[Press Release] Qualiteg to Exhibit at Startup JAPAN EXPO 2026 — AI Agent Capability Comes to Bestllam®

April 13, 2026 Press Release

Press Release: Qualiteg to Exhibit at Startup JAPAN EXPO 2026 (PR TIMES, in Japanese)

Official press release on PR TIMES announcing Qualiteg's Startup JAPAN EXPO 2026 exhibit and Bestllam's AI agent capability.

PR TIMES株式会社Qualiteg

AI Agent Capability Comes to Bestllam®

— Delegate in parallel. Never stop thinking. Toward "AI that works," woven into the business systems of Japanese companies — Generative AI adoption / AI agents / business automation / consulting

Qualiteg, Inc. (headquartered in Chiyoda-ku, Tokyo; CEO: Tomonori Misawa) will exhibit at Startup JAPAN EXPO 2026 (booth 16-16), held at Makuhari Messe from Wednesday, April 15 to Thursday, April 16, 2026.

Our flagship product Bestllam® now ships with a new AI Agent capability. At the exhibition, visitors can try a hands-on demo of the new UX: the moment you make a request, a dedicated team springs into action and processes multiple tasks automatically, in parallel.

We will also be accepting inquiries for our AI business-improvement consulting, aimed at companies that want to use AI but cannot commit to company-wide adoption due to governance concerns, or that have adopted tools without seeing real efficiency gains.

■ Startup JAPAN EXPO 2026: Event Overview

■ Main Exhibits

Exhibit 1: Live demo of Bestllam®, the integrated AI platform

Highlight 1 [New] AI Agent | The moment you give an instruction, your own dedicated team gets moving. The work gets done while your thinking never stops.

While business use of generative AI is spreading rapidly, the AI chat experience remains bound to a turn-based structure: the human waits for the AI's reply before thinking about the next step. It is a paradoxical problem — the more you use AI, the more your thinking gets fragmented.

What if you could simply fire off instructions as they occur to you, and find the work already done before you knew it? Bestllam delivers exactly that experience. The moment you give an instruction, your own dedicated team takes shape. Each sub-agent works in parallel, and a supervising agent assembles the results and delivers them to you.

Work moves forward as if you had become an executive with a talented team at your disposal. Because you can stack new instructions without waiting for the AI to finish replying — even mid-response — your work proceeds without interrupting your train of thought.

Three kinds of value delivered by the Bestllam AI Agent

1. Delegate | Ask, and the work gets done You simply hand over a request, and a group of sub-agents advances the work asynchronously behind the scenes. Research, document creation, data organization, email drafts, internal knowledge search — multiple tasks complete in parallel.

2. Never stop | Send the next request without waiting for a reply A supervising agent takes in every request, and multiple sub-agents behind it process them in parallel. You can stack instructions in whatever order they occur to you, without breaking your flow of thought. This extends the parallel-processing architecture Bestllam has long cultivated for running multiple LLMs simultaneously into an agent-execution platform.

3. Stay protected | Data-leak detection and masking with LLM-Audit™ The risk of confidential or personal information slipping into an autonomously acting AI is handled by Qualiteg's own LLM-Audit™, which detects and masks it on both input and output. For government agencies and municipalities, we also offer a "cross-border protection" option that keeps data inside the country.

• Complex research and business tasks completed automatically through a single point of request. Full operation logs and audit readiness via LLM-Audit™ integration
• Mid-course changes welcome. Additional instructions and changes of direction are handled in real time while keeping the whole effort consistent
• Proprietary MCP connectors already built for system constraints specific to Japanese companies (attendance management, core systems, and more)

👉 Learn more

Highlight 2: The Bestllam foundation | With 20+ AI models available, shadow AI disappears — and agents get stronger.

"Shadow AI" is drawing attention as a growing corporate risk. The AI you want is not approved internally. By the time approval comes through, a newer model is out. Each department has a different ideal AI, yet the company can only contract for one — under these conditions, companies keep reporting cases of employees signing up for AI tools personally and using them for work.

With Bestllam®, a single contract gives you more than 20 LLMs — GPT, Claude, Gemini, Grok, Qwen, DeepSeek, and more — to use as needed. By building the capacity to meet the latest models and each department's differing needs in-house, it helps reduce the risk of shadow AI emerging.

This foundation for running multiple LLMs in parallel powers the agent capability, and it also amplifies Bestllam's value as a conventional AI chat tool. Send the same question to several LLMs at once and compare their answers — decision-making informed by each model's strengths and weaknesses, and reliability gains from cross-checking answers, become a natural part of daily work. Let the agent auto-select the best model and boost accuracy by running up to six models simultaneously; in chat, weigh multiple models' answers side by side. The same foundation supports both use cases.

