It’s important – especially in this age of big-data and (in my opinion) excessive data hoarding – to remember that having data is not nearly the same thing as using it effectively and that it is the proper analysis of it – to get to insights, actionable intelligence and predictive analysis – that really matters.

Business Intelligence is broken

The process of turning data into useful information is costly, time consuming and can be quite elusive. According to Gartner, a mere 20% – 30% of all Business Intelligence initiatives are successful, mostly due to complex technology stacks, communications issues and their reliance on the past. 

Even when successful these BI initiatives – delivering descriptive analytics – fall short in several key areas ranging from self service to their ability to deliver predictive analysis. Their tendency to deliver little qualitative information also means that they tend to (again in my opinion) only deliver abstract view of the business, as seen through a rear view mirror.

Data is messy and data science is costly

Complexity of data means that data scientists spend 60%-80% of their time doing menial data wrangling before they get to the analysis itself and can start extracting value and since data scientists are rarely business experts they often need to be paired with business managers or -specialists to ensure that the results make business sense and proposes viable business actions. 

These challenges contribute to the fact that data scientists are now both a scarce and expensive resource. So much so, that most SMEs can’t afford their work. 

According to McKinsey, this problem will only continue to grow in the years to come and they project that America alone will need at least 250,000 additional data scientists by 2024.

We need a new approach – a new paradigm – for harvesting value from data.

Augmented analytics and data discovery

Augmented analytics will give much needed reprieve as it aims – by automating the process of business insights generation – to greatly minimize this manual labor and alleviate the strong data scientists dependency. 

For this Augmented Analytics relies on machine learning and artificial intelligence to automatically comb through data, clean it and perform analysis on it to deliver valuable information to enterprise executives, marketeers and small business owners alike.

A fairly recent concept

Augmented Analytics is a relatively new field. The first noteworthy mention of the term was in Gartner’s publication of 2017 where it is hailed as the next data and analytics paradigm.  

Experts believe that believe that Augmented Analytics will be the main selling point for most – if not all –  Business Intelligence solutions by the year 2020 and that it will transform business just as much as the Internet has done to date. 

In 2017, the global augmented analytics market size was estimated to be $4,094 million but analysts project the market size will grow to $29,856 million by 2025.

A fast tracking for the path from data to insights

With Augmented Analytics, there will be virtually no lag from the time a data queried to the time real insights are presented. 

For instance, instead of an analyst developing a report that shows overseas sales have gone down by 30%, Augmented Analytics should be able to tell you why it happened and – when full maturity is reached – make suggestions on how to address it. 

To put it simply; Augmented Analytics aim to turn novice BI users into data experts and help businesses of all sizes get more knowledge from their data requiring lot less effort and without direct involvement of data scientists. 

Radical changes with Autonomous Analytics 

Even though effective  business intelligence can give you a head start, the real test now is how long it takes to turn fresh data into insights and actionable intelligence. Businesses that succeeds in producing valuable business insights – in real-time – are likely to attain a sustainable competitive advantage. 

This is where Autonomous Analytics comes in. 

Autonomous- and Adaptive Analytics will dramatically speed up the process of data analysis and deliver actionable insights bundled with proposed actions to the right people at the right time.

The Cisco story

The  journey at Cisco systems is a good way to describe the impact of Autonomous Analytics for a business. 

Approximately a decade ago, Cisco systems started experimenting with artisanal analytics and used it to advise marketing organizations on the accounts they should focus on. 

The models they developed could predict what products customers would be interested in and this helped their clients to spend less on marketing while achieving greater returns. 

After about 5 years, machine learning was becoming more practical in computing and Cisco took advantage of the new wave to develop even more models. By 2014, they were generating approximately 25,000 propensity models every three months based on data from 160 million businesses all over the world. By 2016, they had grown to 60,000 models every three months. 

Today, they have employed several machine learning algorithms and it is now possible for them to see results that are 3-7 times better and faster than those of the earlier propensity models.

The Active Seek

The first things many executives do in the morning is to review their BI dashboards. They want to monitor trends, KPIs and other visuals that might impact their day. Others have set email alerts so they get notifications whenever something important is realized. 

Most of these tools take a long time to be developed and even though they can be somewhat effective, they are – in fact – very statistic and slow to adapt to change. They represent very well what Autonomous Analytics will make obsolete in the next decade.

As the field of Autonomous Analytics advances we will reach a phase of Active Seek which . Active Seak is where data actively seeks its consumer instead of the other way around. 

This represents a seachange which is estimated to become a reality in businesses by the year 2030. 

The AA driven business assistant 

Like fully autonomous vehicles reach a desired destinations – without human intervention – so Autonomous Analytics aims to reach theirs without any or much intervention. 

Once something of value is discovered it will be passed to the responsible individual or even – when this is of sales or support value – directly to the customers. 

Information will find its most valuable consumers regardless of whether they are employees, customers or partners. This distribution will obviously be controlled and monitored for security reasons.

This automation process will be fronted by bots and the main user interface for it will be voice and human-like interactions. These bots will also communicate their findings directly via APIs – to assist in process automation – and via integrated alert systems.

Conclusion

Business Intelligence tools today are – almost solely – driven by users. Current BI visualizations can accurately portray quantifiable information and they can represent the quantitative, “how much”, quite accurately.

BI become increasingly inadequate when dealing with the more elaborate aspects like “what is happening” to “why is it happening”. They have limited predictive abilities and are incompetent when it comes to action recommendations, monitoring of outcomes and then using those outcomes – via reinforced learning – to improve future proposed actions. 

This need – to know why something happened, what could happen next and what actions could/should be taken – is what is inspiring the growth of data science and the birth of both Augmented- and Autonomous Analytics. 

We can expect the fields of data science and business intelligence to merge somewhat as analytics tools become seamlessly integrated into business operations and we see the birth of the super assistants bots which offer personalized assistance to employees, managers and customers alike.