Putting AI to work in your organization: You’ve got to adapt your processes

By Bryan Kirschner, Vice President, Strategy at DataStax

In a previous article, I advocated for taking a personal growth mindset toward generative AI (genAI). I ended it with a promise to offer a guide to adapting processes, building alignment, and pursuing organizational excellence to those in a position to lead an organization-wide genAI journey.

That’s my focus here. It’s an urgent and important job to be done, because while history isn’t exactly repeating itself right now, I see it rhyming.

Here’s what the impact of web and mobile can teach us about where the puck is going with genAI.

From retention and exposure to connection and consumption

After decades of replacing paper-based processes and records, big retailers entered the dawn of digital transformation with a lot of data generated in the course of doing business.

Customer purchase histories and store inventories, for example, were reliably retained. Employees could access them as needed to process a return or place the next month’s orders to suppliers, respectively.

What incumbent retailers could not count on, however, was a rapid round trip from the creation of that sort of information back into real-time e-commerce experiences. In the new world of Web 2.0 and mobile, customer experiences lacking (for example) personalization or suggested substitutes for out-stock items quickly became unsatisfying and uncompetitive.

Processes and tools built for retention and exposure were no longer fit for purpose. New ones needed to drive connection and consumption of what now needed to be managed not simply as business records but also as digital assets that were critical to improving every interaction.

Something similar now applies to every company, regardless of industry, now that we’re at the dawn of the age of genAI.

In employee workflows, leveraging knowledge assets using genAI solves for “I wish we’d known” missed opportunities that lead to disappointment (and frustration). We can get an intuitive handle on this from an insight by organizational learning pioneer Chris Argyris.

He defined error in the  business context as a gap between intended and actual outcomes. It’s both brilliant and actionable, because, in business, we rarely do things out of idle curiosity. From preventing regretted attrition to juicing back-to-school sales, we’ve got metrics or other success criteria in mind.

And if we reflect on postmortems on occasions when there was indeed a gap between what was intended and what was achieved, it points us toward knowledge as both the prophylaxis and remedy. Those conversations likely included statements like these:

“We drew the wrong conclusion from…”

“We never even imagined…”

“If only we’d known…”

GenAI enables conversational access to any codified knowledge, and it makes possible agentic systems that can do work on people’s behalf. These two capabilities make it perfect for incorporating into workflows that would otherwise become unsatisfying and even uncompetitive.

An HR example

Let’s consider the potential for helping prevent regretted attrition if knowledge assets are accessible to retrieval-augmented generation (RAG) and agentic genAI apps.

An HR business partner (HRBP) could get a weekly comparison of trending topics and sentiment comparing internal channels like Slack and email with external sources such as Glassdoor and LinkedIn. Levels and trends in the latter might be benchmarked against top competitors.

The HRBP and each of his client people managers could get a weekly diagnosis of internal comms sentiment among high performers, accompanied by an analysis of how they’ve spent their time in the last week (e.g., percent of time in meetings) and positive and negative events (e.g., a feature shipped versus a feature timeline that slipped).

Each high performer might have a customized plan taking into account internal and external sources that are updated on an ongoing basis. For example: One employee might write frequently on Substack about work-related topics, while also posting Medium about a hobby. GenAI could note if the tenor of the former turns negative or if the frequency of the latter declines coincident with this employee taking an unusual number of sick days.

An “High Performer Risk Synthesizer Analyst” agent could put all these pieces together into briefings for an “HRBP Attention Assistant” agent, a “Manager Attention Assistant,” and an “HR-Manager Coordinator”—the latter providing informed recommendations with rich context that the HRBP and manager could discuss with it about where they might best devote their (finite human) attention to as they wrap up the week or begin the next.

For indications of low risk, that action might entail approving a pre-drafted check-in email. For indications of high risk, it might entail agreeing to a 1:1 meeting for which time has already been found on the employee’s calendar.

The cognitive value chain

This example of “agentic flows” of previously latent or hard-to-assemble knowledge is well within the power of genAI technology as it stands today. Rallying around changing behaviors and putting the right tools in place is the next job to be done.

The result will be a cousin and complement to the modern digital value chain that we call the new cognitive value chain.

I’ll cover the components and competencies required for the latter in an upcoming article.

Learn more about DataStax.

About Bryan Kirschner:
Bryan is Vice President, Strategy at DataStax. For more than 20 years he has helped large organizations build and execute strategy when they are seeking new ways forward and a future materially different from their past. He specializes in removing fear, uncertainty, and doubt from strategic decision-making through empirical data and market sensing.