AI Agents

• Introduction
• Chapter 1 What Are AI Agents and Why Should Managers Care?
• Chapter 2 The Evolution of AI: From Automation to Autonomy
• Chapter 3 Core Components: Understanding LLMs, APIs, and Tools
• Chapter 4 Identifying High-Impact Use Cases for AI Agents in Your Business
• Chapter 5 Building the Business Case: Calculating ROI and Securing Buy-In
• Chapter 6 Risk Management: Navigating the Ethical, Security, and Compliance Landscape
• Chapter 7 Build, Buy, or Partner: Choosing Your AI Agent Strategy
• Chapter 8 Selecting the Right Platform and Technology Stack
• Chapter 9 Assembling Your A-Team: The Skills Needed for AI Agent Implementation
• Chapter 10 Your First Project: A Step-by-Step Guide to a Successful Pilot
• Chapter 11 Data is Fuel: Preparing Your Organization's Data for AI Agents
• Chapter 12 Designing Effective Prompts and Workflows
• Chapter 13 Integrating AI Agents with Your Existing Systems and Processes
• Chapter 14 The Human-in-the-Loop: Fostering Collaboration Between Employees and AI
• Chapter 15 Change Management: Preparing Your Team for an AI-Assisted Workplace
• Chapter 16 Measuring Success: Key Performance Indicators for AI Agents
• Chapter 17 Scaling Up: From a Single Agent to an Autonomous Workforce
• Chapter 18 Governance and Oversight: Establishing Policies for AI Agent Use
• Chapter 19 Training and Fine-Tuning Agents for Specialized Tasks
• Chapter 20 Multi-Agent Systems: Unleashing the Power of Collaborative AI
• Chapter 21 Case Studies: Real-World Examples of AI Agents Driving Business Value
• Chapter 22 The Impact on Leadership: How AI Agents Will Change Management
• Chapter 23 Beyond Efficiency: Using AI Agents to Drive Innovation and Growth
• Chapter 24 The Future of Work: Reshaping Roles and Responsibilities
• Chapter 25 Preparing for Tomorrow: The Long-Term Vision for AI in Your Organization

It’s Monday morning. Your coffee is still too hot to drink, but your inbox is already overflowing. A flashing notification reminds you that a critical progress report for the quarterly review is due by noon. Your top salesperson just sent a text message about a pricing snag with a key client, your head of logistics has flagged a potential supply chain disruption in Southeast Asia, and you have three back-to-back meetings starting in forty-five minutes. You take a tentative sip of the coffee and contemplate the sheer impossibility of being in three places at once.

This scenario, in one form or another, is the daily reality for managers everywhere. You are the central node in a complex network of information, decisions, and human dynamics. Your job is to orchestrate a vast number of moving parts, solve problems you didn’t see coming, and somehow steer the team toward its strategic goals. For decades, the solution has been to work harder, hire more people, or invest in better software—spreadsheets, project management tools, communication platforms. But what if there were another way? What if you had a new kind of team member?

Imagine an assistant who, overnight, had already analyzed the preliminary data for that quarterly report, drafted an initial summary, and highlighted three anomalies that require your attention. Imagine this same assistant had cross-referenced the client’s pricing issue with historical data and internal policies, then drafted a few potential solutions for you to review. Imagine it had also monitored the supply chain issue, summarized the latest news reports, and scheduled a brief check-in with the logistics head. This isn't about a better app; it's about a new actor. This is the promise of the AI Agent.

For many managers, the term "Artificial Intelligence" is both intriguing and intimidating, often conjuring images of complex algorithms or futuristic robots. An AI Agent, however, is a much more practical and immediate concept. Think of it not as a piece of software you use, but as an autonomous entity you delegate to. It is a system designed to perceive its digital environment, make decisions, and take actions to achieve specific goals you have set for it. In short, it’s a digital employee that can handle complex, multi-step tasks with a degree of independence.

Let’s use an analogy. A simple calculator is a tool. You tell it "five times ten," and it gives you "fifty." It is powerful but passive; it waits for your explicit command. A spreadsheet is a more advanced tool. You can build a financial model, but you still have to input the data, design the formulas, and interpret the results. An AI Agent, by contrast, is more like hiring a junior financial analyst. You can give it a goal, such as: "Monitor our department’s spending against the budget and alert me to any potential overages." The agent doesn't just wait for you to input numbers; it actively retrieves the data, analyzes it according to the rules and goals you’ve set, and then takes action—by sending you a notification or even generating a detailed report.

