Why I Built My Own Agentic AI Tool and How Agency Operators Can Use It

Earnings Disclaimer: You have a .1% probability of hitting million-dollar months according to the US Bureau of Labor Statistics. As stated by law, we can not and do not make any guarantees about your own ability to get results or earn any money with our ideas, information, programs, or strategies. We don’t know you, and besides, your results in life are up to you. We’re here to help by giving you our greatest strategies to move you forward, faster. However, nothing on this page or any of our websites or emails is a promise or guarantee of future earnings. Any financial numbers referenced here, or on any of our sites or emails, are simply estimates or projections or past results, and should not be considered exact, actual, or as a promise of potential earnings – all numbers are illustrative only.

I recently launched Utari, my own agentic AI tool. This whole thing started because I was obsessed with a tool called Manus. If you caught the AI news recently, Meta supposedly acquired Manus. That’s when I knew we were onto something worth exploring.

One of my Inner Circle members, Daniel Fazio, introduced us to Manus and changed how I think about AI tools. He explained something that shifted my understanding of how these AI platforms actually work.

Results are not typical. Your results will vary and depend entirely on your individual capacity, business experience, expertise, and level of desire. There are no guarantees concerning the level of success you may experience. The testimonials and examples used are not intended to represent or guarantee that anyone will achieve the same or similar results. We don’t believe in get-rich-quick programs. We believe in hard work, adding value and serving others. As stated by law, we can not and do not make any guarantees about your own ability to get results or earn any money with our information, courses, programs, or strategies.

If you’re looking for structured frameworks on building operational systems for your agency, my 7-week live comprehensive training covers these concepts in depth alongside other core agency infrastructure.

How Credit Based AI Tools Process Data Differently Than Subscription Models

Here’s what most people don’t understand about AI tools.

When you upload files to monthly subscription platforms like ChatGPT, Gemini, or Claude, they’re creating efficient summaries because you’re a flat-rate monthly user. Credit-based tools work differently. Even though they might use APIs from ChatGPT, Claude, or Gemini, the credit-based model means they process your entire upload — every single word and every piece of data.

According to McKinsey’s research on AI adoption, organizations are increasingly looking at how AI tools handle data processing and thoroughness. The difference in approach matters when you’re uploading a year’s worth of sales call recordings. A credit-based system processes all of it. Not just skimming. Not just summarizing.

That’s when I fell in love with the platform. The thoroughness was exactly what I needed for my businesses.

What Agentic AI Mode Does and Why I Wanted It Exclusively

Manus had a feature that changed how I used AI. They had three different modes, and the middle one was called agentic mode.

That’s where I discovered what agentic AI could do. The difference between normal chat-based AI and agentic mode was significant. But then Manus made a decision that frustrated me: they removed the option to use agentic mode exclusively and switched to an automatic mode that toggled between regular chat and agent mode based on what it thought was best.

I wanted control. I wanted agentic mode all the time.

So I called AI developers who were deep in this space. I told them I wanted to build something similar to Manus, but exclusively in agentic mode. I wanted an agent-only AI tool that I owned and controlled. Initially, this was just an internal tool. That’s how Utari was born.

How Agentic AI Works and What Agents Can Actually Do

Let me break down what agents actually do, because most people don’t really understand this.

The simplest way to think about it is this: anything a person can do at a computer, an agent can attempt on your behalf. Gartner’s analysis of agentic AI describes these systems as autonomous software that can perceive, decide, and act to achieve goals.

There are different agentic software platforms out there. Most major AI tools like ChatGPT, Claude, and Gemini have versions of agentic modes. But Utari is exclusively an agentic AI tool. That’s the entire point.

When you use Utari, you’re working with agents that can execute tasks autonomously. Not just answer questions. Not just summarize documents. They actually go out and attempt work on behalf of your organization.

How to Create Your First Worker in Utari

When you log into Utari, you’ll see a clean interface. You can accomplish things just using the basic interface, but the real functionality comes from understanding workers.

In the top left, there’s an option called Workers. This is where you access what Utari can do.

When you click Add a worker, you’re essentially creating an agent. The process is straightforward, but there are some important things to understand. Every agent starts with default instructions. We give you a baseline set of instructions that makes the agent functional right out of the box. You can edit these if you want, but we don’t recommend changing the core instructions. Instead, add to them.

Here’s an example of how I customize my agents. I want every document my agents produce to be in Google Docs — not PDFs, not Word docs. Everything goes straight to Google Docs connected to my business Gmail account.

So when I’m creating a worker for ad copy, VSSL scripts, or webinar presentation slides, I just add a simple instruction: “Every document you create should be converted to a Google Doc and added to the user’s account as a default action.” That’s it. Plain language. The agent understands and executes.

How to Define Your Agent’s Specific Role and Intention

Here’s where most people miss the mark. Every agent needs a specific intention.

The instructions seem random until you define what your agent is actually supposed to do. Think about the specific role you want this agent to fill in your organization. Maybe you need an agent whose job is to write email copy for your business, analyze competitor ads against your own ads, or audit sales calls and provide daily reports to your sales manager.

In my businesses, I’ve built agents that function as sales manager assistants. They analyze every sales call that comes in, compare it against our knowledge base of call recordings, and send reports to me and my sales manager about what training needs to happen.

I have agents that monitor CRM compliance. They log into our CRM and check if leads are sitting in the same pipeline stage for too long. If a sales rep isn’t moving people through the pipeline, the agent reports it to the sales manager.

