Insights Do Not Implement Themselves

By: Steven A. Stanton & Michael R. Neece

Let’s separate myth from reality and ask, “Is Artificial Intelligence (AI) truly useful to you today?”

To get an answer, imagine you had a free month of the latest AI system, such as IBM Watson, and it generated an enormous amount of analyzed data for you. And that analysis yielded a treasure trove of “insights” about your customers and your operations.

What would you do with that?  What could you do with that?

• Could you easily connect those data-driven "actionable insights" into product or service enhancements to better delight your customers?
• Could you translate those insights into well-defined process improvement projects that would quickly lower costs, eliminate errors, improve compliance, and protect data privacy?
• Could you automatically start cyber security processes triggered by threat intelligence to protect your company?
• Could you digitally transform inefficient financial service processes into efficient workflows that reduce time-to-revenue or create new customer solutions?

Probably not. ...... For most organizations, there’s a wide gap between acquiring data-driven insights and turning them into actual value creation.

In reality, “Actionable Insights” from AI or any data source are simply more data “noise” and marketing spin. Achieving the promise of AI, still requires humans to take “augmented action and decisions” that complete the transformation of insights into value. Data by itself can never turn a broken set of tasks into a high-performing business process. 

“Actionable Insights” are simply more data “noise” and marketing spin.

For one example; An AI or IoT (Internet of Things) system might predict the possible failure of a key elevator or jet engine component, but a human must initiate the field service workflows to perform preventative maintenance and avoid equipment downtime.

For a second example; Population health data might identify meaningful interventions that can improve health outcomes for an at-risk patient group. But no results will be achieved unless clinical and non-clinical services are delivered through coordinated efforts of human beings.

Are You a Data Hobbyist?

In the never-ending cycle of SEARCH-COLLECT-ROUTE-ASSESS-and USE data, sadly, the only step that organizations are getting better at is COLLECTION.

The explosion of information creation has turned many of us into data hobbyists, gathering interesting specimens of data for our private collections. As avid hobbyists, we’re busy digitizing all things physical, environmental, and biological, including our own biometrics and locations. What fun!

The Internet of Things (IoT) is creating massive data lakes to make it easier to fish for insights. But even worse, our cyber-security devices that are meant to protect us, often just generate fragmented data streams and beautiful dashboards that inadvertently overwhelm network defenders, and leave no time to find “incidents of compromise” and launch remediation processes.

Just as we worship at altars of data, our data sources are becoming more fragmented. 

This focus on collection is facilitated by cloud computing and IoT technologies that have made data collection cheaper and easier. However, the proliferation of single-purpose apps and the persistence of legacy systems continues to fragment and isolate our data, creating data silos of unprecedented scale.

Just as we worship at altars of data, our data sources are becoming more and more fragmented. Many have experienced this dislocation just like the mythical six blind men who run into an elephant and perceive six different animals. Even after decades of trying, many of us still dream about integrated data to inform us.   We yearn for a single source of truth.

But we continue buying pretty apps that isolate data even more.

As a consequence, ROUTING, ASSESSING and USING the data insights lags far behind. 

The life blood of an organization is its data. Business processes are the arteries that deliver the data to the right people.

The blood that should nourish an organization is data. The pathways that facilitate ROUTING, ASSESING and USING data are well-designed and (change on the fly) business processes explicitly crafted to channel data flows to decision and action points where people create the most value.

To truly leverage data, a new process architecture is required. Like a network of arteries, each process must be connected to neighboring processes. For example, a “change request” in engineering or IT is routed through an authorization process, which then triggers implementation. Finishing the "change request" might then launch other processes in manufacturing, procurement, supply-chain, quality control & assurance, and customer service. Each process orchestrates the flow of data that informs each participant and organization affecting the delivery of value to customers, peers and trading partners.

In summary, organizations don’t yet take full advantage of data because:

1. Department leaders are incentivized to implement single-purpose apps that solve only their data SEARCH and COLLECT needs.
2. Single purpose apps isolate data, making it difficult to share (ROUTE) key information, preventing an integrated view of organizational status.
3. Processes (ROUTES) within each app (or legacy system) are hard-coded making workflow changes difficult and expensive, placing the organization at the mercy of each app vendor.   

Most firms see themselves as the center of the universe, rather than part of an ecosystem (network) of participants that need each other to thrive.

Are You An Innocent Bystander or Dora the Explorer?

As for SEARCHING, most organizations are still basically solipsistic, spending most of their energy internally, viewing themselves as the center of the universe rather than seeing themselves as part of an ecosystem of department, customers and partners. As a result, organizations rarely have well-designed external sensing processes that scans the environment for customer and competitor moves, or risks/opportunities in their supply-chains or value-chains. Instead organizations still rely on traditional customer survey data or anecdotes to perceive their marketplace.

Test Yourself

Here’s a quick test to assess your SEARCHING strength. If one of your employees noticed a new customer behavior who would they tell?  If they discovered a new competitor pricing plan, how could that news be exploited? How would these pieces of intelligence enter your data flows and process flows?

1. Where in your process designs are the intake for external insights?
2. How does your organization/department initiate pre-defined processes that automatically react to internal or external signals?
3. Is your organization still running “dumb” and “blind” while manually sending e-mails, spreadsheets, shared folders and hopping between multiple legacy apps?
4. Why is your organization filled with really smart people using antiquated methods? (To answer this question read Steven Stanton’s Book titled Smart Work: Why Organizations Full of Intelligent People Do So Many Dumb Things and What You Can Do About It https://tinyurl.com/y6vq8wtb )

If data doesn’t drive action, it’s useless.  

Data is only valuable for identifying meaningful opportunities and informing decisions.  If data doesn’t drive action, it’s useless. Most organization have developed scorecards to assess the performance and health of their business processes. But these metrics inevitably focus on outcomes. But decisions are where insights blossom into value. Do you have a way to measure your process efficiencies and decisions?

And, at the end of the day, who are your insighters. This is a tough one. With rampant simple automation and declining investments in training, who will identify the insights you need and who will capitalize them? How well do you train and educate your frontline employees to use information to make better decisions?

Without a clear data process (SEARCH-COLLECT-ROUTE-ASSESS-USE) embedded within your process architectures and populated by well-trained employees organizations can never turn data insights into value.