This distinction from a simple tool to an autonomous actor is the fundamental shift that managers need to grasp. We have spent decades creating digital tools that help us do our jobs better. We are now entering an era where we will be managing digital workers that do the jobs for us. This is not an incremental improvement; it is a categorical change in how work gets done. The agent is the difference between a hammer and a carpenter who knows how to build the house on their own.

The rise of AI Agents is not a sudden phenomenon but the result of a perfect storm of technological advancement. The concept has existed for decades in computer science labs and theoretical papers. However, only recently have three key ingredients converged to make them a practical reality for businesses. First, the development of powerful Large Language Models (LLMs) has provided agents with a sophisticated "brain." These models allow agents to understand natural human language, reason through complex requests, and generate nuanced responses.

Second is the proliferation of Application Programming Interfaces, or APIs. If LLMs are the brain, APIs are the hands and feet. They are the digital connectors that allow an AI Agent to interact with the outside world—to read your emails, access your company’s sales database, book a flight on an airline's website, or post an update to a project management tool. Without APIs, an AI Agent would be a brain in a jar, capable of thinking but unable to act. With them, it can execute tasks across the vast landscape of digital systems your business already uses.

The final ingredient is the accessibility of immense computational power via the cloud. Training and running the massive models that power AI agents requires a level of computing horsepower that was once the exclusive domain of governments and massive research institutions. Today, cloud platforms have democratized this power, allowing businesses of all sizes to tap into world-class AI capabilities without having to build their own supercomputers. This confluence of brain, body, and power is why AI agents are moving from the lab to the office right now.

Of course, for any seasoned manager, this grand pronouncement might sound suspiciously familiar. You’ve weathered the storms of "paradigm-shifting" technologies before. You’ve been told that Big Data, the Internet of Things, or blockchain would fundamentally revolutionize everything about your business. You’ve sat through countless presentations filled with buzzwords, only to see many of these trends fade into the background noise of incremental improvements. So, a healthy dose of skepticism is not just warranted; it’s wise.

The hype cycle is a real phenomenon in the tech world. A new idea emerges, followed by inflated expectations, a trough of disillusionment when it fails to solve every problem overnight, and eventually, a plateau of productivity where its real, practical value is finally understood and integrated. AI is certainly experiencing its own hype cycle, and it is easy to dismiss AI Agents as just the latest buzzword destined for the corporate jargon graveyard.

However, there is a fundamental difference this time around. Previous technological waves primarily gave us better ways to collect, store, and analyze information. They provided dashboards, reports, and insights. But after the insight was delivered, the responsibility to act still fell squarely on human shoulders. The AI Agent is different because, for the first time, the technology itself can take the next step. It can close the loop between insight and action. It doesn’t just tell you there’s a problem; it can be empowered to start solving it.

This is not a theoretical distinction; it’s a practical one that changes the nature of delegation and execution. An analytics dashboard can show you that customer complaints are spiking, but an AI Agent can be tasked with triaging those complaints, responding to common issues with automated solutions, and escalating only the most complex cases to a human support representative. That ability to act autonomously is what separates this shift from a mere upgrade in our analytical tools to the introduction of a new type of workforce.

This brings us to the purpose of this book. This is not a technical manual for data scientists or a philosophical treatise on the future of consciousness. It is a practical guide written specifically for you: the manager. It is for the person who is responsible for budgets, deadlines, and the well-being of a team. Your primary challenge isn't to understand the intricacies of neural networks, but to answer a series of very practical questions.

What can this technology actually do for my team today? How do I identify the right tasks to delegate to an AI Agent versus a human employee? How do I build a business case to justify the investment, and what does the return on that investment (ROI) even look like? What are the risks involved—from data security and privacy to the ethical implications of automated decision-making—and how do I mitigate them? This book is designed to provide clear, straightforward answers to these managerial questions.

It is structured to be a roadmap, guiding you through the journey of integrating AI Agents into your operational reality. We will cut through the jargon and focus on what you need to know to make informed decisions. The goal is to demystify the technology and empower you to lead your team through this transition with confidence. This is not about becoming a technology expert; it's about becoming an expert in managing a new, hybrid workforce of humans and intelligent machines.

