Bloomberg unveiled ASKB: agentic AI built directly into Terminal for finance teams

Bloomberg demonstrated ASKB live — a new layer of agentic AI built directly into Terminal. The presentation took place at the Bloomberg Family Office Summit 2026 in Hong Kong and showed that AI in finance is moving from chat windows to embedded work tools.

What was shown

According to Bloomberg's description, Kevin Shiky demonstrated ASKB as part of the Bloomberg Terminal itself, not as a separate experimental service alongside the main platform. This is an important nuance: the user doesn't need to switch between an external assistant and a window with market data. AI appears where work with quotes, news, analytics, and tool search is already happening. For a professional audience, this is usually more important than a striking demo image: the product that wins is the one that reduces unnecessary steps.

The term agentic AI in this context typically means not just answering questions, but the ability to guide a person through a chain of actions: understand the request, gather the necessary context, open relevant functions, and help reach the result. In Bloomberg's brief announcement, not all ASKB scenarios are revealed, but the emphasis on integration directly into Terminal shows the direction: the assistant becomes part of the workflow, not a separate showcase of capabilities.

Why for Terminal

For Bloomberg, this is a logical development of Terminal, which has been the center of decision-making for financial market participants for decades. In such an environment, the value of AI is measured not by beautiful formulations, but by time savings, reduced friction, and speed of access to needed information. If the assistant is deeply embedded in the interface, it can remove some manual screen navigation, shorten the path from question to answer, and make complex functions more accessible to users who don't remember dozens of commands and abbreviations.

The choice of venue is also telling. Family offices, wealth managers, and investment teams often work at the intersection of macro analytics, news, corporate events, and risk management. For them, tools that don't just summarize text but help quickly assemble a working picture of an asset, company, or topic are especially valuable.

In this sense, embedded agentic AI fits real tasks better than a regular chat living separate from the main system.

Practical value

Even without a complete list of functions, it's clear why Bloomberg is betting on exactly this format. When AI is embedded in infrastructure, it's useful not just for answering questions but for organizing the analysis process itself. In a financial environment this is especially noticeable because the cost of error, delay, or loss of context here is higher than in regular office software. For the user, such an approach typically means several things:

• less manual searching through Terminal screens, menus, and sections
• quick transition from a question to a set of needed data and news
• help with multi-step queries where a single text hint isn't enough
• smoother onboarding for new users who find professional interfaces difficult to master
• keeping work within one environment without constant switches between services

The key question now is not whether such AI will appear in professional terminals, but how deeply it can be embedded in critical financial scenarios. The closer the assistant gets to actions rather than just advice, the more noticeable the effect will be. But along with this, demands for transparency will grow: users need to understand what data was used, how the answer was built, and where automation ends and human decision-making begins.

What it means

The ASKB demo shows a simple shift: the next stage of AI in finance is not separate chatbots, but embedded agents within familiar work systems. If Bloomberg brings this model to reliable everyday work, the market will get not another AI feature for the checklist, but a new interface to financial data and analytics.