Agentic AI
Opportunities for large scale transformation using Large Action Models
Introduction
Agentic AI is a class of Artificial Intelligence systems that can act autonomously with limited or no human intervention. Agentic AI systems can not only understand and analyze human commands or interactions, but also carry out actions that previously required human involvement. The key characteristic that puts Agentic AI way ahead of its predecessors is its ability to ‘carry out actions’. While predecessors like Machine Learning / Deep Learning systems and even GenAI were only limited to suggesting actions and were not capable of executing actions. Agentic AI’s capability to execute actions both in the digital world and even in the physical world, gives them far more potential for transforming many Industries.
What is Agentic AI?
Agentic AI uses a new class of deep learning models called Large Action Models (LAM) to, autonomously, set goals, plan and take actions to achieve those goals. Agentic AI differs from its predecessors, in that, it uses a more iterative and multi-step approach to accomplishing tasks using LAMs. Compare this with GenAI, which takes a prompt to generate a single and direct response; and it becomes clear why Agentic AI can make its predecessors seem primitive.
LAMs do not stop with comprehension, but instead leverage their understanding of language to interpret user intent and translate it to one or more goals which are then mapped to concrete steps and tasks, that are finally executed to achieve those goals. Let us take an example to understand how LAMs work their magic. Imagine that an Agentic AI system called TravelAgentic has been developed to help users with their vacation planning and booking. Users can open TravelAgentic on their phone and type or say “I need help with planning and booking my vacation”. TravelAgentic would ask the Users some questions to understand their preferences for dates, travel, stay, food and activities and then iteratively prepares a vacation plan after taking feedback from the User. Once the User is happy with the plan, TravelAgentic would identify relevant travel websites or apps, navigate their interfaces and finally, complete the bookings.
Gigagentic Opportunity in Vertical AI
Now, let’s see how Agentic AI can work its magic within an enterprise and improve employee productivity. Just as our fictitious TravelAgentic could plan and book a vacation for its users, so can an Agentic AI plan and execute enterprise workflows for employees within an enterprise. Imagine you are a Sales Manager selling networking equipment to an IT manager of a midsized enterprise; and you are using SalesAgentic, an Agentic AI system to improve your sales productivity:
You ask SalesAgentic to help you with a proposal to the prospect.
SalesAgentic can analyze the RFP, if one exists, and access your CRM’s API and automatically generate personalized proposals based on pre-defined templates.
Once the proposal is sent, SalesAgentic can schedule automated follow-up emails with nudges to remind the client about the proposal and its value proposition.
It can also track email opens and clicks, and even analyze the client's online behaviour to identify their level of interest.
It can also prepare and send out revised proposals and contracts with e-signing capabilities for faster approvals.
Upon contract approval, SalesAgentic can automatically generate purchase orders for your fulfillment team.
It can integrate with your inventory management system to track stock levels and prevent overselling; and if an item is on backorder, SalesAgentic can proactively communicate with the client and suggest alternatives.
It can send pre-payment reminders before the due date and offer multiple secure payment options.
Once payment is received, SalesAgentic can automatically generate invoices with all the necessary details.
It can track outstanding invoices and send gentle reminders to chase late payments.
It can gather customer feedback through surveys to gauge satisfaction and identify areas for improvement.
It can prepare detailed reports on sales performance, including win rates, deal size, and sales cycle length; and by analyzing this data, SalesAgentic can suggest targeted strategies to optimize your sales funnel and maximize revenue.
As you can see, unlike traditional sales automation tools, SalesAgentic can operate with a high degree of autonomy. It can learn from your sales interactions and adapt its approach over time, continuously improving the efficiency and effectiveness of your sales process. By working with it, you can transform your sales team's productivity and unlock new levels of growth for your business.
The Future of Work with Agentic AI
Agentic AI is a big leap forward in AI. It can not only understand human intent but also execute actions on behalf of humans to fulfil that intent. Agentic AI systems can automate complex workflows and free up employee time for more strategic work; and can lead to large scale transformation across Industries. Most Vertical AI solutions today are co-pilot based solutions: they use GenAI to generate responses for requests or prompts from users. By bringing in the capability to execute actions, Agentic AI can enable Vertical AI solutions to unlock their full potential and revolutionize industry-specific processes across functions such as finance, HR, and operations. Building on top of Agentic AI, can make a compelling case for adoption of Vertical AI solutions.