I have marketing agents that monitor every email that goes out from our CRM. They track open rates and click-through rates. If an email performs below our threshold, the agent analyzes why and can even check if our domain got blacklisted.

The point is: you need to have a clear intention for each agent. Once you have that, the instructions and setup become simpler.

How to Build and Organize Your Knowledge Base for Different Workers

One of the five key variables you control when creating a worker is knowledge.

You can build a master knowledge base where you upload all of your organization’s materials. This sits in the bottom left of the interface. Anything you add here becomes available to all your workers. But here’s the important part: you get to choose which specific pieces of knowledge each worker has access to.

If I have a sales training worker, I only give it access to sales-related materials: sales call recordings, sales training documents, scripts, objection-handling frameworks. If I have a marketing worker doing competitive analysis, it gets access to SOPs, training material, and any marketing resources that help it make decisions when producing reports or taking actions.

This separation is critical. You don’t want every agent having access to everything. You want focused, specialized agents that excel at their specific role.

According to Harvard Business Review’s research on AI implementation, how you structure and segment your knowledge base directly impacts the quality of AI outputs.

Setting Up Triggers, Tools, and Integrations for Automated Workflows

Beyond knowledge, you have triggers, tools, and integrations to work with.

• Triggers make your agents proactive. Something happens, and the agent automatically goes to work. For example, with my sales manager helper agent, every time a new sales call transcript gets uploaded to our Google Drive, that triggers the agent to analyze the call.

• You can also create workers that wait for you to interact with them. Those wouldn’t need triggers.

• Integrations are external applications. We have pages of apps and APIs you can plug into your workers: CRMs, Slack, Gmail, Stripe, payment processors, social media platforms. If it has an API, it should be capable of integrating with your Utari worker.

And here’s something I appreciate about owning the company: if there’s an integration missing that you need, just reach out. We have our developers sitting in the support chat in real time. They’re not random support reps — they’re the actual people building the platform.

Real World Use Cases for Agentic AI in Agency Operations

Let me give you some examples of how I’m using Utari in my businesses.

One use case came from a personal frustration. My wife noticed several companies were overcharging us for furniture we bought for our new home: double charges, unfulfilled delivery services, various issues. Nobody in my organization was focused on auditing every purchase. My CFO focuses on tax strategy, not auditing daily transactions.

So I created a worker called the AMX auditor. We have about eight different AMX cards across various employees. This agent logs into our AMX portal every day and audits that day’s transactions. If it finds something concerning based on parameters I’ve defined, it sends my wife and me an email with the specific transactions that look off. We’ve even given it the capacity to communicate with those companies on our behalf to dispute charges.

Through integrations with tools like 11 Labs for voice and Twilio for phone numbers, this agent can make phone calls to companies to resolve issues. It’s not just sending emails; it’s acting as a representative of our organization.

Another use case comes from Marcos at the Birdhouse. He helps people run their Twitter and LinkedIn accounts. He built a Utari worker that does competitive analysis of top tweets and LinkedIn content. It scrapes those platforms throughout the day within specific niches for specific accounts. At the end of the day, it creates a detailed report using data from their knowledge base. That report goes to key people in his organization and helps with content ideation for their clients.

These are just two examples. The possibilities are limited by your creativity about what you want your agents to do.

Why Credit Based Pricing Creates Better AI Processing

I keep coming back to this because it’s fundamental to why Utari works the way it does.

With a credit-based model, we’re incentivized to be thorough. We consume credits by doing more work, being more detailed, and processing more information. That’s the opposite of flat-rate subscription tools that are incentivized to do as little as possible to save on costs.

You’ll rarely hit credit ceilings with Utari. And if you do, credits are cost effective. We’re built on Claude from Anthropic as our foundation and always operate on the latest API available. As we grow, we’ll train our own models, but for now, we’re using the technology available.

How to Get Started With Utari and Access Developer Support

The pricing is competitive. We have free plans so you can test it out, though they’re limited compared to paid plans. Our paid plans start at $20 per month, then $50 per month, with our highest tier currently at $200 per month.

The main thing I want you to understand is this: your ability to get value from Utari is limited by your capacity to be creative about what you want your agents to do. How do you want them to support your staff? What tasks do you want them to autonomously handle? What reports do you need them to generate? What compliance do you need them to monitor?

I’d encourage you to at least try the free version and see what you come up with. We have support documentation available to help you learn the platform thoroughly. We’re also building out a use case section where you can integrate specific worker functionality into your account with one click — just add your credentials and modify as needed.

If you have questions, we have a Telegram channel for Utari users. You can also communicate with us through the support chat in the platform. Click the bottom left and press chat. You’ll be talking directly with our developers.

The domain is simple: utari.ai. Try it out, build some agents, and let me know what you come up with.

For agency operators looking to build more sophisticated operational frameworks, my 7-week live comprehensive training covers systems thinking and infrastructure development. And if you want ongoing access to how I’m implementing tools like this across my businesses, the Inner Circle is where we go deep on these implementations.

Results are not typical. Your results will vary and depend entirely on your individual capacity, business experience, expertise, and level of desire. There are no guarantees concerning the level of success you may experience. The testimonials and examples used are not intended to represent or guarantee that anyone will achieve the same or similar results. We don’t believe in get-rich-quick programs. We believe in hard work, adding value and serving others. As stated by law, we can not and do not make any guarantees about your own ability to get results or earn any money with our information, courses, programs, or strategies.

---

Watch the video: