Skip to main content

8 posts tagged with "ai-agents"

View All Tags

Agent Builder Upgrades: Team Editing, Folders, and Version History

Agent Builder Upgrades: Team Editing, Folders, and Version History

Building AI agents is a team sport. The Agent Builder now supports collaborative team editing, folder organization for managing large agent libraries, and version history so you can experiment with confidence.

What's New

  • Agent Team Edit Mode: Multiple team members can now collaborate on agent configuration with the v2 team editing interface. See who's working on what, avoid conflicting changes, and build better agents together.

  • Agent Folder Organization: Organize your agents into folders with drag-and-drop. Move animations provide visual feedback as you reorganize, and folder names are validated to prevent duplicates. Find the right agent instantly, even when you have dozens.

  • Version History Sidebar: Browse and compare previous versions of your agent configuration. See what changed, when it changed, and roll back if needed. Version history gives you the confidence to experiment knowing you can always undo.

  • Updated Agent Templates: Pre-built agent templates have been refreshed with AI-generated system prompts. Start with a template that's already well-crafted and customize from there instead of starting from scratch.

  • Search & Replace in Prompt Builder: Find and replace text across your agent's prompts. Rename a product, update a phone number, or fix a typo across every prompt in your agent with a single operation.

  • Pause or Stop AI Agent Runs Per Contact: Control AI agent execution at the individual contact level. If an agent is behaving unexpectedly for a specific contact, pause or stop it without affecting other contacts in the workflow.

Why This Matters

Collaborative Agent Building: AI agents are too important to build in isolation. Team edit mode lets your best prompt engineers, product experts, and customer success people contribute to the same agent simultaneously.

Iterate With Confidence: Version history and auto-save mean you can experiment aggressively with your agents. Try a new prompt approach, test it, and roll back in seconds if it doesn't work.

Organized at Scale: Folders keep your agent library manageable as it grows. When you have 50 agents across sales, support, and onboarding, folders are the difference between chaos and clarity.

AI Voice Optimization: Faster Responses, New Models, and Vision

AI Voice Optimization: Faster Responses, New Models, and Vision

Speed and intelligence are everything in voice AI. We've optimized response times, added TTS caching, introduced new voice models, and enabled vision capabilities during live voice conversations — so your AI agents sound better, respond faster, and see what your callers see.

What's New

  • Voice Speed Optimization: Two rounds of voice latency optimization make AI responses noticeably faster. Your AI agents now respond with near-human speed, keeping conversations natural and callers engaged.

  • TTS Caching: Frequently used text-to-speech phrases are now cached, eliminating synthesis delay on repeat playback. Greetings, hold messages, and common responses play back instantly.

  • Vision on Voice Chat: AI agents can now process images shared during voice conversations. A caller can send a photo of a product, document, or issue, and the AI agent sees and responds to it in real time — no switching to a separate chat channel.

  • Call Screening: New call screening functionality lets your AI agent qualify incoming calls before routing them to your team. Screen for intent, urgency, or specific criteria, and only connect the calls that matter.

  • New Voice Models: Added the latest AI voice models with full voice configuration settings. Your AI agents can use the most natural-sounding voices available, with settings that the agent respects throughout the conversation.

  • My Voice Setting: A personal voice selection option in the agent voice dropdown lets you assign specific voices to specific agents. Each agent gets its own identity and personality through its voice.

  • AI Call Summaries: Every AI call now generates an automatic summary available as a workflow variable and displayed on the contact's activity timeline. Your team gets instant context without listening to recordings. Summaries are now generated for both inbound and outbound AI calls.

Why This Matters

Speed Wins Conversations: Faster AI responses mean callers stay engaged instead of wondering if the line went dead. The difference between 2 seconds and 0.5 seconds of response time is the difference between a natural conversation and an awkward pause.

Visual AI Conversations: Vision on voice chat opens up entirely new use cases — insurance claims, tech support, product identification — where seeing what the caller sees changes everything.

Instant Context for Your Team: AI call summaries on every contact mean your team never walks into a follow-up call blind. Read the summary, know the situation, and pick up right where the AI left off.

More Chat Widgets, More Customization: Full Screen, Orb, and Flexible Embedding

More Chat Widgets, More Customization: Full Screen, Orb, and Flexible Embedding

The chat widget system gets a major expansion with new widget styles, full-screen mode, and deeper customization options. Whether you want a subtle orb in the corner, a full-screen conversational experience, or a widget embedded inline within your page — you now have the flexibility to match any use case.

