DAPPOS's xBubble Aims to End Prompting with 'AI That Uses AI'

SINGAPORE – May 12, 2026 – AI firm DAPPOS today launched xBubble, a new platform designed to fundamentally change how users interact with artificial intelligence. Backed by over $20 million from prominent investors including Polychain Capital, Binance Labs, and Sequoia China, the company is betting that the future of AI isn't about teaching humans to speak to machines, but about teaching machines to manage themselves.

xBubble is a 'low-prompt' AI agent that promises to deliver professional results without the steep learning curve of prompt engineering. Instead of presenting users with a blank text box and a powerful but inscrutable model, xBubble acts as an intelligent dispatcher, automatically building and deploying task-specific AI agents to get the job done. The company's goal is to close the widening 'usability gap' between AI power users and the general public, making advanced capabilities accessible to anyone with a clear goal.

“Powerful AI no longer requires users to learn AI,” the DAPPOS team stated in their announcement. “xBubble inverts the relationship. We have AI learn AI, and we have AI use AI, so users don’t have to.”

The End of the Prompt-Tuning Era?

The central problem DAPPOS aims to solve is a familiar one for many who have dabbled in generative AI. The same model that produces breathtaking results for an expert can yield disappointing, generic output for a novice. This disparity has turned 'prompt engineering' into a crucial, almost arcane skill, requiring users to study model behaviors, test tool combinations, and constantly relearn processes with every new update.

xBubble proposes a radical departure from this model. Its architecture is built on two core systems: Bubble Engine and Bubble Pilot. The Bubble Engine acts as a behind-the-scenes 'solution factory.' For any given task, such as creating a marketing poster or drafting a research report, it uses AI coding agents to generate and test numerous potential workflows. It combines different AI models and tools, evaluates their outputs against quality criteria, and ultimately establishes the most effective method as a 'Standard Operating Procedure' (SOP).

The Bubble Pilot is the user-facing dispatch layer. When a user makes a request, such as “Create a slide deck about the history of renewable energy with relevant images and speaker notes,” Bubble Pilot interprets the intent. It then checks for a pre-built SOP from the Engine. If a match exists, it executes the highly optimized, pre-tested workflow. If not, it deploys a general-purpose agent, and the request pattern is fed back to the Bubble Engine as a candidate for a new, future SOP. This creates a self-improving ecosystem where the system gets smarter and more efficient with use.

A Two-Pronged Attack on Complexity

xBubble is being introduced with two distinct modes designed to integrate into different parts of a user's digital life: Bubble Computer and Bubble Personal.

Bubble Computer is positioned as an end-to-end project workspace. When the Bubble Pilot detects a complex, multi-step task, it routes the project to this mode. Here, a secure sandbox environment is created, and specialized AI agents are loaded on demand. Within a single session, xBubble can perform deep research, draft documents, generate accompanying visual assets, and deliver a polished final output from a single, high-level user goal. The system handles the intricate coordination of selecting the right model for each sub-task, routing data between them, and assembling the final product.

Bubble Personal operates within the user's local environment, automating tasks across files, browsers, and applications. It's designed to act as a personal assistant for operations that require access to personal accounts or local data. Examples include generating a morning briefing based on your calendar and inbox, organizing photo libraries, or scheduling a nightly task to collect and summarize market data. To address security concerns, DAPPOS has implemented a sandboxed execution model. Heavy computation and potentially risky operations are handled in isolated cloud containers that are destroyed after the task is complete, with only clean, authorized results being sent back to the user's machine.

Supported tasks at launch are extensive, including Voice Dictation, Text to Speech, Talking Avatar creation, Deep Research, Slides and Docs Creation, Fact Checking, Poster and Image Creation, and even basic Website Development.

Navigating a Crowded and Autonomous Future

DAPPOS enters a fiercely competitive AI agent market where companies are racing to define the next generation of human-computer interaction. Its 'low-prompt' philosophy and automated SOP generation system are its key differentiators in a field populated by developer-centric frameworks like LangChain and increasingly agent-like consumer products such as OpenAI's custom GPTs. The company's significant financial backing signals strong investor confidence that usability, not just raw power, will be the deciding factor in winning the market.

The launch of systems like xBubble is part of a broader, more profound industry trend toward 'meta-AI'—AI that builds, manages, and deploys other AI. This paradigm shift promises an explosion in productivity and accessibility, but it also surfaces complex ethical and societal questions that experts are actively debating. While democratizing access to powerful tools is a clear benefit, the concentration of such orchestration power in a few platforms could risk a new form of technological centralization.

Furthermore, the increasing autonomy of these systems brings challenges of control, accountability, and transparency to the forefront. When an AI system autonomously chains together multiple other AIs to produce a result, determining responsibility for errors or biases becomes exponentially more difficult. Ensuring that these complex, self-directed systems remain aligned with human values is no longer a theoretical exercise but a pressing design challenge for companies like DAPPOS. As AI continues its rapid evolution, the industry will be watching closely to see if this new generation of autonomous agents can deliver on its promise of effortless power without introducing unforeseen risks.