AlphaSense vs Bloomberg ASKB

AlphaSense and Bloomberg ASKB put AI on top of two very different foundations. AlphaSense’s AI searches a research library of more than 500 million documents, strongest on broker research and expert-call transcripts. ASKB is the Bloomberg Terminal’s AI, answering across Bloomberg’s market data, news and research and handing back the code behind its analysis. For firms that need AI built around their own data and definitions, with every number computed in code and cited to the line, the third option on this page is Kepler.

We build Kepler, so weigh the last section with that in mind. Facts about AlphaSense and Bloomberg come from their sites, press releases and public reporting as of November 2026, linked inline.

AlphaSense vs Bloomberg ASKB at a glance

	AlphaSense	Bloomberg ASKB	Kepler
Core idea	AI search and agents over a licensed research library	Conversational AI inside the Terminal	Your firm’s ontology, every document read, numbers computed in code
Content included	500M+ documents, 1,500 broker partners, 300,000+ Tegus expert transcripts	Bloomberg data and news, research from 800+ providers, Bloomberg Intelligence	26M+ SEC filings, 40M+ IR documents and transcripts, market data, news, plus your firm’s documents
Your own content	Internal decks, reports and notes in the same search	Instant Bloomberg chats, Worksheets, Bloomberg Drive files	Read in full at ingest and resolved into your ontology
Agents and workflows	SuperAnalyst plans, runs code, builds memos, models and dashboards	ASKB agents in parallel; saved ASKB Workflows	Agents on schedules and events, acting as their owner
Where numbers come from	AI answers over library content and standardized financials; SuperAnalyst writes and runs code	Bloomberg data, with BQL code for data analysis	Pulled from the source and computed by code
What a citation opens	The cited passage in the source document	The source document, chat or file passage	The exact page, line or cell, with formula and inputs
Your firm’s definitions	Prompts, templates and saved workflows	Prompts and saved workflows	Built into your ontology, applied every time
Where it runs	AlphaSense app, Excel and PowerPoint add-ins	The Bloomberg Terminal, with Excel and BQuant via BQL	App, Excel, PowerPoint, Word, Claude, ChatGPT, Slack, API
Reported pricing	Average $10,131 (smaller firms) to $139,455 (enterprise) a year (SpendHound)	$28,320 to $31,980 a year per Terminal (NeuGroup)	Through our team

AlphaSense: the research library’s AI

AlphaSense’s case rests on what comes with the subscription. Tegus.com, now an AlphaSense page, lists 300,000+ Tegus expert transcripts, Tegus expert call services, 4,000+ Canalyst models, a 500M+ document library and 1,500 broker partners. On top of that sit AI search that answers with citations to the source passage, a grid for comparing hundreds of documents side by side, and Deep Research for primers and competitive landscapes. SuperAnalyst, launched on October 6, 2026, takes one objective, plans the work, reconciles the evidence, writes and runs code, and produces investment memos, financial models, comparison tables and monitoring dashboards. AlphaSense also offers Excel and PowerPoint add-ins and connectors that bring internal decks, reports and notes into the same search.

What users say: “The AI-generated insights are helpful, but I still prefer to verify anything important against the original sources,” wrote a G2 reviewer in September 2026 (AWS Marketplace, G2 reviews).

Strongest at: expert calls and broker research in one search with filings, transcripts and news.

Bloomberg ASKB: the Terminal’s AI

Bloomberg introduced ASKB in beta on February 23, 2026 as a network of AI agents, built on several commercial and open-weight models, that work in parallel across Bloomberg’s data, news, research and analytics. Its sources include hundreds of millions of company documents, Bloomberg News, sell-side and independent research from more than 800 providers, and Bloomberg Intelligence, BloombergNEF and Bloomberg Economics. When an answer includes data analysis, ASKB returns the underlying BQL code so users can carry it into Excel or BQuant, and ASKB Workflows saves multi-step tasks such as pre-earnings preparation as reusable templates.

The October 15 general availability release brought the client’s own context in. Once enabled, ASKB can search up to three months of a user’s Instant Bloomberg chats, answer questions about the securities in a Worksheet, and treat files in Bloomberg Drive as a persistent source, citing the chat, file and passage behind each point. Bloomberg says client material stays permissioned and separate from Bloomberg sources.

What users say: “I don’t even try to use the first answer anymore but ask it again to check its previous answer against verified and authoritative sources/documentation and will come up with many corrections,” wrote a Terminal user on Hacker News in October 2026, describing ASKB’s answers about Bloomberg functions and fields.

Strongest at: real-time markets, news and analytics, with AI where the trader or PM already works.

Head to head

Content. AlphaSense wins on expert calls and on the size of its document library. Bloomberg wins on market data, news and its own research arms. Both carry broker research.

Workflow. ASKB lives in the Terminal, next to the screens, chats and BQL a Bloomberg user already relies on. AlphaSense is a research destination of its own, with Office add-ins for the work that follows.

Your own content. AlphaSense pulls internal documents into the same search as its library. ASKB now reads a client’s chats, Worksheets and Drive files, which is a strong fit for desks whose working context lives in Instant Bloomberg.

Agents. SuperAnalyst carries one objective through to memos, models and dashboards. ASKB coordinates agents for each question and saves repeatable work as Workflows.

Numbers. Both attribute answers to sources: SuperAnalyst writes and runs code, and ASKB hands back BQL for its data analysis. In both, house definitions live in prompts, templates and saved workflows, so check any derived figure on your own metric before it goes into a model or a memo.