What's New

  • Full-Screen Widget Mode: Launch the widget in full-screen mode for immersive conversational experiences. Perfect for dedicated support pages, onboarding flows, or landing pages where the AI conversation is the main event — not a sidebar distraction.

  • Orb Widget Style: A compact, modern widget design that sits unobtrusively on your page until clicked. Includes a transcript toggle so visitors can review their conversation history without leaving your site.

  • Chat & Voice Toggle: Visitors can switch between text chat and live voice conversation within the same widget. Start typing a question, then switch to voice for a more natural interaction — or vice versa.

  • Target Div Placement: Control exactly where the widget appears on your page by targeting a specific div element. Embed it in a sidebar, a support page, or inline within your content — not just floating in the corner.

  • Widget Customization Options: Configure widget appearance, behavior, and default settings to match your brand. Control colors, positioning, default open/closed state, and which capabilities (chat, voice, or both) are available to visitors.

  • Website Data Context: Data extracted from your website is sent to the widget session, giving your AI agent real-time context about what the visitor is looking at when they start a conversation. A visitor on your pricing page gets a different conversation than one reading your blog.

  • Widget Session Logging: Every widget interaction — chat messages, voice calls, page navigation — is persisted to the database and appears in the contact's conversation timeline. Your team sees the full picture of every visitor interaction.

  • Support Widget in CRM: Your team now has a support chat widget accessible directly from the CRM left sidebar. Get help or access resources without leaving the platform.

Why This Matters

Widget for Every Use Case: Full-screen for dedicated support experiences, orb for subtle engagement, inline for contextual help — pick the widget style that fits each page and scenario.

Brand-Consistent Experience: Customization options ensure the widget feels like a natural part of your website, not a bolted-on third-party tool. Your brand, your look, your conversation.

Context-Rich Conversations: Your AI agent already knows what pages visitors have been browsing. Conversations start with context, not cold intros.

Complete Interaction History: Widget sessions appear in your CRM timeline alongside calls, emails, and SMS. One unified view of every customer touchpoint.

New Workflow Nodes: Conversation, Availability, Content Generation, and Sub-Agents

New Workflow Nodes: Conversation, Availability, Content Generation, and Sub-Agents

The workflow builder just got a massive expansion. New nodes let your automations check calendars, generate content, search knowledge bases, and hand off between AI agents — all without leaving the builder.

What's New

  • Conversation Node: The Talking Node has been renamed and supercharged. The new Conversation Node supports a "stay on node if no route match" option that keeps the AI engaged instead of dropping off. Full text mode support means it works for chat, SMS, and email workflows — not just voice.

  • Check Availability Node: Query your team's calendar availability directly within a workflow. Use it to offer open time slots, validate proposed meeting times, or route contacts differently based on whether your team is available right now.

  • Generate Content Step: The Generate Content step uses AI to create personalized content on the fly within your workflow — follow-up emails, custom SMS messages, or summaries — using contact data and conversation context.

  • Forward to Agent / Sub-Agent Node: Hand off conversations between AI agents mid-flow. Route a sales inquiry to your sales agent, then forward to your scheduling agent when it's time to book — each agent specializes in what it does best.

  • Search Knowledge Base Step: Query your knowledge base from any workflow and use the results in subsequent steps. Output variables are fully exposed in the variable picker, so you can insert knowledge base answers into messages, conditions, or AI prompts.

  • AI-Powered Call Forwarding: The call forward node now supports AI-picked destinations. Instead of static forwarding rules, your AI agent evaluates the conversation context and routes the call to the best available team member or department.

  • SMS AI-Generated Responses: The SMS node now supports AI-generated message content. Define instructions instead of static text, and the AI crafts a contextual response for each contact based on their data and conversation history.

  • Extract Dynamic Variable Step: Enhanced with more data types and detailed info in test mode. Extract structured data from unstructured text with greater precision and visibility into what the AI is pulling out.

Why This Matters

Build Flows That Think: AI-generated content, knowledge base queries, and sub-agent handoffs mean your workflows can handle complex, contextual interactions that static automation can't touch.

Specialization Through Sub-Agents: Instead of building one massive agent that does everything, create specialized agents and let workflows route between them. Better responses, easier maintenance, and clearer analytics.

