Produktnyheter inom AI

Data Export Feature for Interfaces

Interface Admins can now export data directly to CSV for easier extraction and analysis.

All services operational

All ERP.AI services (Apps Platform, APIs, Website) reported operational.

Create your own playbooks

Users can build negotiation playbooks capturing positions and fallbacks that Genie applies to every contract review.

Bring code-backed screens onto the canvas with variables attached

Variables now automatically bind to colors, typography, and spacing when importing code-backed screens, reducing manual rebuilding work.

Classification export and period lock fixes

Adds classification export capability and fixes issues with period locking.

Time Off API: policyTypeCustomName on get request details

The Time Off API now returns the policyTypeCustomName field when retrieving request details.

ChatGPT desktop app experience updates

The macOS and Windows desktop app gets a clearer layout with a global Chat/Work/Codex switcher, unified Recents, Projects in-app, and cross-device Work sync.

Pipefy AI Agents: New Lite, Standard, and Pro Models

AI Agent behaviors can now use three different AI models tailored to varying task-complexity levels.

IVR Post-Welcome Workflow Execution

You can now trigger a custom workflow right after the IVR bot plays its welcome message, enabling preprocessing such as user authentication before the bot assists.

Build AI Agents With Agent Studio, Now Generally Available

Agent Studio is now generally available, providing a managed platform for building, deploying, and operating AI agents grounded in Algolia data.

Artemis 1.3.0 (Major Release)

Adds real-time Arch AI feedback, arch-loop evaluation, engine tracing and persistent memory, plus voice call recording, DTMF and Genesys Cloud integrations, SOAP/WSDL support, new knowledge base connectors and workspace-aware enterprise SSO.

Support for Audio File Transcription

Users can upload audio files and receive automatic transcripts inside Harvey.

Dubbing 2.0 is here: a new model, a new bar for quality

An improved end-to-end dubbing system with a new state-of-the-art lip sync model and better translations.

Neural Search Summary Backfill

A weekly automated task now detects and reindexes documents missing summaries so the bot always reflects the most current content.

Increased custom instructions limit

Plus, Pro, Enterprise, Business, and Education users can now save custom instructions up to 5,000 characters, up from 1,500.

Xero OAuth and revenue batch approvals

Introduces Xero OAuth integration and batch approval workflows for revenue.

Thread Experience Update

Threads now support multi-step request planning with parallel task tracking.

Connect AI Agents to External Tools With MCP Servers

Agent Studio now supports Model Context Protocol servers, enabling AI agents to securely connect to external tools and services.

Upcoming System Maintenance – 2026-07-15

Scheduled maintenance window completed, resolved at 7:15am UTC.

Configurable Neural Search Models

Administrators can choose between GPT-5.1 and Llama-3.3-70B for document summarization in tenant settings.

Company Overview: User Stats, Near Real-Time AI Usage, and Export

Gives admins direct visibility into user adoption, license distribution, and AI consumption metrics with instant reporting.

Enhanced User Awareness

Applications now extract parameter values from the user's request and prior conversation history, reducing repetitive prompts.

Contextual Fallback Responses

When a request cannot be fulfilled, bots now add relevant contextual information alongside the existing fallback options instead of static responses.

VibeEdit Now Supports Video Translation and Dubbing

VibeEdit can now translate and dub videos into other languages.

Delete and Update Invoice Payments

Adds API support for modifying and removing existing invoice payment records.

Organize pages with folders and nesting

Group related pages into folders with nesting for flexible page hierarchies.

Updated documentation for User Chats API

The create chat endpoint now supports creating case_task chat sessions.

HIPAA configuration is now self-serve

Organizations can independently manage HIPAA readiness and enable configurations without involving Sales.

Camtasia Windows 2026.1.4

UI improvements for Transition and Quiz properties, performance enhancements, and multiple bug fixes.

What's new at Envato?

A running log documenting recent platform updates, new AI models, workflow improvements, and creator-requested features.

See AI credit usage in betas

Admins can download CSV files tracking AI credit consumption across beta features for better spending forecasts.

Back up Softr Databases with snapshots

Manually snapshot database copies to recover from unwanted changes.

Conditionless Rules Can Now Be Created From the Visual Editor

Merchandisers can now create always-apply rules directly in the Visual Editor, previously only available in the Manual Editor.

Build apps with Snowflake data

Native Snowflake integration (Enterprise) to build dashboards and tools on warehouse data.

Statistical journals and fixed asset accounts

Adds support for statistical journals and fixed asset account handling.

Supercharged search across chats, projects, and files

A new unified search on web, iOS, and Android lets users find past chats, projects, images, and documents from one place with content-type filters.

Website downtime

The website experienced a downtime incident lasting about 20 minutes.

DocSearch Hub Centralizes DocSearch Management in the Dashboard

A dedicated DocSearch Hub consolidates crawler access, indices, API keys, and implementation guidance in one dashboard location.

Access your Drafts from the Calendar

Draft posts can now be located and accessed directly from the Calendar view.

New Feature: AI Visibility (AI Search Citation Tools)

New AI Visibility suite (Audit, Answer Check, Query Fan-Out Report) shows where Google AI Overviews and ChatGPT cite your site and where you're missing citations.

