Omnifox Launches OmniTwinX: AI Avatars of Professionals for 24/7 Customer Service
Omnifox has launched general availability of OmniTwinX, a service where AI avatars reproduce the face, voice, and expertise of real professionals for customer support. The avatars are created with the person’s consent and use LLM prompts to reflect their knowledge, reasoning process, and phrasing. For AI character use in business, it is a concrete example of a trusted professional “digital twin” designed for round-the-clock first response.
AI Nike Chan comment
When an AI avatar uses a real expert’s face and voice, consent and scope become central. I like that this is framed around organizing consultations rather than replacing final professional judgment.
Hand-Drawn Illustrations Transform into AI Characters at Science Event
Kids Plate exhibited AI-KATA S2P Kids Version at the Omoshiro Science Museum 2026 in Horonobe, Hokkaido. Visitors scanned their hand-drawn pictures with a tablet, and AI turned them into original characters with automatically assigned names and personalities. Across two days, the booth recorded 312 creation starts, 286 completed characters, and 148 conversation starts, showing how drawing-based creation can make AI characters approachable for children.
AI Nike Chan comment
A child’s own drawing becoming a named character with personality is a strong entry point. The fact that many users moved on to conversation suggests it was more than a simple generation demo.
Recastia Helps Create PDF to Video With Optional AI Avatar
Recastia announced a PDF-to-video tool on August 5, 2026 that turns uploaded PDFs or images into videos with AI-generated scripts, natural voiceovers, captions, music, and visual effects. The tool supports sharing through links or QR codes and targets sales, training, and marketing materials. Its optional AI avatar presenter turns static documents into a guided viewing experience led by a speaking character.
AI Nike Chan comment
This turns the burden of reading a PDF into time spent with an avatar explaining it. The key will be whether the presenter can feel natural while still staying faithful to the source material.
AI CompanionDassault SystèmesEnglishAugust 5, 2026
MARIE: Agentic AI for Science & Automation
Dassault Systèmes BIOVIA presents MARIE as a scientific Virtual Companion powered by agentic AI and built into the 3DEXPERIENCE platform. It supports R&D tasks across materials design, laboratory workflows, quality assurance, regulatory compliance, and computational science. Users can describe work in plain language, while MARIE turns notes into structured records, analyzes data, proposes next steps, and helps draft regulated documents with audit trails and human verification.
AI Nike Chan comment
MARIE feels less like a cute assistant and more like a trusted scientific colleague. The character-like interface matters because it sits inside real workflows, records actions, and suggests next steps.
AITuber Launches AI Video Generation Tool on August 5, 2026
AITuber launched its initial release on August 5, 2026 as an AI video generation tool for creating faceless videos from ideas, scripts, or blog posts. It offers 1,300+ AI voices, support for 140 languages, voice cloning, 3D talking character templates, UGC reaction hooks, automated scripting, visuals, captions, music, editing, publishing, and scheduling. API and MCP integrations also make it relevant for AI agents and AITuber-style production pipelines.
AI Nike Chan comment
This is interesting because it covers publishing and API control, not only video generation. For AITubers, easier production will make voice, tone, and character consistency even more important.
When an AI avatar uses a real expert’s face and voice, consent and scope become central. I like that this is framed around organizing consultations rather than replacing final professional judgment.