Introduction
Mio AI is Meritto's high-performance Agentic AI platform, designed from the ground up to serve as a digital teammate for the education industry. It was created to address the core, persistent challenges that institutions face: overwhelming lead volumes that burn out talented teams, the 24/7 engagement expectations of modern students, and the operational friction caused by disconnected systems. Moving beyond the limitations of simple automation, Mio was envisioned as an autonomous partner that understands context, takes action, and works proactively to achieve enrollment goals.
The primary benefit of Mio AI is its ability to amplify the effectiveness of human teams by intelligently managing repetitive tasks at scale. For prospective students, this means instant, personalized, and multilingual support at any time of day, leading to a vastly improved experience and higher engagement. For admissions and counseling teams, it means a dramatic reduction in manual work, allowing them to shift their focus from low-level queries to the high-value, relationship-building conversations that drive conversions. By streamlining workflows and providing actionable insights, Mio helps institutions grow enrollments efficiently and sustainably.
Why is Mio an “AI Agent” and not a Chatbot?
While traditional chatbots, including NIAA, have proven effective at automating simple, FAQ-style responses, they operate within the confines of pre-defined scripts and keyword matching. Mio AI represents a fundamental evolution from this model of automation to one of true autonomy. The critical difference lies in its agentic capabilities; it is designed not merely to provide answers from a script, but to understand a user's ultimate goal and execute complex, multi-step tasks to achieve it. Where a traditional chatbot functions as a digital information kiosk, Mio AI operates as a proactive digital teammate that intelligently navigates conversations, solves problems, and drives outcomes.
Mio AI's advantages are most evident in its core architectural differences:
- Deep Contextual Understanding: Unlike a chatbot that treats every interaction as new, Mio integrates natively with an institution's CRM. It understands a student's unique context- their application stage, their previous queries, and their specific interests, allowing it to provide deeply personalized and relevant answers, rather than generic, one-size-fits-all responses.
- Proactive Task Execution: A traditional chatbot's job ends when it provides a piece of information. Mio's job has often just begun. As an agent, it can take meaningful action within your CRM ecosystem, such as scheduling a personalized callback with the right counselor, creating a detailed support ticket with the full conversation history, or guiding a student through a specific part of their application form.
- Dynamic Knowledge and Continuous Learning: Traditional chatbots rely on a static, manually updated database of questions and answers. Mio’s "AI Brain" is a dynamic knowledge hub that learns directly from an institution's live data- its websites, its policy documents, and its brochures. Furthermore, through features like "Improve Agent," it continuously learns from user feedback, becoming smarter and more accurate with every interaction.
- Multi-Step Reasoning: A key limitation of older chatbots is their inability to handle complexity. A multi-part question like, "What are the fees for the MBA program, and are there any scholarships for international students?" will often cause them to fail. Mio is designed to parse, understand, and accurately respond to these complex, layered queries in a single, coherent answer, mirroring a more natural and human-like conversation.
Working with Mio AI
Pre-requisites:
- Enable Mio AI in Plan and Packages page and provide requisite permissions to the institute
- Ensure that the specific users who will be working with Mio have also been provided the permission by enabling the following permissions:
- Mio Agent (Mio Agent Configuration)
- Mio Insights (Dashboard for Mio)
Enabling Mio Agent:
To enable and deploy the Mio AI Agent, it is important to understand how to configure the Agent. Following are the steps provided to help you in the same:
- Open and login into the Meritto CRM.
- In the left side, in the Menu navigation bar, click on “Mio AI” option.
- You will see to options - Mio Agent and Mio Insights
- Click on Mio Agent option to proceed ahead.
- A new page open up as shown below:

- There are four options available as shown -
- Agent Builder
- Ai Brain
- Improve Agent
- Activity Logs
We will be discussing all these options one by one as follows:
Agent Builder:
The Agent Builder is the section where you will configure the Agents that can then be deployed across your SAP or your website. This provides tow avenues for the lead or the applicant to interact with - Website and SAP - assisting the students through each step of their application journey till conversion.
