What\u2019s new in AI tools
Automated weekly roundup of release notes and changelogs across every AI tool we track. Currently indexing 317 updates.
This week \u2014 2026-W18
AI Tools This Week: 42 updates across 7 products
-
April Drop 2026
Added slide and interactive HTML page generation features, plus embedded MCP Apps in Glean Assistant for cross-tool actions.
-
Agent Versioning 2.0, Granular Multilingual, Custom Voicemail, Voice Orb
Released improved agent versioning with staging/production environments, multi-language support with automatic detection, custom voicemail handling, and shareable voice agent links.
-
Live Call Monitoring, Custom Dashboards, Built-in CRM, Colloquial Model, Expressive Mode
Retell AI launched five major features including real-time call monitoring, CRM integration with Salesforce/HubSpot, customizable dashboards, and enhanced voice agent naturalness through Colloquial and Expressive modes.
-
Groq Expands to Asia-Pacific with Sydney Data Center
Groq launched a new data center in Sydney, Australia to serve the Asia-Pacific region with AI inference capabilities.
-
Groq Partners with U.S. Department of Energy
Groq established a partnership with the U.S. Department of Energy to advance AI inference and next-generation computing infrastructure.
-
Groq and Nvidia Enter Non-Exclusive Inference Technology Licensing Agreement
Groq partnered with Nvidia through a licensing agreement to accelerate AI inference capabilities at global scale.
-
GroqCloud: Expanding to Meet Demand
Groq announced expansion of its GroqCloud platform to accommodate growing user demand for AI inference services.
-
Groq Raises $650M to Scale Its AI Inference Cloud Business
Groq secured $650 million in funding to expand its AI inference cloud infrastructure and business operations.
-
Mission Big Dipper brings agentic execution to revenue teams
Gong launches Mission Big Dipper, introducing the Gong Revenue Harness as an agentic execution layer that enables AI agents to work alongside humans in revenue workflows. The update includes Custom Agents that can be built without code, expanded Gong Assistant capabilities, and new Gong Enable features like Dry Run for realistic role-play practice.
-
Bringing more control over your connectors
Enhanced connector management capabilities provide users with greater control and flexibility.
-
Leanstral 1.5: Proof Abundance for All
Leanstral 1.5 introduces enhanced proof capabilities for mathematical and logical reasoning.
-
Your Prompts and Skills need a system of record
Studio now provides versioned, owned, and traceable management for AI prompts and skills.
-
Introducing Shieldstral
Mistral AI launches Shieldstral, a new AI safety and security solution.
-
Share context with Custom Agents from the Share menu
Users can now share documents or databases directly with Custom Agents from the Share menu without navigating to agent settings.
-
Copilot usage metrics API adds agent app activity
The Copilot usage metrics API now includes tracking for agent app activity and usage patterns.
-
Copilot code review effort levels are generally available
Code review effort levels feature in GitHub Copilot has moved from preview to general availability.
-
Copilot impact dashboard adds a return on investment section
The Copilot impact dashboard now includes a dedicated section for tracking return on investment metrics.
-
GitHub Copilot weekly releases — August 3
Weekly release update for GitHub Copilot containing various improvements and fixes.
-
Kimi K3 from Moonshot AI is now available on Databricks through Unity AI Gateway
Moonshot AI's Kimi K3 model is now accessible through Databricks Unity AI Gateway. This expands the available AI models for enterprise users on the platform.
-
Introducing OfficeQA Pro V2: A New Benchmark for Enterprise Grounded-Reasoning
Databricks released OfficeQA Pro V2, a new benchmark designed to evaluate enterprise grounded-reasoning capabilities. The benchmark focuses on testing AI models' ability to reason with enterprise data.
-
MCP allowlists in enterprise managed settings
Enterprise administrators can now configure MCP allowlists through managed settings for better control.
-
Databricks joins the Open Secure AI Alliance to advance AI safety and security
Databricks joined the Open Secure AI Alliance as a sponsor to advance AI safety and security initiatives. The company is participating in Black Hat USA 2026 as part of this commitment.
-
Unity AI Gateway is Generally Available
Databricks has made Unity AI Gateway generally available after six months of rapid AI-powered productivity growth. The gateway provides centralized access to AI models and services.
-
Ingest semi-structured data faster and more efficiently with Variant - Now Generally Available
Databricks made Variant generally available for ingesting semi-structured data like JSON, XML, and CSV. The feature addresses previous difficulties in handling these data formats efficiently.
