Why Artificial Intelligence Analytics Works

December 10, 2019
5 min read
Microsoft Italia and iGenius bring the Generative AI into the world of numbers

Analytics & AI is changing business.

Many companies can’t find the right skills and resources to put their analytics in order.

To prepare their data properly and set it up so anybody in a business can use it to make better decisions.

Tough decisions have to be made. What’s the most appropriate Big Data architecture? How should ETL (extract, transform, load) pipelines be designed? What’s the best dashboard tool to show it all to end-users?

This leads to a complex ecosystem of tools — or what iGenius Data Scientist, Manuela Detomaso, calls — a “melting pot of software which can only be used by a small team of experts”.

“Data analytics becomes an elite domain, with low levels of user engagement and awareness of the company’s data,” Manuela added.

What if you could skip all this complication, and just ask for what you’re looking for? Have your own virtual data analyst.

This is what an AI advisor delivers — a conversational interface where data’s processed automatically, displayed simply and scoured by AI to serve up context-specific insights just like an extra colleague.

Here’s how it works.

Simplifies Data Processing

There’re many things to consider when building a data analytics infrastructure, such as:

  • How to connect data sources
  • Whether to move the sources to the cloud
  • What type of information should be selected
  • What graphics are best to visualize the analyses
  • If the analytics workflow should be defined by coding or with a proprietary software

All these decisions are “overwhelming”, Manuela said, “and can get us further away from our main objective to get meaningful insights out of our data”.

An AI advisor “takes the burden of such complexity away from the user” and makes it simple with automatization.

Manuela likens the whole process to buying a car.

“Let’s imagine you want to use a car. You might not have any idea how fuel is burnt and transformed into energy, or care how brakes work, or how the car chassis connects to the suspension and wheels. You just want what a car can offer — a pleasant travel service. Why should it be any different when it comes to data?”

Simplicity is at the core of an AI advisor. It takes over the tricky preparatory processes, and makes the real benefit of your data more accessible, in a shorter time. This “ensures an amazing trip around your data world,” Manuela said.

Once your analytics infrastructure is set up — the AI serves up a range of features to help you get even more from your data.

Goes Extra Mile

All the automation allows an AI advisor to answer a wide range of questions in a simple conversational interface. But how does it go beyond?

🧠 Business Knowledge Graph

This is the brain behind the AI. It organizes a company’s raw data and, with Machine Learning algorithms, makes meaningful links between multiple data sources based on your context. This helps make every response thorough, relevant and personalized. Here’s more about how it works:

Smart Proactive Notifications

Not only can an AI advisor monitor data constantly, detect important changes and tell you about them, its algorithms can learn from the user’s habits and notify them at the time that suits them. A bit like our AI advisor, Crystal's Proactive Notification & Alerting talent.

💪 Response Enrichment

The goal of an AI advisor is to create an “authentic conversational experience”. This is where “automation intervenes to understand and provide enriched responses, beyond what has been exactly requested, so that a specific piece of data can be contextualized with a number of related bits of information,” Manuela explained.

So let’s say you’re a Sales Manager and ask the advisor for a roundup of last month’s sales. The AI advisor might pair the response with an update that sales are forecast to go down next month. It gives you information you need, before you know you need it. Just like a real colleague would.

💡Insight Discovery

This one’s similar to Response Enrichment, but instead based on the visualizations. As Manuela explains, “Maybe you ask for a split of sales by country. Insight discovery would explore this in different dimensions and visualize other aspects of the topic without you directly asking. So, it might offer a graphic on year-on-year sales, or the sales dynamic by sub-region.”

This all helps give you a complete picture of your data — in ways you may not think of straight away. But an AI advisor will.

📈 Self-Service Forecasting

It’s simple to load an AI advisor with prebuilt forecasting models. But, the algorithms can take it a step further by learning from the data and generating its own over time. That means it’ll spot patterns, make its own predictions based on your needs and get smarter as you use it.

❗️Smart Alerting

On top of the user being able to set custom alerts, an AI advisor can “generate alerts based on provided threshold values or even automatically,” Manuela said.

The advanced algorithms can “spot unusual events and notify users”, so they never miss a thing. Like self-service forecasting, this comes from the AI’s ability to grow over time and become more adapted to a user’s particular role and context.

💬 Recommender

Rather than adding information to responses (like in Response Enrichment), the Recommender system uses its conversation training to “generate automatic suggestions of questions the user might want to ask next”. This is another feature covered inContext-Driven Response talent.

Augmented Analytics

From data preparation to visualization and more, AI-powered analytics — or Augmented Analytics — can “democratize access to the knowledge arising from data,” Manuela said.

This is, after all, its “very first mission”. A mission to “provide simple, rich and insightful content that enables anybody to make better decisions, making their lives simpler and more data-driven”.

Similar stories