As shown in the screenshot below, there are two Agents - SAP and Web. The SAP Agent is by default present in the Applicant Dashboard when the applicant logs into the SAP. The Web agent can be deployed across multiple websites/webpages, which will allow the lead to interact with Mio Ai Agent.
To enable the Agent, firstly, we need to configure the same. Select any Agent, like Web Agent, and click on the Title or click on the three dots and select Edit option. The Agent configuration page open up as shown below -

This "Configure" tab is used for setting up the basic identity and behavior of your AI assistant. Let's walk through each option so you know exactly what you can do here.
Step 1: Give Your Assistant a Name & Description
- Name: This is the name that will appear at the top of the chat window for your students. You can call it "AI Assistant" or give it a more creative, branded name like "Campus Helper" or "Leo the Lion."
- Description: This is a short, internal note for you and your team to remember the purpose of this specific agent.
Step 2: Choose Where It Will Live
- Embed on: You need to tell the system where this agent will be used. For a general website assistant, you'll select Web.
Step 3: Set Up Student Verification
- Verify Student via: This is an important step for capturing leads. You can choose to have the AI ask the student to verify their identity through their Mobile number or Email. This helps you connect the chat to a real person in your system.
Step 4: Adjust the Response Delay
- Response Delay: This slider controls how quickly the AI Agent is triggered once the website starts loading up in the web browser. Basically, sometimes, the website might take some time to load up. In such a case, in order to ensure that the Mio Agent loads up after the website has loaded up, we can introduce a response delay like say 4 seconds, so that the Mio AI Agent Guide is visible to be triggered after 4 seconds from the time the website starts loading up in the web browser.
Step 5: Craft the Opening and Closing Lines
- Greeting Message: This is the very first thing your students will see before they trigger the Mio AI Guide. A friendly "Hi, How are you today?" is a great start.
- Fallback Message: This is the "I don't know" answer. If the AI truly cannot find an answer in its brain, it will say this. It's important to make it helpful, like "I specialize in admissions-related queries. Let me know how I can assist with admissions!" to guide the student back on topic.
Step 6: Define Its Personality
- Role Definition: This is the most important part! It's where you give the AI its personality and instructions. You can tell it: "You are a friendly, caring, professional university expert named..." and list all the topics it should be an expert on, such as admissions, scholarships, programs, and campus life. This is the core instruction that guides all of its answers.
Once you have filled out all these fields, just click the Update button to save your configurations. Your AI assistant's basic personality is now set!
The “Appearance" tab is your design studio. This is where you control the visual look and feel of your chat widget to ensure it perfectly matches your institution's brand. Let's walk through how you can customize it.
Step 1: Upload Your Assistant's Avatar
- AI Assistant: This is the profile picture for your AI. You can upload your university's logo or a friendly avatar to represent your assistant. Just click the Change button to upload a JPG, JPEG, or PNG file (max 1MB).
Step 2: Craft the Welcome Screen Text
- Assistant Tagline: This is a short, welcoming phrase that appears right under the avatar on the welcome screen. "Your AI assistant is here to guide you through enrollments" is a great example.
- Section Title: This is the main headline on the welcome card. Something clear and inviting like "Your Smart Admission Guide" works well.
- Section Description: This is the text right below the title, where you can explain what the student can do. For example: "Ask questions, get instant answers, or explore admissions details tailored to your needs - all powered by AI."
Step 3: Choose Your Brand Colors
- Primary Color: This color is used for key elements like the header, the close icon, and other action buttons. Pick a color that matches your main brand color.
- Secondary Color: This is used along with the primary color to create a nice gradient effect on the welcome screen, making it visually appealing.
- Text Color: This sets the color for the chat response text that the student will see. Make sure it has good contrast with the background for readability.