As for the data-leakage risk that comes with agents acting autonomously, our security product LLM-Audit™ steps in as a guardian agent*. It detects confidential and personal information upstream — before anything is sent to an LLM — and blocks, masks, or pseudonymizes it, preventing leaks to external LLMs. The detection engine behind it is PII-Fi™, our in-house foundational technology. Beyond generic personal information, it can be pre-trained on each company's own definition of "what counts as confidential here," enabling highly accurate detection tuned to that company's context.

• 20+ LLMs under a single contract. Usable for both agents and chat, and helps curb shadow-AI risk
• Simultaneous multi-LLM questioning: send the same question to multiple models in parallel and compare their answers
• LLM-Audit™: prevents leaks to external LLMs through detection, blocking, masking, and pseudonymization of confidential and personal information*
• PII-Fi™ (foundational technology): can be pre-trained on company-specific confidential information in addition to generic PII. A fast, high-accuracy detection engine

*LLM-Audit™ is an optional feature.

👉 Learn more

Highlight 3: AI content generation | Images, documents, even video — all from a single piece of text.

Hunting for illustrations for a deck, cleaning up report formatting, building presentation slides — how much time have these production tasks consumed until now? Bestllam generates a wide range of content on the spot from a single piece of text.

AI image generation comes standard as a core Bestllam® feature. Just say "create a product illustration in our corporate colors," and it instantly generates on-brand imagery. AI also auto-generates Word, Excel, and PowerPoint documents, dramatically streamlining routine work and report writing.

• AI image generation: product images, document illustrations, and social media assets, created instantly from a single piece of text
• AI auto-generates Word, Excel, and PowerPoint documents — dramatically streamlining routine work and report writing
• AI avatar video generation: an option connects to MotionVox®, our AI avatar video service. Produce video content with no need to show your face or use your voice. Supports 27 languages

Where Bestllam stands — an AI agent that reaches Japan's corporate front lines

Agent-style UX is already becoming the global mainstream. Building on that, Bestllam aims for a form that truly blends into the day-to-day operations of Japanese companies — connectivity to core systems and internal knowledge bases (MCP and more), of course, but also interactions as natural as collaborating with a human colleague, controls you can use without hesitation, and behavior you can trust with delegated work. We weave into every corner of the product Japan's distinctive sense of omotenashi — anticipating and picking up on the other party's intent before being asked.

And that omotenashi does not stop at the product. Our consulting team walks alongside you from PoC through operational design, on-site training, and business-process redesign. With product and people working in tandem, we take responsibility for the "final step" — until AI genuinely works on your front lines.

Exhibit 2: AI business-improvement consulting (now open to the public, formerly invitation-only)

We offer hands-on, practice-grounded consulting for companies facing challenges like these:

• "We introduced generative AI, but it's just chat — it isn't improving our operations"
• "It ended at the PoC; we haven't been able to commit to company-wide rollout"
• "Concerns about governance and data leakage have kept us from deciding on company-wide adoption"

We provide end-to-end support from PoC through operational design, on-site training, and business-process redesign. A unified team of AI engineers versed in the state of the art and veteran strategy consultants with years of experience driving business transformation at large enterprises delivers field-first, hands-on support that refuses to let AI adoption end at "we installed it."

👉 Learn more

■ Booth Visitor Offers

🎁 Exclusive offers for booth visitors who apply during April

• AI business-improvement consulting: 20% off
• the total of your first contract. Bestllam® initial setup fee: 20% off

*Applications accepted at booth 16-16. Applies to contracts signed by Thursday, April 30.

■ About Qualiteg, Inc.

Qualiteg, Inc. is an AI company founded in July 2023. Under our guiding principle of "deepening the creativity of people around the world," we aim to use cutting-edge AI to remove technical barriers, language barriers, and physical constraints, unlocking the creative potential that everyone possesses.

We are working toward a world where human creativity is amplified and new forms of expression and value — previously unimaginable — come to life. The march of innovation from AGI to ASI and on to digital humans will enable natural collaboration between AI and people, promising a future where anyone can intuitively harness AI to dramatically deepen their own creativity. We are building a society free of divides caused by technology gaps, where everyone can set their creativity free.

Our Services

• The integrated AI platform that accelerates your business Bestllam®
• Generate realistic AI avatar videos from text, audio, and more MotionVox®
• Audit LLM inputs and outputs and defend against data-leakage risk LLM-Audit™
• Detect and mask personal and confidential information with speed and precision PII-Fi™ API
• Build chatbots and inference environments on open-source LLMs ChatStream®
• AI, DX, strategy, new business, and open innovation consulting

■ Inquiries

Qualiteg, Inc. Email: startupjapan2026@qualiteg.com

👉 Contact us / request a demo