top of page
Writer's pictureSharon Gai

The Agentic Future and How it will Change Work



The idea of integrating a bunch of AI agents to complete our work excites one group of people and stirs anxiety in another. While those of us who have been in an AI bubble have largely been brainwashed gradually to accept that we will co-exist working alongside agents, those of us who have not been married to this idea are still on the fence.


It's completely natural for people to resist change. Just like how a disruption in my morning routine or not having my coffee the way I like it can be irritating, larger changes—like the global shift toward AI—can feel overwhelming. In many ways, we're all going through a massive change management process when it comes to adopting AI. Change management experts understand that one of the most critical elements is providing support—having a guide to help navigate the transition. The challenge with change is that it often feels unsettling, especially when you're right in the middle of it.


What is an agent?


The next big change is the advent of agents. We all know that agents will become more popular in the future, but what is an agent? And how does it differ from a regular AI tool?


By definition (and I’ve also heard multiple by now) an agent is a software entity that uses artificial intelligence to perform tasks or make decisions on behalf of a user or system. It operates autonomously, perceiving its environment through sensors or data inputs, processing that information, and then taking actions to achieve specific goals. It’s one step beyond a simple chatbot or a “dumber” AI system because it has a higher level of intelligence and can act autonomously.


What is the difference between a chatbot and an agent?


The terms chatbot and agent are often used interchangeably, but they refer to slightly different types of AI systems, especially in terms of functionality and purpose. Here's a breakdown of the key differences:


1. Scope of Tasks


  • Chatbot: Generally, a chatbot is designed to handle simple, rule-based interactions. It's often programmed to respond to predefined questions and follow scripted conversations. These are often used in customer service, providing basic information or performing simple tasks like answering FAQs.

  • Agent: An AI agent is more advanced, capable of performing complex tasks that often involve understanding the user's intent more deeply. Agents can manage dynamic conversations, learn over time, and integrate with other systems to perform actions beyond simple responses (e.g., booking appointments, executing transactions).


2. Intelligence


  • Chatbot: Chatbots often operate on predetermined scripts or decision trees, limiting their ability to go off-script or adapt dynamically. They rely heavily on keywords and triggers to provide answers.

  • Agent: Agents, particularly AI-powered ones, use more sophisticated technologies like natural language processing (NLP) and machine learning. This allows them to interpret broader contexts, understand user intent more accurately, and evolve as they gather more interactions.


3. Learning Capabilities


  • Chatbot: Most traditional chatbots don't learn from interactions. They provide static, repetitive responses unless updated by their developers.

  • Agent: Agents, especially when using machine learning or reinforcement learning, can improve their performance over time. They adapt based on user behavior, feedback, and data patterns, making them smarter with continued use.


4. Integration with Systems


  • Chatbot: Chatbots are often standalone or have limited integration with backend systems. For example, a chatbot on a website might answer questions but not necessarily interact with the website’s database.

  • Agent: Agents are often deeply integrated into larger systems, allowing them to pull data from various sources, perform actions, and even make recommendations or decisions based on that data. For instance, a virtual sales agent could integrate with CRM tools to recommend products based on customer history.


5. Use Cases


  • Chatbot: Basic customer service, answering FAQs, or assisting with simple queries (e.g., "What’s the weather today?").

  • Agent: More complex applications like virtual assistants (e.g., Siri, Alexa), personal productivity tools, or customer service agents that handle full service inquiries (e.g., helping with a full end-to-end process like insurance claims).


6. Conversational Flexibility


  • Chatbot: Conversations with chatbots tend to be linear and less flexible. If the conversation strays from predefined paths, the bot might struggle to respond appropriately.

  • Agent: AI agents tend to handle more natural conversations, allowing for follow-up questions, clarifications, and multi-step processes within a single interaction.



Big Tech Incorporates Agents



Dreamforce Marc Benioff Keynote


We just came out of two major tech conferences: Dreamforce held by Salesforce and Inbound held by Hubspot. Both companies emphasized agents a great deal.

Salesforce predicts that after the wave of predictive AI and Copilots, the third oncoming wave is agents. Salesforce’s big announcement was about Agentforce where customers can design their own agents for sales and marketing. Salesforce also owns Slack and showed a Slack chat of a merchandising team at Saks Fifth Avenue that depicted agents working alongside human merchandisers to decide on whether or not to purchase an order of dresses.


Notice the number of humans in this Slack chat and the number of agents that are also in there on the upper right hand corner.



Also in September was Inbound, the annual conference for Hubspot, a competing CRM company. The CEO of HubSpot is building a Linkedin for agents. He purports that in the future, similar how recruiters are finding workers to hire for a certain job, we will also be shopping for agent candidates to hire for a certain task. I’ve tried some of his agents myself and my feedback is there are probably a few that works well while others have inferior results. Of course, this is all a work in progress. Almost every agent website I visit is mostly a demo. It’s a cutting edge technology that still has to be given time for testing, deployment and maturity.



Crew AI


One prominent company that have risen in recent months is Crew AI, a platform designed to enhance communication, collaboration, and customer engagement through the use of AI-powered tools and agents. It focuses on optimizing team workflows and automating interactions to boost productivity and deliver better experiences for customers. By integrating AI-driven capabilities such as natural language processing, voice recognition, and chat automation, Crew AI helps businesses streamline their processes and reduce manual tasks.



Multi-on


This is a demo of the Multi-On agent that envisions itself to order food for you, buy airline tickets, and can set up cloud instances in AWS.



Artisan


Artisan is another company that hopes to sell agents to cash-strapped businesses. Artisans are advanced human-like digital workers trained to do specific roles, who integrate alongside human teams. They have unique faces, names, memories & skills, and they continuously improve once they are employed, molding to each company's needs. Their first Artisan, Ava, automates the entire outbound sales process and can be set up with a 10-minute conversation. The goal is to communicate with her in Slack to command her to complete certain actions.



 

How Agents Will Change the Way we Hire



It's worth exploring how hiring will evolve when we can outsource much of our work to AI agents. I’m not sure if HR departments around the world are deeply considering this yet, but the way we write job descriptions will need to adapt. Just as we currently require basic skills like typing or using Microsoft Word and Excel, the next essential skill might be the ability to manage AI agents. In the future, job candidates may need to demonstrate proficiency in working with agents just as they do with software systems and data manipulation today.


We may all need a refresher on communication, specifically in articulating clear instructions to AI agents. The clearer we are, the better results we'll get. If we fail to explain tasks properly, we'll likely end up with subpar outcomes. Instead of traditional final-round interviews that involve case studies, it’s possible future assessments will require candidates to complete a project using AI tools and agents.


Even for individual contributor roles, management skills will become critical. People will need to become proficient in project management, as the human role will shift toward initiating projects and coordinating their completion. For example, in my previous role, we outsourced 700 U.S.-based jobs to Serbia, where employees handled tasks like design, updating product pages, and writing social media posts. A single person in the U.S. managed the recruitment and oversight of these 700 eCommerce professionals.


Our Future Org Chart?



Looking ahead, I think we’ll see more one-person, million-dollar companies. As AI advances, what part of the value chain can’t be outsourced? Perhaps only the CEO role will remain uniquely human—the ultimate decision-maker responsible for choosing which AI agents to use and when to deploy them.


What do you think?

 

 

I’m a keynote speaker on AI and its effects on workers and society. Know an event that is need of a keynote on AI? Tell me about the event or refer me and I will thank you with a gift for partnering with me!

1 view0 comments

Recent Posts

See All

Comments

Rated 0 out of 5 stars.
No ratings yet

Add a rating
bottom of page