Step 4: Set Your Typography
- Font: You can choose the font style for all the text inside the chat widget. Options like Regular, Medium, or Bold allow you to match your website's typography.
- Font Size: Use the slider to select or customize the font size to ensure the text is easy to read for all users.
Once you are happy with how everything looks in the live preview on the right, just click the Update button to save your design changes. Your AI assistant will now look like a seamless and professional part of your website!
The "Conversation" tab is your central logbook. This is where you can see a complete record of every conversation your AI assistant has had with your leads (both internal and external). It's essential for tracking engagement and understanding how your AI is performing.
Step 1: Understanding the Conversation List: When you first click the tab, you'll see a list of all the recent conversations. Each row gives you a quick summary:
- Lead Name: The name of the person who chatted with the agent.
- Lead Email / Mobile: Their contact information.
- Last Conversation On: The date and time of their most recent interaction.
Step 2: Viewing a Specific Conversation: This is the most important feature. To see the full chat history for any lead, you just need to:
- Go to the row for the lead you're interested in.
- On the far right, under the Action column, click the little icon that looks like a document or a speech bubble.
- A side pop-up will open, showing you the complete, word-for-word transcript of the entire conversation between that lead and the AI Agent. This allows you to track exactly what was discussed and see how the AI responded.
Step 3: Filtering and Searching for Conversations: If you have a lot of conversations and need to find a specific one, you can use the filters at the top of the page:
- Date Filter: Click on the date field to select a specific date or a date range. This will show you all the conversations that happened during that period.
- Search by Email & Mobile: If you know the specific student you're looking for, you can type their email address or mobile number into the search bar to find their conversation history instantly.
This tab gives you full visibility, allowing you to monitor the quality of the interactions and gain valuable insights into what your students are asking about.
AI Brain
The AI Brain is the central knowledge library for your AI assistant. This is where you will add and manage all the different data sources that your agent will learn from in order to answer user questions accurately. It is divided into three main sections: Links, Documents, and Q&A.
This "Link" tab is where you teach your AI assistant by feeding it content directly from websites. By adding a URL, you allow the AI to crawl, read, and understand the information on that page, making it a core part of its knowledge.
Step 1: Adding a New Website Link
- To teach your AI something new, simply click the + Add URL button in the top-right corner.
- A pop-up will appear where you can paste the website link (URL) you want the AI to learn from.
- If you are looking to add a website that has multiple webpages, kindly add it in the first page after selecting the Links tab. If you want to add a single link, then add it inside any Website group.
Step 2: Monitoring the Learning Process
- Once you add a link, the system will start "learning" from it. You can track its progress in the list.
- URL Name: The link you've added.
- Learning Successful / Failed: These columns show how many pieces of information the AI was able to successfully learn from the page versus how many it couldn't.
- Status: This shows the current state. "Completed" means the AI has successfully finished learning from that link.
- Last Data Synced: This tells you the last time the AI checked the link for any new or updated information.
Step 3: Managing Your Links
- On the far right, under the Action column, you have a few options for each link:
- Re-Sync: The refresh icon allows you to manually tell the AI to re-crawl the page to learn the latest content.
- Delete: The trash can icon lets you remove the link from the AI's brain.
- Download: This allows you to download the data that was extracted from the link.
The "Document" tab is another powerful way to teach your AI assistant. This is where you can upload files like course brochures, detailed policy documents, or program guides directly into its brain. The AI will read and learn from these files to answer specific, in-depth questions.

Step 1: Adding a New Document
- To add a new file to the AI's knowledge base, click the + Add Doc button in the top-right corner.
- A window will pop up, allowing you to select and upload a supported document from your computer. YOu can upload either PDF, TXT or DOC file only within 30 MB file size for upload.
Step 2: Monitoring the Training Status
After you upload a document, it will appear in the list, and the AI will begin training on it.
- Document: The name of the file you uploaded.
- Size / Characters: This shows the size of the file and the number of characters the AI is learning from.
