AlphaSense’s next-gen SuperAnalyst: the ‘always-on’ agent that monitors your thesis and rebuilds the deck itself
AlphaSense launched a next-generation SuperAnalyst on Tuesday — an always-on AI agent for finance teams that plans research, runs the analysis, monitors the assumptions behind an investment thesis, and updates the work product on its own. The release adds a Context Graph, reusable Skills, and a preview of no-code SuperAnalyst Apps.
AlphaSense announced a next-generation version of SuperAnalyst on Tuesday, its always-on AI agent for financial and strategic work — pitching it as an evolution from answering research questions to continuously executing workflows, 24 hours a day, on live market data.
SuperAnalyst is designed to be delegated an objective, not managed through a sequence of prompts. The agent plans the work, finds and reconciles the relevant evidence, performs analysis and calculations, writes and runs code, and produces decision-ready outputs — investment memos, financial models, comparison tables, monitoring dashboards, and board presentations that refresh on their own.
“The primary determinants of AI workflow success are the quality of the data the agents can leverage and the capability of the tools they can use,” said Jack Kokko, Founder and CEO of AlphaSense. “With SuperAnalyst, they can execute the workflows behind those decisions deeply and broadly at superhuman speed, grounded in trusted intelligence.”
It keeps working after the first answer
Unlike a static deliverable, SuperAnalyst monitors the assumptions underpinning an investment thesis, portfolio, transaction, or strategic decision; identifies material changes in the market; and updates the work product with new information — without requiring a new prompt.
Two early users describe the practical effect. Tiago Rodrigues, Investment Director at Aberdeen Investments Limited, said the agent built “a consistent, organized model from AlphaSense’s Canalyst data and company filings, much like the reasoning a human analyst would follow, enriched with comprehensive qualitative commentary around it.” Ian Lieberman, Partner and Director of Research at Talos Eurisko Asset Management LP, said it lets him “cover more ground faster and with more confidence” — “it’s not replacing analysts’ work.”
The Context Graph and reusable Skills
Two architectural pieces underpin the release. Context Graph converts unstructured content into structured, verified knowledge, with every data point tied to its source evidence — and bundles in the tools, skills, permissions, and memory each workflow requires, reasoning inside the customer’s own environment under enterprise-grade permissions.
SuperAnalyst Skills are reusable, role-specific workflows. AlphaSense ships a pre-built library developed by former practitioners — Buyer Universe Development, Earnings Preview, Forecasting, Competitive Mapping — and organizations can create custom Skills in natural language, encoding a firm’s methodology, templates, and house style into repeatable workflows that operate on live data. As market conditions change, outputs — from dashboards to Excel models — update automatically.
SuperAnalyst Apps: describe it, get software
At its annual AlphaSummit conference, the company also previewed SuperAnalyst Apps — living, shareable applications created with natural language and powered by AlphaSense’s intelligence layer. The pitch: describe the app you need, the agent builds and runs it, and AlphaSense keeps feeding it live intelligence, execution, and governance — so users can, in the company’s words, build the software they need without hiring a development team, operating an AI stack, or compromising data governance.
Why it matters
The platform underneath combines over 500 million premium business documents — equity research, earnings calls, expert interviews, filings, news — and a library of 300,000 proprietary expert interviews. AlphaSense counts Bank of America, J.P. Morgan, Morgan Stanley, Google, Intel, Microsoft, Nvidia, and Pfizer among its customers, per the release.
The launch lands as the agent wars shift from answering to executing — and as finance-grade governance becomes the moat. AlphaSense’s wager is that the winning agent isn’t the smartest model but the one wired into proprietary data, persistent memory, and audit-ready attribution. Whether that delivers “superhuman speed” will be judged by the analysts who still have to defend every data point back to its source.