The core challenge presented by AI Agents is, fittingly, a management challenge. The technology is simply a new resource. The real work lies in figuring out how to deploy that resource effectively. How do you "onboard" a digital worker? What does a performance review look like for an algorithm? How do you foster collaboration between your human staff and their new AI "teammates"? These are not technical problems; they are leadership problems.

Answering them requires a shift in mindset. For generations, management has been about directing human potential. Now, it must also include directing machine potential. This involves learning a new form of delegation, where instructions must be precise, goals must be quantifiable, and the "common sense" that you rely on with human employees must be explicitly programmed or taught. It’s a new frontier for leadership, and the managers who navigate it successfully will be the ones who thrive in the coming years.

This book will not offer a one-size-fits-all solution. The right way to leverage AI Agents in a marketing department will be different from their application in a logistics-heavy supply chain. A small business will have different needs and constraints than a multinational corporation. Instead of providing a rigid prescription, our focus will be on providing a flexible framework for thinking, planning, and executing.

We will equip you with the mental models to assess your own business processes and identify the highest-impact opportunities for agent-based automation. We will explore the strategic choice between building your own custom agents, buying an off-the-shelf solution, or partnering with a specialized vendor. Each path has its own trade-offs in terms of cost, speed, and customization, and we will help you determine which is right for your specific context.

This guide will walk you through the entire lifecycle of an AI Agent initiative. We will begin with the fundamentals, establishing a clear understanding of what agents are and how they evolved from simple automation to autonomous systems. This foundation is critical for communicating effectively with technical teams, vendors, and your own leadership. You don’t need to be a coder, but you do need to speak the language of possibility.

From there, we will move into the strategic phase. Chapters Four through Nine are dedicated to the critical upfront work that determines success or failure. This includes identifying high-impact use cases, building a robust business case, navigating the complex landscape of risks, choosing your overall strategy, selecting the right technology, and assembling the necessary team. Rushing through this phase is the most common mistake organizations make.

Next, we will dive into the practicalities of implementation. Chapters Ten through Fifteen cover the nuts and bolts of getting your first AI Agent off the ground. We will provide a step-by-step guide to running a successful pilot project, preparing your organization's data (the fuel for any AI), designing effective prompts and workflows, and integrating the agents with your existing systems. We’ll also tackle the crucial human element: fostering collaboration and managing the change within your team.

Finally, we’ll look toward the future. Once you have a successful pilot, how do you scale up from a single agent to a coordinated workforce of them? The later chapters of the book address the long-term considerations of governance, advanced training, multi-agent systems, and the profound impact this technology will have on the nature of leadership itself. We will explore how agents can be used not just for efficiency, but as engines for innovation and growth.

Throughout this journey, our focus will remain relentlessly practical. Theory is useful, but results are what matter. This book is grounded in real-world applications and the emerging best practices from companies that are already on this path. We will translate abstract concepts into tangible examples that you can apply directly to your own operational context. According to a recent G2 report, nearly 60% of companies already have AI agents in production, demonstrating a rapid move from testing to scaling.

Imagine an AI Agent in your customer service department that can handle 70% of inbound queries, freeing up your human agents to focus on the most complex and emotionally charged customer issues. This not only cuts costs but can also dramatically improve customer satisfaction by providing instant responses to common questions, 24/7. Your human team, relieved of monotonous, repetitive work, can then provide a higher level of service where it counts most.

Consider the implications for a sales team. An AI Agent could be tasked with lead qualification, automatically researching potential clients, scoring them based on predefined criteria, and even initiating the first outreach with a personalized email. Your highly-paid salespeople would then spend their time on what they do best: building relationships and closing deals with well-qualified, engaged prospects, a process that can lead to significant gains in speed-to-market.

In the realm of project management, an agent could act as a tireless project coordinator. It could monitor deadlines, track task completion across different software platforms, nudge team members for updates, and automatically generate daily progress reports. This would free up the human project manager to focus on strategic problem-solving, stakeholder management, and removing roadblocks, rather than chasing down status updates.

Or think about internal operations in a department like Human Resources. An agent could manage the initial stages of the hiring process, from screening resumes to scheduling interviews. It could also handle the onboarding process for new employees, ensuring all their paperwork is complete, their IT access is granted, and their initial training modules are assigned. This creates a more efficient and consistent experience while freeing the HR team to focus on culture and employee development.

