Q: Software Clarification Questions
Hello Support Team,
1.) Great product! Can Diaflow preform tasks in other software like a human? If so, please provide a video training or instructions?
2.) Is it possible to seperate automations via workspaces for each business or company department?
3.) How can we keep AI Agents, Apps & Workflows organized and seperate with the option to collaborator with teams?
4.) Are tasks completed using the prompts given to the AI agents or created workflows? Can you explain how your No-code flow builder works?
5.) Do you have any use cases of how an AI can complete a tasks from beginning to end in Quickbooks or for sales calls, if possible?
6.) Is Pydantic technology being used to create AI agents?
Dung_Diaflow
Apr 24, 2025A: Hi thomas.anthony,
We’re thrilled you’re excited about Diaflow—thanks for the kind words and great questions! I’ll address each one based on the latest info we have:
1. Performing Tasks in Other Software Like a Human: Diaflow doesn’t natively mimic human actions in other software (e.g., clicking or typing in apps like Selenium does). Instead, it automates data processing and integrates with other tools via APIs or webhooks. We don’t have a specific video for this, but you can explore our YouTube channel (https://www.youtube.com/@diaflow_io) for general flow-building tutorials or check https://docs.diaflow.io for node instructions.
2. Separating Automations via Workspaces: Currently, all team members in a workspace share access to automations, apps, and agents unless restricted via flow logic. Separate workspaces per business or department aren’t fully live yet. For now, you can organize flows within one workspace (especially in Tier 4 with 50 users) and use naming conventions to keep them distinct.
3. Organizing AI Agents, Apps & Workflows: You can keep things organized by naming projects clearly in your workspace (e.g., “Marketing_Chatbot” or “HR_Automation”). All users in the workspace can collaborate, but there’s no native feature to isolate specific flows for certain team members yet.
4. Tasks via Prompts or Workflows & No-Code Flow Builder: Tasks can be completed using both AI agent prompts and workflows. AI agents rely on prompts to process inputs (e.g., “Summarize this email”), while workflows combine nodes (e.g., Webhook, LLM, Output) to automate multi-step processes. The no-code flow builder lets you make nodes to create flows visually—think connecting a Webhook trigger to an LLM node to process data, then saving it to a Table node. Check https://docs.diaflow.io/flows for a detailed guide.
5. Use Cases for Quickbooks or Sales Calls: There’s no specific Quickbooks integration or sales call use case, but you could connect Quickbooks via its API using the HTTP Request node to automate tasks like data entry. For sales calls, you could build an AI agent to transcribe calls (using Whisper) or generate follow-up emails based on call notes. Explore the Templates section for similar examples.
6. Pydantic Technology for AI Agents: Diaflow’s agents are built using our hosted LLMs (e.g., GPT-4o, Claude) and configured via prompts and flows. If Pydantic is critical for you, let us know more at support@diaflow.io, and we can explore workarounds.
We’re here to help you get the most out of Diaflow! Join our Discord (https://discord.com/invite/pcfWGjApVC) for quick support or email us to dive deeper into your use cases.
Best,
Diaflow team.