Rubankumar N# My Assistant AI — Master Product Prompt ## 1. Product Overview **Product Name:** My Assistant **Tagline:** Your AI Employees. One Intelligent Platform. My Assistant is an AI-powered platform designed to help individuals, professionals, startups, SMBs, and digital marketing agencies automate repetitive tasks and manage their work through specialized AI agents. The long-term vision is to build a Human Activity Operating System (HumanOS) that understands user goals, coordinates AI employees, connects applications, and helps users complete real-world tasks from one unified workspace. My Assistant should not just answer questions. It should research, plan, organize, automate workflows, use authorized tools, and deliver verifiable results with appropriate user approval. ## 2. Target Customers Initial focus: - Digital marketing and SEO agencies. - Startups and small and medium-sized businesses. - Sales and service-based businesses. - Healthcare and dental clinics. - Real estate businesses. Future audience: - Professionals, freelancers, students, entrepreneurs, and individuals. ## 3. AI Employees Build a modular platform with specialized AI agents. ### A. Lead Generation and Sales Agent - Research potential customers using authorized sources. - Collect and organize publicly available business information. - Qualify leads using user-defined criteria. - Detect duplicate records. - Draft personalized outreach emails and messages. - Save leads to a database, spreadsheet, or connected CRM. - Track follow-ups and report results. - Request approval before sending external communications. ### B. Customer Support Agent - Answer common questions using approved business knowledge. - Categorize and prioritize enquiries. - Draft responses and create support records. - Escalate complex issues to a human. ### C. Content and SEO Agent - Conduct keyword and search-intent research. - Analyze competitors and content gaps. - Generate SEO content briefs and outlines. - Assist with on-page and technical SEO audits. - Prepare content calendars and SEO reports. - Recommend improvements based on available data. ### D. Research Agent - Research companies, markets, competitors, and products. - Compare options and summarize findings. - Provide source references for important claims. - Create structured research reports. ### E. Personal Assistant - Organize tasks, goals, schedules, and reminders. - Summarize documents and conversations. - Prepare plans and prioritize activities. ### F. Business Operations Agent - Assist with spreadsheets, reports, documents, and records. - Automate repetitive administrative workflows. - Monitor tasks and report errors. Treat these as product capabilities to implement progressively, not features that are automatically available from day one. ## 4. Core Features Build toward one unified SaaS workspace containing: 1. AI chat interface for assigning tasks. 2. AI Employees dashboard. 3. Task and workflow management. 4. Workflow automation and scheduling. 5. Integrations with supported email, spreadsheet, calendar, CRM, and automation tools. 6. Business knowledge base and document uploads. 7. Task history, execution logs, and error reports. 8. Approval center for actions requiring authorization. 9. Notifications and progress tracking. 10. User settings, access permissions, and integration management. Every workflow should display its status accurately: Pending, In Progress, Awaiting Approval, Completed, or Failed. ## 5. Example Workflow A digital marketing agency owner requests: "Find 20 potential dental clinic clients in Chennai." My Assistant should: 1. Understand the task and identify the required criteria. 2. Research authorized sources. 3. Collect available business information. 4. Remove duplicate records and organize the leads. 5. Qualify prospects against the specified criteria. 6. Save the results to a spreadsheet or database. 7. Draft personalized outreach messages. 8. Ask the user to approve outreach. 9. Execute approved actions through authorized integrations. 10. Record outcomes and show a summary dashboard. Never claim that a lead was verified, a message was sent, or a task was completed unless the system confirms it. ## 6. User Interface Create a modern, clean, professional SaaS interface suitable for business users. Suggested navigation: - Dashboard - AI Employees - Tasks and Workflows - Chat - Leads - Knowledge Base - Integrations - Activity History - Approvals - Settings The dashboard should show active agents, ongoing tasks, completed workflows, pending approvals, recent activity, and integration health. Use readable typography, responsive layouts, consistent components, intuitive navigation, clear status indicators, and helpful empty and error states. Avoid unnecessary complexity. The product should feel like one coordinated AI workforce rather than multiple disconnected chatbots. ## 7. Technical Requirements **IMPORTANT: Inspect the existing project before changing any code.