Several posts caught my eye last week while I was at the Business Rules Forum on the topic of Business Intelligence (BI) and made me wonder if “BI” is the best way to build an intelligent business. Claudia Imhoff started me off with an article on Operational Business Intelligence – A Prescription for Operational Success in which she said that

“Business intelligence has “invaded” the operational space in a big way, offering in-line analytics, real-time or near real-time decision-making support for all employees in the enterprise.”

She made some good points, especially on some of the challenges of upgrading your data infrastructure to cope with being more real time and more available so as to support operational processes but she left me wondering if the same old tools, upgraded, were really going to help. Would it really make your company act more intelligently, at an operational level, if you gave ever more people the same kinds of software tools? While wondering about this I saw Curt MonashRob Meredith post this one on The Myth of BI for the Masses in which he took issue with the idea that a single BI infrastructure could really support all the decision makers in a company. He felt that different investments might work better than broad deployment of the same tools. He also made a great point, as he said that the real value for BI comes “if the decision is a strategic one (the real sweet-spot for BI)”. His point was that these kinds of strategic decisions require a powerful but ad-hoc collection of decision support tools rather than a broad-based BI framework. This made me think about the kinds of problems that do not meet his definition and I came back to the whole idea of operational BI. Looking back to the quote from Claudia’s piece I have to wonder if the need is for decision-making by the systems of the enterprise rather than the employees. Not supporting more people in making decisions but automating more of those decisions instead. Like CurtRob, I think the power of BI comes in helping with bigger, less structured decisions both in terms of seeing what might happen and reporting on what actually happened. Reports, dashboards, cubes - none of this seems very helpful when you are considering operational processes. Partly this is because these operational processes often require automation and that means that the decisions within them must be made by systems not employees. Partly this is because the staff who are involved are not analytically trained or inclined, making it hard to see how they could use these tools anyway.

I was gearing up to write a post on this when, over on the Enterprise Resilience Management Blog, there was this post on Algorithms and Business which reiterated that “mathematical models are better than humans in making a number of predictions” (as Ian Ayres noted in Super Crunchers). This was a great post on the power of algorithms, predictive analytics and data mining that is, to make better decisions than people once there is any kind of data history to use. Indeed, studies even show that a human supported by an algorithm, while still doing better than an unassisted human, does not do as well as the algorithm alone (the post also referenced this article). This trend to use technology, and algorithms, to automate more decision making is the core of enterprise decision management - automate and improve the operational decisions that drive your business. Three quotes really struck me in this article:

• “The newest space, and the one that’s most exciting, is where machines are actually in charge, but they have enough awareness to seek out people to help them when they get stuck,” he said — for example, when making “particularly complex, novel, or risky” decisions.”
  This is the classic use of what we call enterprise decision management. Handling the vast majority of decisions of a certain type while referring the remaining exceptions to a human for review. Insurance underwriting, claims processing, fraud detection and originations.
• “embedded business logic allows organizations “to codify and centralize its hard-won knowledge in a concrete and easily transferable form, so it stays put when the experts move on.”
  This is one of the key values of using business rules, and predictive analytic models, to automate a decision. As the population ages and more and more knowledgeable workers retire, companies can no longer rely on people to make decisions even if they want to. There simply will not be enough people with the right skills as the boomers retire and, even if there were, it is not obvious that the new generation would want to do the same job or that the companies could reliably train all these new workers in the necessary skills. Decision automation and management may well be the only option.
• “As far as most businesses are concerned, these problems typically fall into two types: improving various processes, such as how a network is configured and a supply chain is run, or analysing (sic) data on things such as customer spending.”
  I think the range of decisions to which this approach is being applied is growing rapidly. Not only do companies now have more and more data to analyze, they are also automating more processes and they want more of those processes to run without human intervention. It is no longer just that a company will analyze customer spending, now they will use that analysis to drive decisions that affect how a customer is treated, what offers get made to them, what price is used. Real-time embedded decisions.

Just when I thought I was done I saw that Jill Dyche had this post SAS and Teradata Show Their Hand in Vegas in which she referenced the new alliance between SAS and Teradata. One of the most interesting areas for me in this alliance is that the two companies will work to ensure that SAS analytic models can be executed, with great performance, in the Teradata engine. This means that the combination of the two firms is more powerful than ever for the delivery of enterprise decision management solutions. Rapid, in-line execution of analytic models is a critical component of the more sophisticated decision management solutions I see.

So, if you are trying to build an intelligent business, your way forward may not be to focus on your BI tools and on how to get more people to use them. Instead, you focus should perhaps be on how you can turn your data into algorithms that can drive your systems to make decisions automatically. To make your systems smart enough.

Technorati Tags: algorithm, BI, boomers, business intelligence, business rules, data mining, decision management, EDM, enterprise decision management, operational BI, operational business intelligence, predictive analytics, SAS, Super Crunchers, intelligent business, Teradata

Related posts:

1. SAS improves its support for EDM
2. Live from DAMA - A Reference Architecture for Integrating an Active Data Warehouse into the Real-Time Enterprise
3. New Research - Dynamic Workloads and Data Warehouses
4. BI’s New Frontiers - “not your father’s BI”
5. Using decision management to deliver intelligent business performance