GPT-5.6 Sol Now Available in Harvey

The new GPT-5.6 Sol model is selectable in Harvey's Model Selector.

ChatGPT returns to WhatsApp in the EEA

ChatGPT is available again on WhatsApp in the European Economic Area via the verified 1-800-CHATGPT contact, with optional account linking for higher limits.

AI Copilot reports, QuickBooks reconciler

Ships AI Copilot reporting and a QuickBooks reconciliation tool.

Test Move Card Automations Before Publishing

Simulates automation execution against real configurations to verify behavior without affecting production data.

Address validation and attachment overrides

Adds address validation and the ability to override attachments.

design_types and design_type fields now generally available

The design_types field on Design responses and the design_type field on DesignPage moved out of preview and are now included in every response.

Test Fin for Sales at scale

A new simulation feature lets teams bulk-test Fin for Sales playbooks before deployment.

Updated memory for Claude

Memory now works as individual categorized entries Claude reads and updates during conversations rather than daily summaries.

Export your macro content to CSV

Macro content can now be exported to CSV for auditing and backup.

45+ New Legal Research Sources

Harvey expanded its available legal research data sources by more than 45.

Audio models list endpoint / Simba model family

Changelog entry documenting the GET /v1/audio/models endpoint and the available Simba family of text-to-speech models.

Seedream 5.0 Pro

New Seedream 5.0 Pro AI image-generation model added to the AI Toolkit.

GPT-5.6 is now in Figma Make

OpenAI's latest model is available in the Figma Make chatbox across all subscription tiers.

A new way to reflect with Claude

A monthly recap surfaces conversation topics, activity patterns and work insights, alongside new break reminders and quiet hours.

Zapier MCP Connector

Fin can execute Zapier actions during conversations without custom code via a new MCP connector.

More ways to snooze and see what's coming back

Enhanced snooze functionality including macro-based snoozing and bulk operations.

BYOM: Connect Your Own Claude Models from Anthropic

Anthropic is now available as a Bring-Your-Own-Model provider, letting users connect their own Claude models via API key.

Introducing ChatGPT Work

ChatGPT Work is an agent for longer tasks that researches, works across connected apps and files, and produces finished documents, rolling out to paid plans.

What's New in June 2026: Close the Loop on AI Search

Introduces GEO Hub for tracking brand visibility across AI search engines, a Translation Agent for 27 languages, enhanced Grid scheduling, and Monday.com/Asana approval-workflow integrations.

Filter Journal Entries by GL Account

Adds the ability to filter journal entries by general ledger account.

API: New simba-3.2 streaming model (recommended)

Speechify released simba-3.2, a streaming-native text-to-speech model with improved latency and expressivity, now the recommended option for new English integrations.

Artemis 1.2.2 (Patch Release)

Adds public-key JWE authentication for the Hosted Exchange Web SDK, enabling secure authentication with customer-managed public keys.

Translate side conversations automatically

Side conversations now support automatic translation between languages for cross-team collaboration.

Advances in AI, Smarter Search, Richer Content, & More: What's New with Signal AI

Quarterly platform roundup announcing Signal AI MCP integration, AI Citations tracking, expanded content coverage, DeepL translation support, Search 2.0, and Risk Scanner.

Conversational search using the Coveo Search Agent

Conversational search via the Coveo Search Agent is now generally available, using agentic capabilities to orchestrate multiple rounds of retrieval and answer generation from user follow-up questions.

Assistant Anywhere

The Assistant now operates as a persistent sidebar throughout the Ivo web app, accessible without leaving the current page.

Vespa 8.719.5

Latest tagged Vespa/Vespa CLI release for managing local, self-hosted and cloud Vespa deployments.

Turn any video into a Scribe

Automatically converts uploaded videos into step-by-step documentation.

Saved Custom Views

Users can preserve customized content filters as saved Views in the sidebar for quick reference.

Comment-Aware Review

Review can now analyze counterparty comments in documents and give recommendations based on comment threads.

Seedance 2.0 Family

New Seedance 2.0 family of AI video-generation models added.

Generate PowerPoint and Excel Files

Assistant can create formatted PowerPoint presentations and Excel spreadsheets with formulas directly from conversations.

Playbook Builder: Non-English Playbooks

Playbook Builder now accepts language specifications, enabling playbooks in languages other than English.

Claude Cowork on web and mobile

Cowork expanded to web and mobile with remote sessions and a shared project interface with Chat.

Run multiple AI image edits and keep designing while they load

Users can now generate several AI image edits simultaneously without waiting for each to complete.

Write tools for the Microsoft 365 connector

The connector now lets Claude draft/send email, manage calendar events and update files in OneDrive/SharePoint.

Security: DNS Verification for SSO

Workspace admins can verify domain ownership via DNS as part of single sign-on configuration.

Developer logs for the Interactions API

Users can now view Interactions API call logs in the AI Studio dashboard.

New in Kittl: Agentic AI

Introduces an AI agent that automates design decisions, reducing manual choices about prompts, models, and formats.