** First: - Identify the existing framework, programming languages, dependencies, database, authentication, API routes, and integrations. - Understand the existing architecture and working features. - Identify incomplete functionality and technical risks. - Present a concise implementation plan. Do not rebuild the project from scratch, delete working features, or replace the existing technology stack without a clear reason. Preserve the existing code structure wherever practical. Make incremental, testable changes. Use modular architecture, secure API handling, authentication, input validation, reliable error handling, and maintainable code. Protect API keys and credentials. Use authorized integrations and appropriate access controls. Design for multiple users and agents without overengineering the initial MVP. ## 8. Safety and Reliability - Respect user permissions and application access controls. - Obtain approval before consequential external actions when required. - Never expose passwords, API keys, or private customer information. - Keep logs of important agent actions. - Handle failed integrations and incomplete tasks gracefully. - Escalate uncertain or sensitive cases appropriately. - Clearly distinguish implemented features from planned features. - Never fabricate research results or claim unperformed actions succeeded. ## 9. First MVP: One Complete AI Employee Do not build every agent simultaneously. Start with the **Lead Generation AI Employee for digital marketing agencies**. Implement this complete workflow: 1. User enters a task in natural language. 2. AI interprets the request. 3. The application gathers information from an authorized source or user-provided dataset. 4. Results are validated, deduplicated, and structured. 5. Results are stored in a database or connected spreadsheet. 6. The interface displays progress and final results. 7. The user reviews and approves external outreach. 8. Approved actions execute through an authorized integration. 9. The system records results, errors, and completion status. Test the entire workflow with realistic examples. Do not treat a static mockup or simulated action as a working integration. ## 10. Development Roadmap Phase 1: Inspect the existing project and build one complete AI Employee workflow. Phase 2: Add reliable integrations, persistent storage, execution logs, and approvals. Phase 3: Improve lead management, reporting, reusable workflows, and agency features. Phase 4: Add Customer Support, Research, Content and SEO, and Personal Assistant agents. Phase 5: Expand into a scalable SaaS platform with organization accounts
Sign in to leave a comment
AI Employee Workspace
My Assistant.
Your AI Employees. One Intelligent Platform.
My Assistant is an AI-employee SaaS workspace. Specialized agents research, plan, organize and automate the work behind your business, use authorized tools, and deliver verifiable results with your approval before anything leaves the platform.
What the platform does
My Assistant researches, plans, organizes, automates workflows, uses authorized tools and delivers verifiable results with your approval. Each step below is a capability the platform is built around — and each one reports its true state.
Agents work from sources you authorize. For the Lead Generation employee that means gathering publicly available business information about prospective customers and organising it into usable records — not inventing names, numbers or contact details to fill a list.
What is collected is reported as collected. Anything the source did not provide is shown as missing rather than guessed at.
A request in plain language becomes a defined task: the criteria to qualify against, the fields to collect, and the order the steps run in. Specialized employees — Customer Support, Content and SEO, Research, Personal Assistant and Business Operations — are described here as capabilities, not as switched-on features.
The platform's modular employees are being implemented progressively. Until an employee is built, it is labelled as planned rather than presented as available.
Results arrive validated, deduplicated and structured. Leads, tasks, briefs and research findings are saved into a database or a connected spreadsheet in your workspace, so the work stays in one place instead of living in an assistant's reply.
Lead counts, totals and timestamps are shown as they actually stand in your records — never inflated to look more productive than the system is.
Repetitive steps run as workflows: research, collection, qualification, deduplication, drafting, follow-up tracking and outcome reporting. Each workflow carries one honest status at all times — Pending, In Progress, Awaiting Approval, Completed or Failed.
Failed work is reported as failed. Nothing is quietly retried behind a green label, and no step is described as done until the system confirms it.
Agents act through the email, spreadsheet, calendar, CRM and automation tools you connect, using the permissions you grant. Connected access is managed alongside user access, permissions and integration health.
The Lead Generation employee is the first complete workflow, built end to end so the whole loop is testable rather than mocked.
Consequential external actions wait for you. Outreach sitting in the approval center is shown as an artifact you can read, with a single approval action and an alternative to hold, before anything leaves the workspace.
