2022 - The Year of the Decision?

Most people are well familiar with terms such as digital twins, RPA, IoT, 5G, smart sensors, blockchain, predictive analytics, AI, data science and machine learning - but far fewer have heard the terms ‘cognitive automation’ and ‘decision intelligence.’

2022 is likely to be the year that changes.

The term 'decision intelligence' was first coined by machine-learning pioneer Dr Lorien Pratt (see her TEDx Talk from 2016 on the subject here) and in this talk, she declares it as the ability to "understand the connection between decisions we make today and the outcomes they produce tomorrow".

Cognitive automation and decision intelligence (CA/DI) are often used interchangeably and refer to the recently acquired ability to take vast amounts of enterprise-wide data and apply advanced machine learning capabilities to create simulations, models and predictive and prescriptive analytics that can then be used to augment human decision making or automate the execution of decisions based on pre-defined rules.?

Unlike tools such as RPA, which are focused on the automation of knowledge worker activities such as data capture and back-office system updating, CA / DI applies machine-based intelligence to the task of making and automating decisions. Because let's face it, that should be the ultimate goal of any investments in digital technology – make better and smarter decisions. ?

These can be:

• Operational decisions, such as how to prevent stock-outs or ATP failures while minimising stock obsolescence, or how to minimise the cost of delivery while maximising truck fill rates.
• Tactical decisions, such as when and where to procure goods to ensure availability while correspondingly minimising carbon emissions, or how to maximise the effectiveness of trade promotion spending while understanding the projected impact on instore availability and profitability.
• Strategic decisions, such as guiding decision-making around where to source key components, utilising forward projections around commodity prices, geopolitical forecasts, shipping container rates and exchange rate predictions combined with other factors such as vendor sustainability scores.

The really important point, as Gartner highlighted in its recent report on Decision Intelligence, is that in our current messy VUCA world these decisions are no longer nicely separated but are interconnected and have tactical, strategic and operations consequences. They also highlight that as we move to an ever-more connected, contextual and continuous world, the complexity of these decisions will increase exponentially.

People make thousands of these decisions every day across the enterprise, but they are disconnected, often counter-productive due to this disconnection, and we never really analyse and consider how we made these decisions, what the outcomes were, and whether we could make much better ones.

Well, now we can.

Decision intelligence enables companies to not just digitise their current processes but gives them an opportunity to understand why they do what they do and to codify this rationale, so it becomes embedded into the organisation's memory bank, allowing it to be analysable and usable in the future.

This codification enables the machine to predict what the likely outcome of such decisions will be, based on multiple criteria, so that the user is presented with recommendations as to what actions to take and why this course of action is proposed compared to others. These recommendations could be as a result of simulations, modelling, data science and machine learning, and include data sources that traditionally would have required inputs from across the enterprise, involving multiple meetings and extensive preparation. Unlike the past, now the decisions are made using a forward-looking consideration of the potential impact of the decision on numerous elements, not just a single one. These decision intelligence systems are also constantly on, analysing the supply chain 24/7 across all SKUs, able to access data at a volume, frequency, granularity and accuracy that is far outside the capability of humans.

For example, while making the decision as to what mode of transport to use to ship goods, the system could consider the impact on CO2 emissions, inventory, service, lead times and cost using each potential transportation mode. It can also look at all other planned shipments from the relevant locations and consider whether to wait and consolidate, change the mode of transportation, or change vendor. It can do this analysis in full appreciation of the impact each option has on each of the strategic elements. The final decision/recommendation, therefore, is made in conjunction with defined strategic imperatives and with complete visibility as to the impact on carbon, cost, lead-time, inventory, service and profit. The ability to undertake this multivariate analysis at an operational level is far beyond the capability of humans to analyse other than for only a select number of SKUs. The weighting allocated to each of these strategic attributes can also be defined by the business based on need, allowing for strategically focused decision making – for example, in some supply chains, service will take priority over cost and carbon, while in more cost-sensitive supply chains, the reverse may be true.

Over time, once the system is trusted, it can be left to its own devices, becoming effectively ‘self-driving’ and able to update the underlying transactional systems without human intervention. But it never becomes a black box, constantly providing visibility of the logic and data behind the decisions it makes, the actions undertaken, and the outcomes achieved. All of this is then stored in a cognitive data layer that is analysed and used by the models and machine learning algorithms to create ever-smarter outcomes, creating a thinking and learning digital brain whose memory serves to create an organisation-wide increase in intelligence - both human and machine.

The market seems to be waking up to the potential of this new technology:

• Several articles appeared last year highlighting the potential of Decision Intelligence. CIO magazine called it ‘The New BI’ and Raconteur did a piece on how it could democratise analytics.
• Gartner has officially recognised Decision Intelligence as a new category and one of the Top Strategic Technology Trends for 2022 and predicts that by next year more than a third of all large organisations will have analysts practising decision intelligence, including decision modelling.
• And this April, Decision Intelligence will receive its ultimate recognition – a ‘For Dummies’ guidebook.

DI/CA represents not just another digital tool in the transformation armoury but the ability to create an exponential increase in an organisation’s decision-making capability. Too many companies approach digital transformation with a focus on the digital element but little focus on the transformation part, with the resultant outcome being that they focus on digitising their current processes rather than take the opportunity to re-evaluate why they do the things they do - and how they can use these new tools to achieve results never before possible.

While it is natural to think that this means that human knowledge workers are now redundant, the reality is that their importance is elevated and their time redirected to more important, value-focused tasks. This technology enables the augmentation and amplification of human intelligence and capabilities, not their replacement. The success of any digital transformation project will always require a combination of people, processes, systems and data, and often the efforts put into data analysis and modelling are lost because of a disconnect between these tools and the people who use them. As many a data scientist has found, sometimes the biggest challenge is getting the business to understand the outputs. It is therefore imperative that data intelligence systems can translate their recommendations and the logic behind them into a language that a human can understand.

The companies that are leading the way in developing cognitive/decision intelligence operating systems, such as Aera, have understood the importance of this fact and include graphical user interfaces, decision-explanation elements and Natural Language Processing (NLP) capability to ensure that all this cleverness is not lost in translation, and results in real change.

Organisations capable of embracing the opportunity to use this technology to re-evaluate the critical decisions they need to make in order to achieve successful long, medium and short-term outcomes will be able to finally free themselves from the confines of functional thinking, imperfect information and reactive, gut-feeling based decision making. This will bring a level of competitive advantage far greater than what we have seen to date; much more than can be achieved from the raft of process-focused systems or data visualisation tools. They will have truly 'digitally transformed', not just digitized.

Over the coming weeks, I will be writing a series of posts explaining more about the potential of this new technology and its ability to transform organisational performance, how to become decision-driven, the shifts that need to take place, how to address the natural cultural challenges, and what the benefits could be to your organisation.

Sean.