Register now for - "Turbocharging Data Management and Analytics with GenAI Assistants" - Thursday, March 21, 2024 at 11:30 a.m. Eastern Time.

Five Steps to Succeed with Your Continuous Intelligence Strategy

It’s easy to buy into the vision of continuous intelligence. What decision makers would not want to use real-time insights to improve their operational command of the business?

As defined in my first blog in this series, continuous intelligence applies streaming and historical analytics to both internal and external data, to adjust operational processes in real-time. It often integrates with application or machine data architectures, and assists ITOps, DevOps and customer engagement processes.

The challenge, of course, is making continuous intelligence work in today’s complex enterprise environments. You need to implement new technology, manage heterogeneous architectures, and teach busy staff to take on new work. Sometimes, the cost outweighs the benefits, but many companies have discovered how to reap business value.

This blog offers techniques to tip the cost-benefit scales in your favor. To succeed with continuous intelligence, data analytics leaders need to start with simple but scalable use cases, adapt their teams’ skill sets, and sequence analytics and operational workflows to reinforce one another. They also need to vigilantly monitor and retrain machine learning models, especially when business conditions change, and seek external data to enrich their insights.

Let’s examine these five steps, with a focus on customer engagement examples, to understand how to minimize the cost and maximize the benefits of continuous intelligence.

  • Start with simple use cases that scale. Minor changes often have large impacts – which brings both opportunity and risk. To minimize the risk, try to start with simple use cases; for example, that add just one new task, tweak one machine learning model, or integrate analytics output with the robotic process automation you already have. Test these changes in smaller environments– one region, one line of business, one customer subset – and gauge their results. Then maximize the opportunity by scaling regions, business lines or customers.
  • Adapt the skill sets of existing teams. Data leaders should build an inventory of team skills, then compare those to your use case requirements. Play to your bench strengths and objectives, and try to avoid disrupting jobs or forcing significant re-training. Recruit outside talent where necessary to fill gaps.
  • Sequence your analytics and operational workflows. Continuous intelligence can be particularly effective with online customer engagement. For example, ZestMoney provides shoppers with no credit history the ability to borrow and spend money. Their platform analyzes applicants’ creditworthiness in real-time, using contextual demographic data. It then activates an appropriate line of credit, so they can set up monthly payments and shop in ZestMoney’s ecosystem of more than 3,000 brands. So operations trigger analytics, which trigger operations, creating a seamless workflow that maximizes customer value.
  • Continuously check and retrain ML/AI. The e-commerce site Flipkart, hospitality site MakeMyTrip and food aggregator Zomato all use machine learning to generate real-time recommendations as part of their continuous intelligence systems. They must monitor results to ensure accurate predictions in the fast-changing COVID economy. To identify data drift quickly in your environment, consider using BI software to go back and visually compare machine learning predictions with actual customer decisions. You can intuitively spot issues, then re-train and iterate your model in a continuous loop.
  • Find new external data. Publicly available datasets, for example related to market behavior, social media trends or the weather, can enrich many of your continuous intelligence use cases. Think creatively to identify these opportunities, then tap free open data platforms or commercial data exchanges such as data.world to procure datasets that improve the impact and accuracy of your real-time analytics.

Continuous Intelligence offers both significant upside and significant costs and risks. Define what you have today – teams, skills, processes, architecture and tools – and focus maximum effort on the least disruptive opportunities to generate results in that framework. In this way, you can achieve small wins that multiply and grow over time.

Kevin Petrie

Kevin is the VP of Research at Eckerson Group, where he manages the research agenda and writes about topics such as data integration, data observability, machine learning, and cloud data...

More About Kevin Petrie