Dual Dynamics: Harnessing and Adapting to Agentic AI for Tomorrow's World

Leadership in the Loop Volume 3

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

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

Our first two blogs in this series explored the importance of strategic intent and focus on workflows, leading us naturally to explore agentic AI, a rapidly evolving new capability in the world of AI. Recently, Andrew Ng of AI Fund talked about what's next for agentic workflows.

Let’s define agency and Agentic AI briefly. In the world of AI, “agency” refers to a system’s capacity to make autonomous decisions and take actions, based on programming and learned experiences. Agency is at the heart of agentic AI. OpenAI researchers define agentic AI systems as those that can pursue complex goals with limited direct supervision. They go beyond content generation, and have the ability to:

• Execute goal-oriented autonomous workflows across systems with minimal human supervision.
• Exercise human-like judgment and decision-making.
• Make dynamic goal and plan adjustments based on changing conditions.

In the physical world, we have good examples. Tesla and Waymo are already deploying autonomous vehicles that can make real-time driving decisions. AI tools are used for performing robotic surgeries. AI is already used widely in high-frequency trading by financial institutions to make fast, efficient trading decisions based on market data analysis.

Since we’re on LinkedIn, let’s illustrate the capabilities of agentic AI with a simpler example. Consider this: normally writing a research paper and posting a summary on LinkedIn could take weeks, or at least a few days. Now, imagine an agentic AI accomplishing all these steps in just a few minutes and posting it on LinkedIn upon your approval or meeting certain parameters.

Recent developments in agentic AI have been catching the attention of Enterprise AI watchers and business leaders. Google recently announced the launch of Vertex, and many speculate that OpenAI has agentic AI enhancements coming soon. It’s still early days for agentic AI. However, in just weeks and months, agentic AI will feature increasingly in routine business functions and workflows.

Inside-Out: Focus on Your Organization’s Productivity

Agentic AI will undoubtedly bring about a massive productivity boost within your organization. As you begin experimenting, clear strategic intent and a focus on workflow are indeed critical, as we have pointed out before.

Some of the questions you might ask are:

• “How can we reimagine select workflows within my enterprise functions and operations, to be run by autonomous or semi-autonomous agents?”
• “How can we execute complex transactions at scale with AI agents, including planning, review, judgment and trade-off decision making?”
• “How and where should we involve humans in the loop and which humans?”
• “How can we create multiple AI agents that work with each other to drive even more productivity and still ensure responsible AI (RAI)?”

Mirror Mirror on the Wall: Who’s the Fairest AI Agent of Them All?

However, there’s another dimension to this.

If we look beyond your enterprise into the value chain your company is part of, we realize that it’s not only your company that is planning to implement agentic AI. More than likely, your suppliers and customers are also experimenting with agentic AI.

Which brings us to ask a fundamental question: What do our core business functions (product development, marketing, procurement, finance, sales, customer support) look like in an agentic AI world? How does marketing for example, need to adjust, when I am now also marketing to AI agents?

Therefore, the question is not only, “Do you have AI agents ready to work within your enterprise?”, but also, “Is my organization, its technology and data ready for agentic AI driven customers, suppliers and partners?”

Outside-in: Thrive in an Agentic AI Value Chain

Let’s take an example: Imagine you're the VP of Marketing or CEO of a B2B company that specializes in high-quality, eco-friendly office supplies. Your company prides itself on sustainable practices and has a solid customer base. However, you've noticed a shift in the marketplace. One of your key accounts, a large corporation committed to sustainability, has recently deployed an AI purchasing agent named "EcoProcure." EcoProcure is tasked with finding the most eco-friendly, cost-effective office supplies. It analyzes vast amounts of product data, reviews, and sustainability credentials to make purchasing decisions. Your company's products are a perfect match for EcoProcure's criteria, but you've noticed that orders from this account have started to dwindle. You realize that while your products are superior, the way they're presented online doesn't cater to the analytical nature of AI agents like EcoProcure.

AI agents may have different information needs, decision-making criteria, and communication preferences compared to human buyers. To address this, as the VP of Marketing, you need to make changes to your systems and processes to be ready for your customers’ agentic AI systems and processes.

• Develop Buyer Personas and Optimize Product Listings for agentic AI: Create detailed buyer personas for AI agents, focusing on their goals, pain points, and preferred content. Overhaul your product listings to ensure all information is comprehensive, accurate, and formatted for AI comprehension, including detailed specifications and sustainability credentials.
• Adapt SEO and Content Strategy for agentic AI-Driven Searches to cater to AI-driven searches. Structure your data and use relevant keywords that AI systems prioritize. Align with emerging standards like Google's Search Generative Experience (SGE).
• Build Engagement and Strategic Partnerships with AI Agents. Develop strategies to engage directly with AI agents by ensuring your content and product updates are prominently featured on platforms and services they frequently use. Establish partnerships with these platforms to enhance your visibility. This ensures your marketing efforts are aligned with the environments where AI agents operate.
• Adapt your Training Methods and Sales Tools on how to interact with AI agents and understand their decision-making processes and information needs. Equip them with AI-friendly sales tools, such as interactive demos and self-service proposal generators.
• Establish Mechanisms for Feedback and Performance Metrics for measuring the effectiveness of your strategies, continuous improvement, A/B testing refinements and adaptation based on insights from AI agents and their human operators.

Similar transformations will be required in other business functions as well, and as you move from agentic AI experimentation to at-scale roll outs, you will need a thorough understanding of your AI readiness, i.e., leadership, capabilities, operating model and technologies. (We will explore AI readiness in a future blog.)

This dual nature of the agentic AI opportunity brings the focus back to preparing with strategic intent and a focus on workflow. As you experiment, evaluate and learn making agentic AI work, consider the bi-directional approach. Think “inside-out” and “outside-in”, i.e., take your entire value chain into scope. Such an approach can help you not only reap the productivity benefits within your organization with agentic AI, but also ensure readiness for agentic AI-driven customers, suppliers and partners.

Who knows, AI agents may require us to rethink or adjust some of favorite business frameworks: Product-Market Fit; 4Ps, 7Cs, etc.

#AI #GenAI #AIAgents #Workflow #LeadershipintheLoop

?