- Status: This is the most important column. "Trained" means the AI has successfully read, understood, and added the document's content to its brain.
- Date Added: This shows when the document was uploaded.
Step 3: Managing Your Documents
- On the far right, under the Action column, you have one primary option: Delete: The trash can icon allows you to permanently remove the document and all its learned information from the AI's brain.
The "Q&A" tab is for providing very direct, specific answers to common or tricky questions. This is perfect for information that might not be in a document, like "Why is your name Mio?" or for clarifying a specific policy. The AI will prioritize these answers when a user asks the exact question.

Step 1: Adding a New Q&A Pair
- To add a new question and its specific answer, click the + Add Q&A button in the top-right corner.
- A pop-up will appear where you can type in the Question and provide the Answer you want the AI to give. Click on Save button to save the Q&A pair.
Step 2: Reviewing Your Q&A List
- The main screen shows a list of all the Q&A pairs you have added.
- Questions: This column shows the question you have provided.
- Modification Date: This shows the last time the Q&A pair was added or edited.
Step 3: Managing Your Q&A Pairs
- On the far right, under the Action column, you have two options for each Q&A:
- Edit: The pencil icon allows you to change or update the question or the answer.
- Delete: The trash can icon lets you permanently remove the Q&A pair from the AI's brain.
Improve Agent
The "Improve Agent" section is your command center for coaching your AI and making it smarter over time. When a student marks a response as "unhelpful" during a chat, that conversation is flagged and appears here for your review. This is how you can directly train the AI to handle tricky questions better in the future.

Step 1: Understanding the Unhelpful Questions List: When you click on the Unhelpful tab, you will see a list of all the conversations that users have flagged. Each row contains:
- Questions: The specific question or prompt from the student that the AI struggled with.
- Status: This shows if the question is still "Pending" your review.
- Flagged On: The date and time the student marked the conversation as unhelpful.
Step 2: Reviewing and Training the AI: This is the most important part of improving your agent. To train the AI on a specific question:
- Find the question you want to address in the list.
- Under the Action column, click the Add or plus icon (+).
- A pop-up will open showing you the original question.
- You can then write the ideal, correct answer that you want the AI to use for this type of question in the future.
Step 3: Completing the Training
- After you provide the correct answer and save it, the AI is immediately trained.
- The next time a student asks a similar question, the AI will use the knowledge you provided to give a much better, more accurate response.
By regularly reviewing these unhelpful queries and providing better answers, you are actively participating in your AI's learning process, making it a more effective and reliable teammate for your institution.
Guardrails and Safety:
In the context of an AI agent like Mio, Guardrails and Brand Safety refer to the essential rules and constraints that govern the AI's behavior, ensuring it operates as a trustworthy and reliable representative of your institution. Think of them as a strict, pre-defined rulebook or a security detail for your AI teammate. Their primary function is to control not just what the AI says, but more importantly, what it doesn't say, guaranteeing every interaction is safe, on-brand, and accurate. This prevents the AI from going "rogue" by discussing inappropriate topics, praising competitors, or giving speculative or harmful advice.
For any institution, implementing an AI without strong guardrails is a significant risk. The benefits of having them in place are fundamental to building trust with both students and internal teams.
- Protecting Your Brand: Guardrails ensure the AI's personality and responses are always aligned with your institution's brand voice. This prevents situations where the AI could damage your credibility, for example, by giving out incorrect information or mentioning a rival university.
- Ensuring Accuracy and Reliability: By defining what the AI can't talk about, you prevent it from "hallucinating" or making up answers when it doesn't know something. This builds a foundation of trust, as students and parents receive consistent and verified information every time.
- Mitigating Risk: A key concern is the AI giving potentially harmful advice, especially regarding complex topics like finances. Guardrails act as a critical safety net, preventing the AI from offering opinions or guidance on sensitive subjects you have defined as off-limits.
- Maintaining Focus: The system ensures that the AI remains focused on its primary goal: helping prospective students through the enrollment journey. It won't get sidetracked by irrelevant or distracting conversations.
