Putting Analytics and AI in Context for Better Outcomes
How do organizations really gain value from analytics? Deliver results and drive change at your enterprise using these experts' advice.
- By David Stodder
- July 15, 2019
One of the errors of traditional business intelligence (BI) has been its standalone, passive relationship to human decisions and business processes. A BI report or dashboard delivers information, but how that information is connected to the decisions and actions that users need to take is often unclear and unstated. The information may or may not be relevant; it could even be out of date or misleading. As analytics applications expand beyond specialists, organizations need to ensure that data insights are better connected to what humans and automated applications and services will do with them.
This issue was discussed extensively at the recent TDWI Solution Summit in Coronado, CA, which focused on the theme of "strategies for delivering results with analytics and data science." In his keynote, James Taylor, cochair of the Summit and CEO of Decision Management Solutions, offered the clearest context for analytics: the decision.
Rather than dive immediately into collecting often ill-defined requirements for developing analytics models and algorithms, Taylor suggested that an organization should identify decision points in its business processes, customer engagement, and strategy development and think about how these decisions could be improved. This approach will naturally link analytics to business goals. For example, an organization could look at whether changes in decisions about pricing, claims handling, and service renewals could increase customer satisfaction.
Although many organizations are looking at how to use artificial intelligence (AI) and process management technologies to automate decisions, in most cases decisions are formulated and executed with considerable human involvement.
In his talk at the Summit, Rob Horrobin, AVP of Advanced Analytics and Planning at Pacific Life Insurance, drove home that organizations have to balance people, process, and technology aspects if they are to achieve good outcomes applying analytics to real-world situations. He recommended that organizations composing data science teams to develop analytics models and algorithms should ensure that they include personnel with domain expertise, communications skills, and an understanding of the user experience.
Embedded Analytics and Recommendations Systems
In the technology realm, we are seeing exciting developments in the use of AI to provide prescriptive insights in the form of recommendations to users of BI tools and other applications. For users of BI tools, AI can provide recommendations about which data sets to use in predictive analytics. The software can be smart enough to understand, for example, that the user is trying to examine the effectiveness of marketing campaigns across channels and therefore should include a source that has relevant customer behavior data. AI can even take things a step further by not waiting for user actions and instead applying machine learning to automatically find patterns in big data and deliver answers to users that are relevant to their roles and responsibilities.
In some ways, the trend toward AI-driven recommendations represents the next generation of embedded BI and analytics. Organizations have long had an interest in bridging the gap between BI and analytics tools on the one side and business applications on the other. Traditionally, users have had to step out of their CRM, ERP, or other business application environments to get the full functionality of a BI or analytics solution. Embedded versions of BI and analytics tools have tended to be primitive, offering simple reporting-oriented dashboards, alerts, and limited query capabilities.
Although the limitations can be frustrating to power users, the simplicity is appropriate for the majority of business application users who typically do not want to climb the learning curve of a more complex BI or analytics solution just to consume information. With AI-driven recommendations and information delivery, organizations can still keep things simple but allow users to tap richer sources of data and learn answers to questions they may not have thought to ask.
Some modern BI solutions use AI to discover and surface simple visualizations of information relevant to the user's decision. An example is MicroStrategy's HyperIntelligence, part of the company's 2019 release, which offers "hover over" insights that pop up as users look at forms, files of customer information, or other business application interfaces. Competing solutions such as ThoughtSpot help nontechnical users explore data related to their decisions through natural language searches. AI can enable such systems to learn from user behavior and data characteristics to shorten the path to relevant, accurate answers, including presenting the insights in the form of alert messages as users are involved in business processes.
AI-Driven Performance Management
Performance management has long been a way to tighten the connection between BI reports, dashboards, and alerts and users' roles and areas of accountability. Using key performance indicators (KPIs) and other metrics, organizations can communicate strategic corporate objectives, sometimes as part of the implementation of business performance methodologies, to guide decisions and actions.
AI can drive improved understanding of metrics by bringing relevant information to users automatically rather than waiting for them to write a query. Organizations can also use AI-based features in solutions to develop recommended actions. For example, an organization may use these recommendations to address problems in processes and behavior that are causing customer satisfaction metrics to fall below corporate objectives.
Performance management depends on data quality and consistency. If users cannot trust the data, they will not trust the KPIs and other metrics. Data quality, data cataloging, and other data management solutions are using AI to help organizations accelerate the improvement of data quality through quicker discovery of discrepancies, anomalies, and inconsistencies across sources.
Analytics in Context: Essential to Achieving Success
As many speakers at the TDWI Solution Summit pointed out, most organizations invest in analytics and AI to drive change. They want to use data effectively to make better decisions, improve customer engagement, and run processes and operations smarter and more efficiently. However, these initiatives will fall short if AI and analytics development is not well integrated with how humans make decisions and take actions.
Prescriptive, AI-driven recommendations will misfire if their development does not take into account how human decision makers will employ them, whether users can trust the underlying data behind them, and whether the recommendations are relevant to the outcomes they are trying to achieve. Organizations that take these human factors into consideration will be able to move beyond the limitations of traditional embedded BI systems.
About the Author
David Stodder is director of TDWI Research for business intelligence. He focuses on providing research-based insight and best practices for organizations implementing BI, analytics, performance management, data discovery, data visualization, and related technologies and methods. He is the author of TDWI Best Practices Reports on mobile BI and customer analytics in the age of social media, as well as TDWI Checklist Reports on data discovery and information management. He has chaired TDWI conferences on BI agility and big data analytics. Stodder has provided thought leadership on BI, information management, and IT management for over two decades. He has served as vice president and research director with Ventana Research, and he was the founding chief editor of Intelligent Enterprise, where he served as editorial director for nine years.