-
June Drop 2026
Introduced meeting notes functionality that converts meetings into searchable context and interactive content for team collaboration.
-
July Drop 2026
Added voice content creation capabilities, enterprise workflow scaling features, and Snowflake/Databricks data querying with business context.
-
Agent Versioning 2.0, Granular Multilingual, Custom Voicemail, Voice Orb, and more
Enhanced agent management with version control, multi-language support with automatic detection, custom voicemail handling, and shareable agent links.
-
Live Call Monitoring, Custom Dashboards, Built-in CRM, Colloquial Model, Expressive Mode, and more
Major platform update introducing real-time call monitoring, customizable dashboards, native CRM integration with Salesforce and HubSpot, and enhanced voice agent capabilities with more natural speech patterns.
-
Prompt Caching
Prompt caching automatically reuses computation from recent requests when they share a common prefix, delivering 50% cost savings for cached portions and improved response times. The feature works automatically with no code changes required.
-
Moonshot AI Kimi K2 Instruct 0905
Kimi K2-0905 brings Moonshot AI's cutting-edge model to GroqCloud with day zero support, featuring a 256K context window and enhanced agentic coding capabilities. The model offers prompt caching with up to 50% cost savings and runs at 200+ tokens/second.
-
Leanstral 1.5: Proof Abundance for All
Mistral AI released Leanstral 1.5, an updated version of their lean model focused on proof generation capabilities.
-
AI Meeting Notes can now trigger Custom Agents
Custom Agents can automatically run after AI Meeting Notes are completed, enabling automated workflows like updating project trackers and posting recaps to Slack.
-
AI Meeting Notes can now trigger Custom Agents
Custom Agents can now automatically run after AI Meeting Notes are completed, enabling automated workflows like updating project trackers, posting recaps to Slack, or creating tickets.
-
Enterprise teams model policy targeting in public preview
Enterprise teams can now target specific model policies in public preview.
-
High contrast mode
New accessibility display option that makes text, icons, and borders easier to read, available on desktop and web.
-
Convert proprietary code to open ANSI SQL with Genie Code
Databricks introduces Genie Code, a tool that helps migrate legacy data warehouse code to open ANSI SQL. This addresses the complexity of migrating from proprietary data warehouse systems.
-
High contrast mode
Added a new accessibility display option that makes text, icons, and borders easier to read. Available on desktop and web through Settings > Preferences > Appearance.
-
Say hi to new Ojai in-car experience
Waymo unveiled new in-car features for the Ojai vehicle including Gemini in Waymo and a redesigned user interface. The Ojai is designed as a premium ride-hailing vehicle with a flexible cabin space.
-
Say hi to new Ojai in-car experience
Waymo unveiled new in-car features for the Ojai vehicle including Gemini AI integration and a redesigned user interface. The Ojai is designed as a premium ride-hailing vehicle with a flexible cabin space.
-
Say hi to new Ojai in-car experience
Waymo introduces a new in-car experience called Ojai for riders.
-
Agents for production lines: Trusted decisions in real time
Databricks introduces AI agents designed for manufacturing production lines to make trusted real-time decisions. The solution helps line managers respond quickly to production issues.
-
Bringing real-time fraud prevention to government benefits
Databricks announces capabilities for real-time fraud detection and prevention in government benefits programs. The solution addresses improper payments that cost federal benefits programs significantly.
-
Copilot code review: Agent skills and MCP now generally available
Agent skills and Model Context Protocol (MCP) for code review are now generally available.
-
March Drop 2026
Launched remote MCP server to integrate ChatGPT and Claude with company context, and expanded Assistant and Agents with third-party MCP tools.
-
April Drop 2026
Added slide and interactive HTML page generation features, plus embedded MCP Apps integration within Glean Assistant for cross-tool actions.
-
May Drop 2026
Enhanced Assistant with real-time voice capabilities, company-aware Skills, and expanded action capabilities across existing tools.
-
June Drop 2026
Introduced meeting notes functionality that converts meetings into searchable context and enables interactive content creation for teams.
-
July Drop 2026
Added voice content creation capabilities, enterprise workflow scaling features, and integration with Snowflake and Databricks for contextual data querying.
-
Agent Guardrails, Webhook Testing, Auto-Failover, Billing Transparency
Security improvements with jailbreak protection and harmful content filtering, enhanced webhook testing capabilities, and improved billing transparency features.
-
ChatGPT app, Dynamic voice speed, A/B testing
Integration with ChatGPT for building voice agents, adaptive voice speed matching caller pace, and A/B testing capabilities for comparing agent performance.