Enterprise AI Deployment: It's All About the Workflow

Amir Hartman , Managing Director,?Dasteel Consulting?| Director AI Strategy Research Experience Alliance , Fidere.ai, Praxis AI

Venkataraman Lakshminarayanan |?Board Member, CRON AI, Former ServiceNow Value Leader

In our last blog, we discussed leadership’s AI strategic intent, and how organizations seem to be deploying enterprise AI roadmaps in several ways ranging from a “Decorator”, “Remodeler”, “Builder”, “Architect” to “Syndicator” profile. In this article, we turn our attention from strategy to execution.

From Strategy to Execution… to “POC Paralysis”

We pointed to the POC paralysis (the inability to move from experimentation to enterprise scale) many companies we studied seem to face with GenAI. Part of the problem seems to be the concerns and uncertainty leaders have. Several studies, including our study on enterprise AI initiatives with the Experience Alliance have shown that these concerns are still top of mind for enterprise leaders.

This recent article in the Harvard Business Review explains why adopting GenAI may be so difficult. Even if you resolve the gap between AI capabilities and the use cases, the article points out that GenAI by itself may not be sufficient to realize its potential. The State of Generative AI, 2024 brings both optimism and caution together - noting the security concerns that are stalling projects, and a looming AI hype plateau.

With the current sense of cautious optimism prevailing in the world of AI, it is no surprise that many companies find themselves experiencing POC paralysis. Enterprises may decide to experiment and evaluate rather than fully adopt. Many leaders are faced with a dilemma: On the one hand, experiment and evaluate, and on the other, show early results of outcomes that matter. How might they solve for this?

Spray and Pray vs. Workflow Execution

Let’s take a moment to understand a couple of different ways GenAI is often ‘implemented’ in an enterprise:

· Spray and Pray: GenAI out of the box (e.g., ChatGPT, Gemini, Copilot) is easily accessible, fun and helps employees and enterprises explore its potential. While it offers AI-enabled content generation and concision, it may be harder to translate that to bottom line savings. With extensive prompt engineering, Retrieval Augmented Generation and predictive AI techniques, specific problems can be solved more effectively, but enterprise-scale productivity may still prove elusive.

· Targeted Workflow: Moving a level up in the AI integration and capability ladder, custom-built applications can combine Predictive AI and Generative AI, integrated with other applications into a digitized workflow. Since the solution fits in well into an existing functional workflow, it is easier to drive productivity and actions with this approach. When this is aligned to business goals and combined with the redesign of the business process, enterprises are better positioned to take advantage of the full potential of GenAI – and predictive AI.

What seems to be common amongst those companies that are seeing meaningful results from AI and GenAI, is their focus on integrating them into the day-to-day operations and decision-making processes of their organization, rather than being treated as a standalone solution. By embedding GenAI into workflows, companies can unlock the power of AI.

The targeted workflow approach is an important aspect of preparing for what is emerging. Just listen to Andrew Ng talk about how Agentic AI systems will drive massive progress, by planning, executing, reviewing, and acting autonomously, making decisions and performing tasks with little or no human intervention.

In other words, by embedding AI into workflows, companies can show meaningful results from POCs. This remains true as they scale their rollouts. As companies translate their strategic intent and adopt the profiles that are right for them (“Decorator”, “Remodeler”, “Builder”, “Architect” to “Syndicator”), we encourage leaders to make sure execution is grounded in embedding AI into the workflow of the organization, so that these AI experiments can be part of your organization’s DNA.

Tell us about the success or challenges you are experiencing in your AI journey.

#AI #GenAI #AIAgents #Workflow #LeadershipintheLoop

From the Leadership in the Loop Team