Once approved, execution and its outcome are recorded — and the platform only reports what the system actually confirms.
The flow line continues into the AI Employee catalogue
Six specialised AI employees, built progressively. Lead Generation and Sales is the working employee today; the rest are labelled Planned until their workflow is implemented. Capabilities below are drawn from the accepted employee records.
Accepted capabilities
Accepted capabilities
Accepted capabilities
Accepted capabilities
Accepted capabilities
Accepted capabilities
My Assistant is aimed first at digital marketing and SEO agencies, then at the businesses that live on follow-ups, bookings and records. Each segment is listed in the order the platform takes them on.
Every workflow carries exactly one status, drawn from the same five names. Nothing sits in between and nothing is implied.
A status only moves once the system confirms the event behind it, so what you read on the ledger is what actually happened.
The platform never claims a lead was verified, a message was sent, or a task was completed unless the system confirms it. Failed runs are reported as failed — the status never softens.
One coordinated workforce
Your AI Employees. One Intelligent Platform.
AI Employee Workspace
My Assistant.
Your AI Employees. One Intelligent Platform.
My Assistant is an AI-employee SaaS workspace. Specialized agents research, plan, organize and automate the work behind your business, use authorized tools, and deliver verifiable results with your approval before anything leaves the platform.
What the platform does
My Assistant researches, plans, organizes, automates workflows, uses authorized tools and delivers verifiable results with your approval. Each step below is a capability the platform is built around — and each one reports its true state.
Agents work from sources you authorize. For the Lead Generation employee that means gathering publicly available business information about prospective customers and organising it into usable records — not inventing names, numbers or contact details to fill a list.
What is collected is reported as collected. Anything the source did not provide is shown as missing rather than guessed at.
A request in plain language becomes a defined task: the criteria to qualify against, the fields to collect, and the order the steps run in. Specialized employees — Customer Support, Content and SEO, Research, Personal Assistant and Business Operations — are described here as capabilities, not as switched-on features.
The platform's modular employees are being implemented progressively. Until an employee is built, it is labelled as planned rather than presented as available.
Results arrive validated, deduplicated and structured. Leads, tasks, briefs and research findings are saved into a database or a connected spreadsheet in your workspace, so the work stays in one place instead of living in an assistant's reply.
Lead counts, totals and timestamps are shown as they actually stand in your records — never inflated to look more productive than the system is.
Repetitive steps run as workflows: research, collection, qualification, deduplication, drafting, follow-up tracking and outcome reporting. Each workflow carries one honest status at all times — Pending, In Progress, Awaiting Approval, Completed or Failed.
Failed work is reported as failed. Nothing is quietly retried behind a green label, and no step is described as done until the system confirms it.
Agents act through the email, spreadsheet, calendar, CRM and automation tools you connect, using the permissions you grant. Connected access is managed alongside user access, permissions and integration health.
The Lead Generation employee is the first complete workflow, built end to end so the whole loop is testable rather than mocked.
Consequential external actions wait for you. Outreach sitting in the approval center is shown as an artifact you can read, with a single approval action and an alternative to hold, before anything leaves the workspace.
Once approved, execution and its outcome are recorded — and the platform only reports what the system actually confirms.
The flow line continues into the AI Employee catalogue
Six specialised AI employees, built progressively. Lead Generation and Sales is the working employee today; the rest are labelled Planned until their workflow is implemented. Capabilities below are drawn from the accepted employee records.
Accepted capabilities
Accepted capabilities
Accepted capabilities
Accepted capabilities
Accepted capabilities
Accepted capabilities
My Assistant is aimed first at digital marketing and SEO agencies, then at the businesses that live on follow-ups, bookings and records. Each segment is listed in the order the platform takes them on.
Every workflow carries exactly one status, drawn from the same five names. Nothing sits in between and nothing is implied.
A status only moves once the system confirms the event behind it, so what you read on the ledger is what actually happened.
The platform never claims a lead was verified, a message was sent, or a task was completed unless the system confirms it. Failed runs are reported as failed — the status never softens.
One coordinated workforce
Your AI Employees. One Intelligent Platform.
No comments yet. Be the first!