These are not futuristic fantasies; they are applications being built and deployed today. Businesses are using AI agents to streamline everything from financial analysis to software development. The technology is maturing at an astonishing pace, and the barrier to entry is dropping just as quickly. The question for managers is no longer if this technology will impact their business, but when and how.

This book is designed to be your trusted advisor through this transformation. The tone is intended to be straightforward and engaging, without the sermonizing or proselytizing that often accompanies discussions of new technology. The facts will be presented plainly. Where there are risks and downsides, they will be addressed head-on. Where there is uncertainty, it will be acknowledged.

There will be a touch of humor where appropriate, because navigating profound technological shifts can be a bewildering experience, and sometimes you just have to laugh. The goal is to create a resource that is not only informative but also accessible and, hopefully, enjoyable to read. We will avoid getting bogged down in overly technical details while providing enough depth for you to make intelligent, well-informed decisions.

The perspective is always that of the manager on the ground. Your reality is one of limited resources, competing priorities, and the constant need to deliver measurable results. This guide is written with that reality in mind. Every concept, every framework, and every piece of advice is intended to be directly applicable to the challenges and opportunities you face every day.

We will not be preaching about the dawn of a new era or making grand, unsupported claims about the future. Instead, we will focus on the tangible present and the achievable near-future. The aim is to give you a solid, pragmatic foundation for action. You will come away from this book not as a programmer, but as a leader equipped to strategically deploy one of the most powerful new tools available to business.

This transition will not be without its challenges. It will require new skills, new ways of thinking, and a willingness to experiment and learn. Like the introduction of the personal computer or the commercial internet, the rise of AI Agents will reshape workflows, redefine job roles, and alter the very texture of daily work. Preparing for this change is not just an IT issue; it is a core strategic imperative for any forward-thinking manager.

Some jobs will be automated, but many more will be augmented. History has shown that technology tends to create new roles even as it makes old ones obsolete. The key is to proactively manage this transition, helping your team members adapt their skills to collaborate with these new digital colleagues. The manager's role will become even more critical, shifting from task supervision to the orchestration of a hybrid human-AI workforce.

The opportunity before you is immense. By thoughtfully integrating AI Agents, you can unlock new levels of productivity and efficiency within your team. You can free your most valuable asset—your people—from the drudgery of repetitive tasks and empower them to focus on the creative, strategic, and deeply human work that machines cannot replicate. This is the true promise of this technology.

It’s about more than just cost savings or efficiency gains, though those are significant benefits. It’s about elevating the nature of work itself. It’s about creating an environment where technology handles the routine, and humans handle the exceptional. It’s about giving yourself and your team the leverage to accomplish more than you ever thought possible.

This book is your first step on that journey. It is a guide for managers who are ready to move beyond the headlines and start building the future of their organizations. The world of AI Agents is here, and it is time to get to work. Let’s begin.

Your team just wrapped up a major product launch. It was a classic "all hands on deck" effort, culminating in a frantic, caffeine-fueled final week. Now, the dust is settling, and a different kind of work begins. You need to compile a post-launch report for leadership, which means gathering performance metrics from the sales dashboard, sentiment analysis from the marketing team's social media tools, customer support ticket volume from the CRM, and bug reports from the engineering team's tracking software. Then, of course, you need to synthesize it all into a coherent narrative. It's a task that is both critically important and mind-numbingly tedious.

You know the drill. You’ll spend the next two days chasing people for data, copying and pasting numbers between spreadsheets, and trying to align conflicting reports. It’s the kind of administrative black hole that consumes a manager’s most valuable resource: time. Now, imagine a different scenario. You open a chat window and type: "Compile a post-launch report for Project Nova. Include sales data, customer sentiment from Twitter, support ticket trends, and all critical bug reports. I want a summary of key takeaways and a draft presentation by tomorrow morning."

You've just delegated that entire multi-day, multi-system scavenger hunt not to a person, but to an AI Agent. The agent doesn't need to ask you where to find the data; it already has access to the relevant systems. It doesn't need you to clarify the format; it understands the conventions of a business report. It simply gets to work, perceiving its digital environment, executing a plan, and acting to achieve the goal you’ve set. This is the fundamental difference between the tools you’ve been using and the agents you will be managing.

