For years, getting value out of data has required a scarce profile: someone who masters the BI tool, knows what to ask and has the time to explore the data chart by chart. Augmented analytics exists precisely to ease that bottleneck: to let the machine help with the more mechanical parts of analysis so that people can focus on what matters.
In this article we explain exactly what it is, with real examples you can use today, what it is not (important), and how to start making the most of it.
What augmented analytics is
Augmented analytics is the use of machine learning techniques to assist people in the data analysis process. Instead of the analyst having to do everything by hand, the system collaborates on several tasks:
- Automatic insights. The system explores the data on its own and highlights patterns, correlations or relevant figures you might have missed.
- Anomaly detection. It identifies outliers or behaviour outside the norm without you having to go looking for them.
- Assisted data preparation. It suggests how to clean, combine or enrich data, reducing the manual work that precedes analysis.
- Basic natural-language queries. It lets you type a question in plain words —“sales by region last year”— and get a visualisation, without building the chart step by step.
The underlying idea is to reduce human effort and bias in exploration: the machine reviews far more combinations than a person could check by hand, and it does so without preconceived ideas.
A term coined by Gartner
The term augmented analytics was popularised by Gartner, which flagged it as one of the major forces reshaping the analytics and BI market. Its thesis: machine-learning-assisted capabilities would move from being an extra to being embedded in the core of the platforms, changing how analysis is produced and consumed.
A few years on, that prediction has come true: the three leading tools on the market now include these capabilities as standard.
Real examples in 2022
This is not future theory. You can use augmented analytics today in the most widely used tools:
- Qlik Insight Advisor. Qlik’s assistant generates visualisations from questions, suggests relevant analyses and highlights insights automatically over your data model.
- Tableau Explain Data and Ask Data. Explain Data analyses a specific mark in a chart and proposes statistical explanations for why it has that value; Ask Data lets you query the data by typing questions in natural language.
- Power BI: Q&A and quick insights. The question-and-answer feature (Q&A) responds in natural language with visualisations, and quick insights automatically searches a dataset for patterns.
Three different tools, the same direction: bringing analysis closer to non-specialists and speeding up the work of those who are.
What augmented analytics is NOT
Here it pays to be clear, because there is a lot of confusion: augmented analytics does not replace the analyst, it assists them.
- It does not make decisions for you. It highlights a pattern; interpreting it in the context of the business is still human work.
- It does not guarantee the insight is relevant. The machine finds correlations; telling coincidence from cause, and the important from the anecdotal, requires judgement.
- It does not fix bad data. If the data model is poorly defined, automatic insights amplify the error rather than correct it.
It is also worth placing it correctly: we are talking about machine learning applied to assisting analysis, not systems that write conclusions on their own. It is a support tool, powerful but bounded, that pays off in the hands of someone who can read its results critically.
How to start making the most of it
You do not need a big project to get started. A few practical recommendations:
- Start from a well-defined data model. Augmented analytics shines on clean data and clear metrics. Without that foundation, it produces noise that looks like rigour.
- Begin with the features you already have. If you use Qlik, Tableau or Power BI, these capabilities are probably already in your licence. Turn them on and try them with a real case.
- Use it to accelerate, not to delegate. Let the system do the initial exploration and anomaly detection; reserve interpretation and decision-making for people.
- Train your team. The advantage is not having the feature, but knowing when to trust an automatic insight and when to distrust it. That is where Business Intelligence training makes the difference.
The machine explores, the person decides
Augmented analytics is one of the best pieces of news in recent years for data teams: it removes mechanical work and puts more analysis within reach of more people. But its value depends entirely on two things: well-governed data underneath and people with judgement on top.
At Digital Fox Data we help activate these capabilities on a solid data foundation and train teams to get the most out of them, through our Business Intelligence consultancy. If you want your tool to work for your team and not the other way around, let’s talk.