Schedule-Aware Automation: The Check Availability node eliminates the "sorry, that time is taken" back-and-forth. Your workflows only offer times that are actually open, reducing friction and no-shows.

Workflow Personalization: Smart Automation That Feels Human

Workflow Personalization: Smart Automation That Feels Human

Generic outreach is dead. We've added powerful personalization across your entire workflow builder—from dynamic AI questions to personalized emails that automatically insert contact details. Your automation can now feel as personal as a 1-on-1 conversation.

What's New

  • Contact Variables Everywhere: Use {{first_name}}, {{email}}, {{company}}, or any custom field directly in your workflows. Personalize every touchpoint without lifting a finger.

  • Dynamic Ask & Save Questions: Your AI agent can now ask "Hi {{first_name}}, are you still interested in {{product}}?" instead of generic questions. Works in voice and chat.

  • Personalized Email Automation: The Email Node now supports variables. Write one template, automatically personalize it for every contact: "Hi {{first_name}}, following up on our call about {{topic}}."

  • Smart Calendar Auto-Retry: Booking failed? The Smart Calendar Node now automatically retries with alternative time slots instead of giving up after one attempt.

  • MMS Image Support: Send product images, flyers, or screenshots directly in SMS workflows. Upload once, send to thousands.

  • Gmail Workflow Triggers: Forward emails to your Votel inbox to automatically trigger workflows. Perfect for support tickets, lead notifications, or custom automations.

Why This Matters

Personalization at Scale: Send 1,000 messages that feel like 1,000 individual conversations. Variables let you automate without sounding robotic.

Higher Response Rates: Personalized messages get 26% higher response rates. Now every email, SMS, and AI question is tailored to the recipient.

Smarter Follow-Ups: Auto-retry on calendar bookings means fewer missed appointments. Image support in SMS means richer, more engaging conversations.

Set It and Forget It: Build your workflow once with variables, then let it run. Whether you have 10 contacts or 10,000, the system handles personalization automatically.

Native MCP Server Integration for AI Agents

Native MCP Server Integration for AI Agents

AI Agents are now super-powered with native MCP (Multi-Connect Protocol) support. This allows your agents to connect directly to other MCP-enabled platforms, enabling deep, real-time data synchronization and workflow automation across your entire tech stack.

What's New

  • Native MCP Support: Connect MCP servers directly inside the Agent module's integration settings
  • Cross-App Connectivity: Seamlessly integrate with MCP-enabled platforms, including popular tools like GoHighLevel and CloseGPT Subscriptions
  • Automated Agent Triggers: Use MCP to sync data, trigger external workflows, fetch real-time context from other apps, and more

Why This Matters

Breaks Down Silos: Your AI agent is no longer an island. It can now talk to your other core systems.

Real-Time Context: Agents can fetch live data (e.g., "What's this user's subscription status?") before answering.

Powerful Automation: Trigger complex, multi-app workflows that start in Votel and finish in GoHighLevel (or vice-versa).

Launching the Votel.ai Agent System

Launching the Votel.ai Agent System

This is the big one. We're officially launching the Votel.ai Agent System. You can now build, train, and deploy advanced conversational AI agents with structured logic and deep knowledge.

What's New

  • Prompt Builder: Build advanced conversational prompts with structured logic
  • Knowledge Base: Train agents with documents, FAQs, and custom datasets
  • Agent Profiles: Configure personality, tone, and behavior for each agent

Why This Matters

Go Beyond Chatbots: Build true, task-oriented AI assistants.

Custom Knowledge: Your AI can finally answer specific questions about your business and products.

Brand Voice: Ensure your AI sounds like your brand, whether it's professional, witty, or empathetic.

New: Workflow Builder for Agents

Workflow Builder for Agents

An agent that can talk is one thing, but an agent that can take action is another. Launching alongside the Agent System is the new Workflow Builder, allowing you to connect your AI to automated, multi-step actions.

What's New

  • Multi-Step Workflows: Automate follow-ups, decisions, and internal actions
  • Conditional Branching: Create dynamic flows that adapt to user responses (e.g., "If intent is 'sales_qualified', create a deal in the pipeline")
  • Execution Preview: Review workflow steps and data inputs for accuracy

Why This Matters

Full Automation: Your agent can now do more than just talk—it can qualify leads, update the CRM, and send follow-ups.

Smarter Agents: Create complex logic that guides the agent's behavior based on real-time conversation data.