Defining Your New Digital Colleague

For years, managers have been told that every new piece of software will "revolutionize" their workflow. In most cases, these tools are just better hammers—more powerful, perhaps, or easier to use, but they still require a human to swing them. An AI Agent is not a better hammer. It is a digital carpenter to whom you can describe the birdhouse you want built. An AI agent is a software system that uses artificial intelligence to pursue goals and complete tasks on behalf of a user, with a degree of autonomy to make decisions, learn, and adapt.

To grasp what makes an agent different, it’s helpful to break down its core characteristics. These aren't just technical features; they are the traits that define its role as a new kind of worker within your team.

First and foremost is autonomy. An AI agent can operate and make decisions independently to achieve a goal. Unlike a simple automation script that follows a rigid, predefined set of steps, an agent can be given a high-level objective and then figure out the steps to get there. This is the crucial distinction. You don't tell a sales agent, "Click on the CRM, then open the contacts tab, then filter for new leads, then compose an email." You say, "Follow up with all new leads from the conference." The agent understands the intent and orchestrates the necessary actions across different applications on its own. It doesn't require constant human input for every step of the process.

Second is goal-directed behavior. Traditional software is instruction-based; an agent is objective-based. You provide the "what," and the agent determines the "how." For instance, you could task an agent with the goal of "reducing our team's response time to customer inquiries by 15%." The agent could then analyze existing workflows, identify bottlenecks, and even propose and implement solutions, such as automatically routing tickets or drafting responses for common issues. This ability to break down a complex goal into specific tasks and subtasks is a defining feature.

Third, agents are perceptive and interactive. They are designed to monitor their digital environment and act within it. If an LLM is the agent's "brain," then APIs (Application Programming Interfaces) are its senses and limbs. An agent uses APIs to "see" new data in a spreadsheet, "read" incoming emails, and "operate" other software applications. This allows it to gather information and execute tasks in the same digital ecosystem your human team uses every day.

Finally, advanced agents possess the ability to learn and adapt. Through a process of feedback and reflection, an agent can improve its performance over time. It can remember past interactions, learn user preferences, and refine its approach based on the success or failure of its previous actions. This is not just about following a script more efficiently; it's about getting smarter and more effective with experience, much like a human employee.

The Agent vs. The Bot vs. The Dashboard

The corporate world is already awash with technology that promises automation and intelligence. As a manager, you’re likely familiar with automation scripts, chatbots, and analytics dashboards. It's crucial to understand how AI agents differ, as this clarity will inform where and how you can deploy them for maximum impact.

Automation Scripts and RPA: Robotic Process Automation (RPA) has been a valuable tool for automating repetitive, rule-based tasks for years. An RPA bot is excellent at mimicking human keystrokes to do things like copy data from one system to another. However, it operates on a strict, predefined script. If the user interface changes or an unexpected error pops up, the bot typically fails. AI automation is the digital equivalent of an industrial conveyor belt; it follows predetermined rules with precision. An AI agent, by contrast, is more resilient and adaptable. Faced with an unexpected change, it can use its reasoning capabilities to problem-solve and find a new path to its goal, much like a human would.

Chatbots: Most people have interacted with a chatbot for customer service. These are typically designed to simulate a conversation and answer common questions based on a predefined script or knowledge base. While useful, they are largely reactive and limited in scope. An AI agent is a significant step up. It can understand context across multiple conversations, integrate with various systems to perform actions (like actually booking a flight, not just telling you the flight times), and handle far more complex, multi-step tasks. A chatbot is a conversationalist; an AI agent is a doer.

Analytics Dashboards: Business intelligence (BI) tools and dashboards are fantastic for visualizing data and uncovering insights. They can tell you that sales in a certain region are down. They might even help you diagnose why by showing you related data. But that’s where their job ends. The responsibility to decide what to do next and then execute that decision falls entirely on you. An AI agent is different because it can close the loop between insight and action. An agent tasked with monitoring regional sales could not only flag the downturn but also be empowered to take initial steps, such as scheduling a meeting with the regional sales manager, pulling relevant reports for that meeting, and drafting an email to the team summarizing the issue.

Think of it as a hierarchy of autonomy. A dashboard provides information. A script follows rigid instructions. A chatbot conducts a structured conversation. An AI agent, however, is given an objective and is trusted to operate independently to achieve it.