Pricing. Neither publishes prices. For AlphaSense, SpendHound reports average spend of $10,131 a year for smaller customers and $139,455 for enterprise customers, and Vendr notes that contracts commonly auto-renew unless canceled 60 to 90 days ahead. For Bloomberg, NeuGroup reported $31,980 a year for a single Terminal from 2025 and $28,320 per Terminal for multi-Terminal clients.

Who each one is for

AlphaSense fits research teams at hedge funds, asset managers and corporates who spend the day reading: expert calls, broker notes, transcripts and filings across many companies.

Bloomberg ASKB fits traders, PMs and analysts who already run their day in the Terminal and want AI over Bloomberg data, news and their own chats without leaving it.

The third option: Kepler

We build Kepler.

Both products start from content someone else licenses. The questions that decide an investment start somewhere else: your holdings, your coverage list, your definition of adjusted EBITDA, the model your PM signed off on last quarter. We built Kepler to start there, and to make every number in the answer provable.

Our engineers built this layer before at Palantir. On site with your team, they build your firm’s ontology: each company, fund, holding, peer set and internal model resolves to one record, every house term carries one definition, and every relationship points to the sentence that established it. Public data arrives already loaded: 26M+ SEC filings across 20+ form types, 40M+ IR documents and earnings call transcripts, market data and news for 14,000+ companies in 120+ countries. Your SharePoint, Box, email, CRM and research notes connect read-only in the first week, and extraction models we train read every page of every document before anyone asks, so a question returns the complete set and a follow-up is a lookup.

The model never writes a number. Code pulls each figure from its source and computes the rest with the formula visible; statements assemble from each filer’s XBRL as filed or as restated; every citation is checked against its passage before you see it, and conflicting sources appear side by side. We route each job to the model best suited to it, with small models we train doing the reading and frontier models handling the question and the write-up.

Agents run overnight on schedules and new filings, acting exactly as the person who set them up, and your playbooks run the same way every time. The work reaches the Kepler app, Excel, PowerPoint, Word, Claude, ChatGPT and Slack, and your engineers can build on it through our API and MCP server. One permission check governs every read, every query is logged with its lineage, and we hold SOC 2 Type II.

Questions teams bring to Kepler

• “Every holding in our long book whose last 10-Q added a going-concern, covenant or liquidity warning, with the exact language.”
• “Net debt to EBITDA for our 25 largest positions on our definition, last eight quarters, every input cited.”
• “The morning after any company we own reports, a one-page summary in our template with what changed against our model.”

“I can see exactly where every number and conclusion comes from, which means I can defend the work.” (Vice President, Global Investment Bank)

Not the right fit if: your desk mainly needs a terminal for live prices, chat and execution, or a library for reading expert calls.

For deeper comparisons, see Kepler vs AlphaSense, AlphaSense alternatives and AlphaSense vs Hebbia.

How to decide

1. Name the job. Reading experts and brokers points to AlphaSense. Working inside live markets points to Bloomberg. Figures that go into the firm’s models and memos, on its own definitions, point to Kepler.
2. Check where your team works. A desk that lives in the Terminal will use ASKB more than a separate destination.
3. Test the numbers. Ask each tool for a derived figure, click through to the source, recompute it and ask again. Our guide to checking AI numbers lists the full set of tests, and the best AI tools for hedge funds cover the wider field.

Bring a coverage list to Kepler at app.kepler.ai, or book a demo on your own documents.

Frequently asked questions

What is the difference between AlphaSense and Bloomberg ASKB?

AlphaSense is a research platform whose AI searches a library of more than 500 million documents, including broker research from 1,500 partners and 300,000+ expert transcripts. ASKB is the Bloomberg Terminal's conversational AI: agents that answer across Bloomberg data, news, Bloomberg Intelligence and research from 800+ providers, and hand back BQL code for any data analysis. One starts from a library, the other from a market data terminal.

Is Bloomberg ASKB generally available?

Yes. Bloomberg introduced ASKB in beta on February 23, 2026 and announced general availability on October 15, 2026. The release added the option to bring a client's own Instant Bloomberg chats (up to three months of history), Worksheets and Bloomberg Drive files into ASKB, with every answer citing the chat, file or passage it drew on.

How much do AlphaSense and Bloomberg cost?

Neither publishes prices. SpendHound reports average AlphaSense spend of $10,131 a year for smaller customers and $139,455 for enterprise customers. NeuGroup reported a single Bloomberg Terminal at $31,980 a year from 2025, or $28,320 per Terminal for multi-Terminal clients. Bloomberg hasn't published a separate price for ASKB.

Which is better for expert calls and broker research?

AlphaSense, for expert calls: the Tegus library of 300,000+ transcripts and Tegus expert call services come with it. For broker research both are strong. AlphaSense says it works with 1,500 broker partners, and Bloomberg says ASKB reads sell-side and independent research from more than 800 providers alongside Bloomberg Intelligence.

Is there an alternative to both AlphaSense and Bloomberg ASKB?

Yes. Kepler, which we build, starts from the firm instead of a content library. Our engineers build your firm's ontology, extraction models read every filing, transcript and internal document in full, and code computes every number and cites it to the exact page, line or cell. Agents run on schedules and new filings, and Kepler works in Excel, PowerPoint, Claude, ChatGPT and Slack.