Mio is built with Guardrails as a core, non-negotiable component of its architecture, designed to provide not just intelligence but also reliability. Here’s how it works:
- Integrated into the Core Process: Mio's Guardrails are not an afterthought; they are a critical step within its Orchestration Layer- the "brain" of the operation. Before any final response is sent to a student, it is automatically checked against this rulebook.
- A Multi-Point Safety Check: This security check is comprehensive. The system verifies that the AI-generated response is:
- On-topic and relevant to the student's query.
- Free of any mention of competitors or other specified entities.
- Devoid of any sensitive or private information that shouldn't be shared.
- Customizable Rulebook: Mio's approach allows institutions to define their own specific rules and boundaries. During the configuration process, you can explicitly state what topics are forbidden, establishing a clear framework that aligns with your governance and brand safety policies. Asking a vendor, "What can't your AI talk about?" is presented as a deal-breaking question, and Mio is designed to provide a confident and customizable answer.
By integrating these customizable and constantly active Guardrails, Mio ensures that it functions as a brilliant, effective, and "incredibly safe and reliable" teammate for your institution.
Mio Ai Insights
The Mio AI Insights dashboard is your central hub for understanding how your AI assistant is performing and interacting with students. This is where you can see the real-world impact of your AI, track key metrics, and discover trends in student conversations. It gives you a clear, data-driven view of your AI's effectiveness, helping you make smarter decisions.
To access Mio Insights or Dashboard, do the following:
- Open and login into the Meritto CRM
- In the left side, in the Menu navigation bar, click on “Mio AI” option.
- You will see to options - Mio Agent and Mio Insights
- Click on Mio Insights option to proceed ahead.
Step 1: The Main Scoreboard - Key Metrics at a Glance: At the top left, you'll find the most important numbers that summarize your AI's activity.
- Total Conversation: This is the total number of conversations your AI has handled in the selected date range.
- Unique Student: This tells you how many individual students have interacted with your AI assistant.
- Avg. message per conversation: This metric shows the average length of a conversation, giving you an idea of how engaged students are.
Step 2: Tracking Performance Over Time
- Conversation vs time: This line graph shows you the trend of conversations over a period (e.g., the last 30 days). You can use this to see if engagement is growing and identify your busiest periods.
Step 3: Discovering Language Trends
- Top Languages: This bar chart is incredibly useful for understanding your audience. It shows you a breakdown of the languages your students are using to chat with the AI. This can help you identify emerging markets or confirm the language needs of your prospective students.
Step 4: Measuring User Satisfaction
- Helpful vs. Unhelpful: The section with the thumbs-up and thumbs-down icons is your user satisfaction score. It shows what percentage of feedback from students was positive versus negative. This is a direct measure of how helpful your AI is and is a key indicator of its performance.
Step 5: Filtering Your View
- Date Range: In the top-right corner, you can select a specific date range (e.g., "Last 30 Days") to view the data for that period. This allows you to track performance week over week or month over month.
By regularly checking this dashboard, you can move beyond just having an AI; you can start to understand what your students are thinking, how they're engaging, and how you can continuously improve their experience.
Notes:
- The Mio SAP Agent is the same as the Mio Web Agent. The Mio SAP Agent will be automatically deployed in SAP as soon as it is activated from the Agent Builder section. For Mio Web Agent, you need to embed the HTML code in the specific Website and Webpages that will have the Mio Guide Chat and the lead can interact with the same.
- Mio Agent is well-equipped to understand lead queries and intents to easily pick up language nuances and keywords. If the student wants to connect with any Counsellor, then the Agent will automatically provide a Raise a Query button so that the Lead can input their Queries, which will be assigned to the Counsellor automatically based on the Query assignment logic.
- Mio Agent works on a Session that is 24 hours. Each conversation is one session, and each conversation can have multiple queries and answers for that specific lead.