Why Should You, the Manager, Care?

Understanding the definition of an AI agent is one thing; understanding its practical relevance to your daily challenges is another. The rise of AI agents is not just another technological trend to monitor; it is a fundamental shift in the resources available to you for getting work done. For managers, the implications are profound and can be broken down into four key areas.

1. A Quantum Leap in Productivity and Efficiency

Every manager is tasked with doing more with less. AI agents offer a path to efficiency that goes far beyond simple automation. By handling complex, multi-step workflows, they can liberate your team from the high-volume, low-judgment work that consumes a disproportionate amount of their time and energy. Repetitive tasks like data entry, report generation, and scheduling can be delegated, allowing your human employees to focus their efforts on more strategic responsibilities.

The impact is measurable. Research from institutions like MIT has shown that teams augmented with AI agents can see dramatic productivity boosts—in some cases as high as 60%—without sacrificing the quality of the work. For example, support agents using AI tools can handle a significantly higher volume of customer inquiries per hour. This isn't just an incremental improvement; it’s a redefinition of what a team can accomplish in a day. It means faster deliveries, lower operational costs, and enhanced efficiency across the entire organization.

2. Augmenting Your Team, Not Just Automating Tasks

One of the most common fears surrounding AI is job replacement. A more constructive and accurate framework for managers is job augmentation. AI agents should be viewed as digital teammates designed to amplify human capabilities, not replace them. They excel at tasks that are structured, data-intensive, and repetitive—the very tasks that often lead to burnout and disengagement in human employees.

By automating this "work about work," you empower your people to focus on what they do best: strategic thinking, creative problem-solving, building relationships, and handling complex, nuanced situations that require human judgment. An AI agent can sift through thousands of sales leads to identify the top 10 prospects, but it’s the human salesperson who builds the rapport and closes the deal. The result is not just a more productive workforce, but a more engaged and fulfilled one.

3. From Data-Driven Insight to Data-Driven Action

Managers have been inundated with data for years. The challenge has never been a lack of information, but a lack of time and resources to act on it. AI agents bridge this critical gap. They serve as an "action layer" on top of your data, turning analytics into automated workflows.

An AI agent can monitor key performance indicators in real-time and trigger actions when certain thresholds are met. Imagine an agent overseeing a supply chain. It doesn’t just show you a dashboard indicating a potential delay; it can proactively search for alternative suppliers, calculate the cost implications, and present you with three viable solutions before the delay becomes a crisis. This proactive capability transforms your team from being reactive to data to being proactive and driven by it.

4. Gaining and Maintaining a Competitive Edge

In today's fast-paced business environment, speed and agility are paramount. Organizations that can respond to market changes, customer needs, and operational issues faster than their competitors will win. AI agents provide a powerful engine for this agility. They can accelerate processes that are currently labor-intensive and time-consuming, from customer support to marketing campaigns to software development.

By automating lead qualification, AI agents can help sales teams focus on the most promising prospects, accelerating deal closures. In customer service, providing 24/7 instant responses to common queries dramatically improves the customer experience. Early adoption of this technology offers a critical advantage, enabling businesses to scale faster, innovate more rapidly, and operate with a level of efficiency that is difficult to match with human labor alone.

A Note of Pragmatism

While the potential of AI agents is immense, it's essential for managers to approach this new frontier with a healthy dose of realism. These are not magical black boxes that will solve every problem overnight. As with any new employee—human or digital—they require clear direction, proper oversight, and a well-defined role.

Deploying AI agents effectively is not just a technical challenge; it is fundamentally a management challenge. It requires you to rethink workflows, manage the integration of human and AI labor, and establish new systems for governance and quality control. The initial learning curve can be steep, and not every task is suitable for an agent. A poorly defined goal or a lack of access to clean data can lead to poor outcomes. However, the organizations that are willing to invest in redesigning how work gets done will be the ones to thrive.

The purpose of this chapter was to establish a clear answer to two questions: "What is this?" and "Why should I care?" An AI Agent is an autonomous, goal-directed digital worker that can perceive and act within your business's software ecosystem. You should care because this technology represents one of the most significant levers you will have in the coming years to boost your team's efficiency, empower your employees, and accelerate your operations. The following chapters will provide the practical roadmap for how to do it.