July 2026 - Release note

Suggested Contacts filter panel expanded from 4 to 15 filters with new include/exclude operators and criteria like years of experience, certifications, languages, and vacancy stage.

Interactions API developer logs

Developer logs for the Interactions API are now available in the AI Studio dashboard.

Envato MCP Server: A new way to search Envato via your LLM of choice

A new MCP server connects AI assistants like ChatGPT and Claude to the Envato library for real-time, structured asset search.

VibeEdit Now Supports Automatic Text and Emoji Overlays

Automatic text and emoji overlays were added to VibeEdit's editing pipeline.

New AI Models: Claude Sonnet 5 & Claude Fable 5 (DeepPolish+)

Added two premium Anthropic models with full DeepPolish+ treatment across Advanced, Webpage, Rewrite, and Optimize modes.

Customise Grid Layout so That Your Preview Matches Your Storefront or Site

The Visual Editor now supports custom grid layout configurations, letting merchandisers preview results in the tile pattern used on their storefront.

Read-aloud for AI Summary

Visitors can now listen to AI Summary responses with a built-in play/pause button for more accessible consumption.

Runway Agent Skills

New agent skills let users build ad campaigns, create commercials, and localize ads.

Brain² takeover

A full ClickUpdates dedicated to Brain², the rebuilt AI that self-improves, routes across Claude/Gemini/GPT models, generates slides, dashboards and apps, adds cross-conversation memory, a Brain MAX mobile app, autonomous Super Agents, and MCP integrations.

Runway Agent Skills

New agent skills let users build an ad campaign, create a commercial, and localize ads with a simple command.

Apps are here: connect the tools you already use, right inside Kittl

Users can now integrate external tools like Pinterest, Dropbox, and Printful directly within Kittl's interface.

Batch APIs for Translations, Lipsyncs, and Assets

High-volume workflows now support native batch processing for video translations, lipsyncs, and asset uploads, accepting up to 100 items per request.

Product Newsletter — June 2026

Monthly roundup of Leena AI product developments and releases from June 2026.

Voice content creation, enterprise workflows, and data warehouse queries

Create content with voice, scale enterprise-ready workflows, and query Snowflake and Databricks data with full business context.

Nano Banana 2 Lite

Image generation model now available on paid plans.

Bob MCP Server: Now uses OAuth with expanded tools

The Bob MCP Server was upgraded to OAuth authentication with additional tool capabilities.

Organize your teams into Departments

Lets administrators group teams into organizational units for better structure.

HeyGen Studio Template API

Developers can now programmatically discover Studio templates, inspect their variable schemas, and render videos by populating template variables.

Nano Banana 2 Lite

Enables rapid image generation while maintaining quality across all paid tiers.

Batch Video Creation

The API now permits submitting multiple videos in a single request that accepts up to 100 payloads and returns one batch identifier.

Website outage

The website had a significant outage lasting 2 hours 16 minutes.

7/2026 'Interval'

Upgraded the Persona editor to support Markdown and stopped conversation starters from interrupting active chats.

July 2026 Release Notes

New features across Channels, Conversations, Salesloft Connect, Clari AI, and Data & CRM.

AI Transcripts

More accurate AI-generated transcripts now auto-generate when videos finish processing.

AMA on Reddit all about Brain²

ClickUp ran a Reddit Ask-Me-Anything with a product lead about Brain², the ground-up rebuild of ClickUp Brain, covering how it handles autonomy vs. human control and what early-user feedback changed.

Gemini Omni Flash public preview

Omni Flash enters public preview for high-speed video generation and conversational video editing, supporting 3-10 second clips at 720p.

Early access: Gemini Omni Flash live in Synthesia

Google's Gemini Omni Flash model is now available in public preview for video generation and editing in the platform.

Gemini Omni Flash public preview

A high-performance model for fast video generation and conversational editing from text or images entered public preview.

Gemini Omni Flash

Users can now produce and modify videos using prompts, images, or existing video content.

Gemini Omni Flash

Added video generation and editing capability.

Nano Banana 2 Lite

Nano Banana 2 Lite AI image model added to the catalog of available models.

Build apps with Microsoft Excel data

Use Excel spreadsheets as a native data source for portals and dashboards.

Send push notifications

Send custom push notifications to logged-in users via workflows on desktop and mobile.

Claude Sonnet 5, Now Live in Harvey

Anthropic's Claude Sonnet 5 model is now available for use within Harvey.

Release log 3.51 (29.6.2026)

Introduces Dashboard 2.0 with personalized customization and Attendance 2.2 with a complete approval workflow, plus usability improvements.

New preview design analytics APIs

Added preview APIs to retrieve engagement metrics for a shared design, including views over time, page views, viewers, and links.

Post Composer Improvements

A bundle of three enhancements to streamline post creation and reduce friction.

Motion Graphics: from script to animated visuals, no design expertise needed

Users can create branded animated infographics directly in Synthesia without separate design tools.

Seed Audio 1.0

Generate up to 120 seconds of dynamic speech, sound design, and music from text prompts.

Seed Audio 1.0

Generate up to 120 seconds of dynamic speech, sound design and music.

Website downtime

The website was down for approximately 22 minutes.