Consider an AI Agent, a digital employee that you would onboard into your organization. You would share with this “new hire” domain knowledge of your business, and give it tools so that it can take actions in its role. VIVI allows you to create AI agents for your business without any specialized training. 

• Reflex agents: Remember when the doctor would tap your knee with a small hammer, and it would jump? It couldn’t help but react. This type of agent is governed by pre-defined scripts. They are meant to handle simple inputs and give automated responses. 

• Model-based reflex agents: These agents have a simple idea of how things work around them. It’s kind of like how we use our knowledge of the world to make decisions every day. For example, a Roomba has a basic idea of a room’s layout. It uses sensors to avoid bumping into furniture and walls while cleaning.  A thermostat knows what temperature it should be. It checks the current temperature and turns the heating or cooling on or off to reach the target. These agents combine their basic knowledge with information from their surroundings to make decisions. 

• Goal-based agents: These agents act to achieve specific goals. For example, a marketing AI agent that identifies potential leads and automates outreach campaigns. 

• Utility-based agents: These agents understand what a good outcome or a bad outcome is in pursuing their goals. If the AI is created to play chess, it aims for checkmate. It moves its pieces, accepting losses if they bring it closer to victory. It makes choices to drive the best outcome, to achieve the highest utility. 

• Learning agents: Think of how you get better at something with practice. Learning agents are similar! They learn on the job. For example, a customer service learning agent notices which answers help customers and which ones don’t. If it notices customers aren’t happy with a certain answer, it might try giving more details or suggesting a different solution. It adapts to become better at its job. 

Yes, ChatGPT is an AI agent, of sorts, but it operates differently than agents created with VIVI. You ask ChatGPT to summarize quarterly performance numbers, and it produces a summary. Without your prompting, nothing happens. A VIVI agent by contrast could analyze sales data based on its internal instructions and identify potential opportunities, then automatically send targeted offers to customers. ChatGPT is largely reactive. 

The best AI agent is the one that solves your unique business challenges. VIVI provides the tools and templates to build agents that fit your specific requirements. 

Yes, Alexa is an AI agent, but it mostly reacts to your commands. You say, ‘Alexa, set a timer,’ and it sets a timer. Simple request, simple answer. VIVI allows you to create agents that can go deeper. Your agent could monitor your company’s inventory and automatically reorder supplies when they’re running low, or it could predict future demand based on the items being sold. While Alexa can connect to a limited number of devices, VIVI is more flexible with a wider range of tools for more complex tasks. 

Is this really a question? No, humans are not AI agents. Humans possess consciousness, emotions, can understand figurative language and abstraction. They sometimes make decisions based on intuition or a gut instinct without consciously understanding why. 

You do. That’s why we built VIVI. It’s easy to be intimidated by the technical language surrounding AI. You don’t need to be a programmer. You don’t need to be a developer. VIVI makes building an agent intuitive and simple. 

Think about it like preparing a meal with a recipe. 

You begin collecting ingredients and then you combine them into something that tastes good.  

VIVI starts you off with a “meal-kit” in our pre-made templates that you can adjust to taste, or you can create something completely from scratch. 

Here’s the recipe for your VIVI agent: 

• Prompt: Give it a personality, a role, and a set of instructions. 

• Knowledge Bases: Give it information to make informed decisions. 

• Integrations: Introduce it to your environment and tools. 

• Channels: Point it toward your customers or employees. 

VIVI makes creating an agent as easy as creating a satisfying meal. It’s like having a master chef guiding you every step of the way. 

AI agents aren’t much different than us. They’re taught information. They learn from experience. As they perform their role, some of their actions are successful, some aren’t. From this, they identify patterns and begin to anticipate what comes next. Experience teaches them not to repeat mistakes. Through trial and error, they become faster, more efficient, and each success guides them toward the actions they should take in the future. 

Most people use ‘chatbot’ and ‘AI agent’ interchangeably, but they aren’t the same. Lots of chatbots use rigid scripts and struggle when humans go outside the lines. Smarter chatbots, like those using LLMs (large language models), can have normal conversations. AI agents, like VIVI, can do more than just have a conversation; they can work toward a goal and take action, not just talk about it. 

Machine learning is the method by which agents learn. It’s why an agent can think flexibly. AI agents aren’t rigid chatbots. They don’t follow scripts or need step-by-step instructions. AI agents learn to identify patterns, make decisions, and take action from the conclusions they reach. This makes machine learning a much more effective way to learn. It’s what makes AI agents smart and adaptable. 

That is a question you determine with the VIVI platform. But in general, it depends on the context. It depends on the task. With some agents it makes sense to work independently. This could be a customer service agent providing tracking numbers or answering routine questions, or an inventory management agent checking min/max levels at forward stocking locations. However an AI agent monitoring surveillance cameras may require a human to interpret a disturbance.   

VIVI lets you control how much freedom each agent has. This way, you get the right mix of automation and human control. 

AI agents like the ones you can build on VIVI are used to automate customer service, generate leads, analyze marketing data, assist with research tasks, and much more. 

No one questions the ethics of using presentation software or spreadsheets, but AI agents are different. They’re more than simple tools. They are tools with a personality, instructions, and goals. They are tools that we create. 

While agents look neutral, they carry the biases of their creators. That’s one of the reasons that the VIVI platform shares its prompt templates. You should know what’s under the hood of your agent. The decisions it makes shouldn’t be mysterious.\ 

Without transparency, it is impossible to build responsible AI agents. The VIVI platform is committed to ethical development in the areas of fairness, transparency, privacy and security, accountability, and social impact. 

VIVI provides tools and resources to help you address these ethical considerations and build AI agents that are fair, transparent, and beneficial for everyone. 

Completely secure. 

Companies have sensitive information, trade secrets, and their agents are trained on proprietary data. All of that data has to be secure, and with VIVI, it is.    

VIVI is built on Microsoft Azure Open AI and all customer data is stored in your own private data lake. VIVI’s security features include: 

• Data Isolation: Customer interactions are logically isolated and secured. 

• Compliance and Control: VIVI integrates with Microsoft’s security and compliance controls.    

• Data Sovereignty: You retain ownership and control over your data. 

With VIVI, you can have confidence that your data and AI agents are secure. 

AI agents are not yet what they fully will be. There are still gaps in their capabilities. 

For example, they can have trouble with the emotional subtleties of conversation. AI agents are like that friend who doesn’t quite get humor. Jokes go over their heads, and sarcasm is taken literally. Words with multiple meanings can confuse them. They can lose the context of a complex conversation easily. 

Also, they aren’t inventors. There’s no aha moment where they create the lightbulb or cold fusion. They’re comfortable mimics at best. 

But they’re improving. 

They handle complex decisions by using all the tools they have. They tap into the knowledge bases and connect to the systems you provide (think ERP, CRM, support notes – and so on). Their prompt? It tells them who they are, what their role is, and what they need to do. Plus, they learn by doing. Machine learning helps them see what works (and, let’s be honest, what doesn’t), so they’re always getting better at handling the complex stuff. 

Agents will be much smarter, needing less human input and integrating directly with the software you use. In fact, you probably won’t use software directly anymore. It will all filter through the agent. They’ll understand the way you work – your quirks, your preferences, the whole nine yards. They’ll collaborate on projects with you and a team of other agents. They’ll start turning up everywhere: our homes, our businesses, our cities. We’ll be trying to keep up with the changes. This whole thing is going to redefine work, and a big part of that is how it changes the skills we need. As AI takes over some tasks, those human skills – the things we can do that AI can’t – will be even more valuable. Ethical development? Oh, you better believe it. The future is rapidly approaching, and we need to be ready to adapt.