Agentic AI for Maximizing Workflows and Business Growth

 

---

We’ve all seen the dramatic impact of Artificial Intelligence (AI) in general, and if you’re like us, you already use tools like ChatGPT, Gemini, or Copilot. But wait, there’s more – and it’s a game changer: Agentic AI. This isn’t just another incremental update; it’s taking AI capabilities much further, creating exciting new opportunities for us to grow our skills, enhance our workflows, and leverage AI in more meaningful ways.

What Makes Agentic AI Different?

To understand Agentic AI, let’s first quickly look at where it fits in your existing AI toolbox. You’re likely familiar with:

• Standard LLMs (Large Language Models): Tools like ChatGPT and Gemini that respond to your prompts but don’t do anything until you ask.
• Custom GPTs: These are instances of LLMs that you train with your own specific information, though they can still “make stuff up” and pull in external context.
• RAG Tools (Retrieval Augmented Generation): Like Google’s NotebookLM, these limit the AI to the information you provide, making them great for tasks where you need the AI to stick strictly to your content, such as generating blog posts from webinar recordings without adding external context.

Agentic AI is the next evolution. While traditional AI tools wait for your command, Agentic AI can think and act autonomously. It takes initiative and makes decisions to achieve a goal you set for it, operating independently without you needing to prompt it for every single step.

There are four key traits that make Agentic AI truly stand out:

1. Autonomy: Once properly set up, it can operate on its own. You give it a goal, and it acts when needed to achieve that goal, saving you significant time because it executes tasks on an ongoing basis without constant prompting.
2. Goal-Driven: Instead of mapping out every single step, you give it an objective, and it researches and makes decisions to reach that outcome.
3. Proactive: It can initiate actions, rather than waiting for you to tell it what to do next.
4. Decision-Making: As it works to complete tasks, it can make its own decisions about the best way to proceed.

Think of Agentic AI as a smart executive assistant. If you tell a traditional AI, “Book me a flight,” it needs specific details. But with Agentic AI, you could theoretically say, “I need to be in New York on Thursday,” and an AI agent would then figure out all the necessary steps: booking a flight, reserving a hotel, adding it to your calendar, and even arranging rideshare transport.

A note of caution here: While this sounds amazing, it’s crucial to understand that such a task requires significant setup. Just like a new employee, an AI agent needs training and clear guidelines on your preferences (e.g., hotel type, flight preferences, budget). It also needs to be integrated with your specific systems (like your hotel booking site or calendar). This is where the power—and the work—come in.

Practical Applications with AI Agents: Real-World Examples

We’re seeing two main types of Agentic AI today:

• Task-Based AI Agents: These are cognitive agents that can think and act for you, and you can train them with parameters. Automation agents also fall into this category, focusing on connecting tools and systems to complete tasks independently across them.
• Specialized Agents: These are more emerging and are almost like pre-trained instances for specific functions.

Let’s dive into some of the tools we use and how they work in action:

1. Zapier’s AI Agents: Automating Cross-System Workflows

Zapier is an automation platform that connects systems that don’t natively talk to each other. Its new AI agents add an autonomous layer, completing tasks that require connections across multiple systems. We use Zapier constantly to get our systems to communicate, from webinar registrations flowing into our CRM and then to our learning management system. The AI agents now take this a step further.

Here are some ways we’re leveraging Zapier’s AI agents:

• Automated Blog Post Creation: We set up an AI agent to trigger whenever a new audio transcript appears in a specific Zoom account. The agent then:
  • Finds the meeting transcript.
  • Uses ChatGPT (integrated via Zapier) to determine three unique blog post topics based on the transcript content.
  • Creates each blog post, applying our brand tone of voice and blog writing style guides (which we uploaded as resources). This is important: we spent time creating an AI tone of voice guide that can be uploaded to any AI tool to ensure consistency.
  • Creates a draft of each blog post directly in WordPress. We review and publish them manually, but the automation saves us tremendous time. This process used to be entirely manual, from downloading transcripts to uploading content.
• Sales Call Feedback: After a sales call in Zoom, the agent reviews the transcript and provides specific feedback on the salesperson’s performance, then emails a summary to them. This goes beyond basic Zoom summaries, offering actionable insights.
• Automated Lead Scoring and LinkedIn Connections: When a new contact is added to our CRM (Keep), the agent automatically:
  • Filters out non-corporate email addresses.
  • Finds the contact on LinkedIn, using their company name, first name, and last name, and prioritizing marketing roles.
  • Adds their information to a Google Sheet.
  • Scores them based on our “hotness score” criteria, determining how well they fit our ideal customer profile. This involves the AI thinking and evaluating, not just automating data transfer.
  • Emails our sales team weekly with an organized list of new, high-quality leads.
  • Optionally, finds their LinkedIn profile, likes their posts, and invites them to connect with me. Many of you may even get an invitation from me after this webinar!

The key to using Agentic AI, and any AI for that matter, is to identify tasks you do regularly and look for ways to automate them. While setting up these automations can take 30 minutes to a couple of hours to test and perfect, the long-term time savings for repetitive tasks are immense.

Zapier’s strength lies in its ability to connect with many systems through APIs, which traditional LLMs like ChatGPT cannot do directly. While Zapier has its own LLM built in, we can also integrate our specific ChatGPT account, which is trained on our writing style, for certain tasks.

