company KPIs and system KPIs. Insights become pervasive, commercial intent passes from an idea to a reality, results are expected, actions are recommended, but humans continue to act.

Level 5 – Autonomous: everything becomes really guided by data, with the best subsequent actions performed on the basis of forecasts, insights and intents. The system is the engine of change.

Where are we at?

I believe that for the moment we have stabilized between the third and the fourth stage, data management starts to be really “augmented”, it is becoming easier to integrate and manage a large amount of information and take the first steps towards automatic identification of KPIs.

Completing the fourth stage and finally reaching the fifth, the “autonomous” total, will be the challenge of the coming years. The way of conceiving Business Intelligence will change drastically, but the challenges to be faced are always more beautiful. We are ready?

Leave a Reply

Your email address will not be published. Required fields are marked *