2. Relevance AI: Advanced Web Scraping and LinkedIn Integration

Relevance AI is similar to Zapier in its automation capabilities but excels in its ability to “read the web” more effectively than other tools. For example:

• SEO FAQ Generation: We use Relevance AI for SEO by giving it a Google search query. It then identifies the “People Also Ask” questions, goes to the linked source for the answer, and then writes a better, optimized answer for us, aligning with our tone of voice. This helps us outrank competitors. This specific capability—going to a source link and analyzing its content—is something other advanced AI tools struggle with.
• LinkedIn Insights: While LinkedIn’s API limitations make deep integration challenging for most tools, Relevance AI, particularly with its higher-tier packages, can do more on LinkedIn, such as scraping comments from viral posts and adding enriched data to a CRM.

A quick tip: When generating content with AI, we always recommend using an AI copy checker during the setup phase to ensure the output doesn’t sound too much like it was written by AI. This helps refine your instructions and brand tone of voice to produce more natural, human-sounding content.

3. CRA AI: Pre-Trained Agents for Specific Tasks

CRA AI operates differently from Zapier and Relevance. Instead of building tasks from scratch, CRA AI offers “brain AI” where you add information about your business, and it provides pre-trained AI agents (or “robots”) that are proficient at specific tasks.

For instance, we asked its “customer inquiry agent” to create FAQ responses for our All Access Pass, and it did an exceptional job, pulling relevant information like pricing questions that other tools missed. This is because CRA’s “brain” learns your business at its core, separate from individual tasks, which leads to better, more relevant outputs. While it has fewer integrations and pre-trained tasks, its out-of-the-box functionality is quite strong.

4. Manis AI: Handling Complex, Multi-Step Strategic Tasks

Manis AI stands in a different category, excelling at very complex tasks with many steps and diverse skill requirements, all at once.

Consider this: if you wanted a SWOT analysis presented as a professionally designed slideshow, with clear layouts and visual elements, in ChatGPT, you’d have to prompt it for the analysis first, then for the formatting, and possibly more steps. Manis AI can execute all these actions as a single, complex task.

Here are examples of how Manis excels:

• Comprehensive Analyses and Action Plans: We asked Manis to analyze the digital marketing education space and provide trends, even generating an interactive website with the findings. We also had it perform a full audit of our website (including SEO and technical aspects) and generate an action plan based on the audit.
• Strategic Deliverables: It can create things like go-to-market strategies or even custom icon sets based on your branding.

Manis is particularly useful when you have multi-step tasks that would typically “break the brain” of traditional LLMs if you asked them to do everything at once. It streamlines strategic work that usually requires numerous separate prompts and much manual effort.

Building Your Own AI Agents: Our Process

The most important takeaway for creating effective AI agents is to approach it systematically:

1. Identify Repetitive Tasks: Look for things you do daily, weekly, or monthly that are time-consuming or manual.
2. Define the Outcome: Be very clear about what you want the AI agent to achieve.
3. Train and Refine: This is critical. For instance, when setting up an agent in Zapier, you provide the trigger (what starts the automation) and then detailed instructions. You can add various tools (like ChatGPT, WordPress, your CRM) and upload resources (like your brand tone of voice guide or style guides).
4. Test and Optimize: It takes time—typically 30 minutes to a couple of hours per automation—to test and refine agents to ensure the outcome meets your objectives. This iterative process helps you learn what the AI is good at and where it might need more direction.

For example, when drafting meeting recaps, Zoom provides summaries, but they often lack our specific writing style. Our AI agent now looks at the transcript, applies our brand guidelines, writes key takeaways and action items in our style, and drafts it as an email. This eliminates 15 minutes of manual editing work after every meeting.

Our Boot Camp Digital Brand Tone of Voice Guide is a living document that we continuously refine. We started by asking AI itself what information it would need to write in our tone. We focus on:

• Core Attributes: Practical, direct, specific, action-oriented, trust-building, educational, and no hype.
• Language: Plainspoken, confident, conversational (not casual), and aimed at a 6th-grade reading level. We use simple words like “use” instead of “utilize”.
• Things to Avoid: Vague buzzwords (“supercharge”), overly academic language, over-promising claims, and emojis. We prefer tangible results and proof-driven statements.
• Structure: Using frameworks, bullet points, short paragraphs (2-3 sentences), short sentences (5-15 words), and clear subheadings.
• Examples: Grounding content in real businesses, results, and metrics.
• Prompting: We use specific prompts for AI tools, such as: “Write in the tone of Boot Camp Digital: direct, practical, and results-focused. Avoid jargon. Use numbered lists and frameworks. Be specific, structured, and provide examples.” We also create AI personas like: “You are a digital marketing expert trained by Boot Camp Digital. Speak in a clear, structured, practical tone. Avoid vague phrases. Focus on action steps, proven methods, and frameworks.”

The ultimate goal is to teach and drive action, never just to fill space. We want to give you real expert advice, focused on what actually works, with demos, templates, and playbooks.

Embrace the Future of Work

Agentic AI truly is a powerful next step in how we interact with artificial intelligence. By allowing AI to operate with autonomy, make decisions, and proactively work towards your goals, you can offload complex, multi-step tasks that previously consumed valuable time. Start by identifying your repetitive workflows, choosing the right tools, and investing time in training your AI agents. The payoff in efficiency and increased capacity for strategic work is immense. We’re incredibly excited about the potential of Agentic AI for our business, and we hope you are too!