This interview covers how AlphaSense CEO Jack Kokko uses AI to boost investment research efficiency by 5-10x. He sees AI as an accelerator, not a disruptor, making search smarter. He's bullish on expert interview data (interviewing former execs, customers) for uncovering truths not in public filings. Key holdings: Tegus (acquired for nearly $1B, filling a gap in expert interviews), OpenAI's GPT series (tech supplier, models update every 1-2 months), and Google (exited enterprise search).
AlphaSense CEO Jack Kokko shared on the program how the company started by aggregating fragmented financial data and, with the evolution of large language models (LLMs), has fundamentally transformed the research experience. The core thesis is that AI-driven market intelligence search can improve re
Jack Kokko is the CEO and founder of AlphaSense. The company started by aggregating fragmented financial data, and with the evolution of large language models (LLMs), it has fundamentally transformed the research experience. The central theme of this episode is how AI is pushing investment research from keyword search toward high-conviction decision-making tools. The most impactful takeaway from the entire episode: Jack Kokko believes AI-driven market intelligence search can improve research efficiency by 5–10 times, and after AlphaSense acquired Tegus, its business expanded from serving investment institutions to all types of enterprises, increasing its addressable market by 10 times.
Jack Kokko believes the emergence of large language models (LLMs) has been a "massive accelerator" for AlphaSense, rather than a disruption to its existing advantages.
Kokko reviewed the company's technology evolution path: from early small language models (capable of understanding short phrases and local context), to models better at grasping long paragraphs and themes, and then to the "truly big breakthrough" in the LLM era. LLMs brought two major capability leaps: first, the ability to truly understand users' natural language query intent (rather than just keyword matching); second, the ability to comprehend text meaning across broad contexts. He described LLMs as "brains in a box," and the company spent its first few months thinking about the most direct applications — document summarization, cross-company insight aggregation, and so on.
Key mechanism: Kokko emphasized that AlphaSense's core advantage is not the model itself, but "a truly intelligent search engine that feeds precise information to the generative AI layer above." He described the RAG (Retrieval-Augmented Generation) architecture: the search layer first semantically finds all relevant information, then feeds it to the LLM for summarization. This architecture preserves AlphaSense's original core value — users can see the original sentences and snippets from each insight source, quickly verify them, and access the full context.
Competitive landscape: Kokko acknowledged that LLMs have indeed given the market new tools — large companies' IT teams are asked by their CEOs to "do something with generative AI," and the most natural idea is to solve internal content fragmentation. Even Google previously exited the enterprise search market. However, he believes this actually makes AlphaSense's advantages more prominent: "When technology evolves at breakneck speed, just as you feel you've built something, a new model appears — this gives fast-deploying startups an even greater edge."
Data support: The platform now hosts over 300 million high-quality documents, including corporate filings, earnings reports, press releases, and more.
Kokko believes the decisive factor in enterprise financial search competition is neither pure tools nor pure data, but the integration of "tools + data" — allowing users to complete all searches within a single system.
Kokko drew from his own experience: as an investment banking analyst, he had to log into dozens of different systems and even ask the library to gather information from more sources. Fragmentation has always been the core pain point. "Our strategy from day one has been: one place, one search, done. If you can't find it on our platform, it doesn't exist." He cited client feedback: clients often say that entering one system gives them the peace of mind that they "don't need to go to 20 different systems anymore."
The value of proprietary data: Kokko specifically highlighted expert interview content as the most exciting dataset. There are two reasons:
1. Providing entirely new perspectives: Public market information (SEC filings, sell-side research) is already relatively abundant, but companies always show their best side. Expert interviews — speaking with former executives, customers, partners, and competitors — yield honest answers from parties "without a vested interest."
2. Filling the private market gap: The private market is becoming increasingly important, yet there is almost no systematically accessible information. Expert interviews fill this enormous void.
Strategic logic behind the Tegus acquisition: AlphaSense was previously a "distant second" in the expert interview space, while Tegus innovated the business model and achieved scale. After the acquisition, the combination produced a "1+1=4 or 5" effect. Kokko revealed that although the deal size was close to $1 billion (significant for a startup valued at tens of billions), investors supported it "with almost no additional due diligence" because the strategic logic was extremely clear.
Kokko believes the hardest part of M&A is not the deal negotiation, but how to quickly build trust between two organizations after integration.
Kokko admitted that although he started his career as an investment banking analyst, he still couldn't find a "playbook" when actually doing M&A. He summarized several lessons:
Investor relations: Patiently build an investor base with "deep pockets," prepare capital when it's not needed, and be ready to act quickly when opportunities arise. In the Tegus acquisition, investors supported the deal "with almost no additional due diligence" because the strategic logic was extremely clear.
Cultural integration: Kokko emphasized that the least guidance exists for the "soft factors" — how to make acquired employees feel at ease from day one. His approach:
Key data: AlphaSense had completed several acquisitions prior to this, but Tegus was "a much bigger bite," with a deal size close to $1 billion.
Kokko believes AlphaSense's most defining moment was "breaking into the enterprise market" — proving that its product is not just a tool for financial professionals, but a search platform for all business professionals.
Initial strategy: Starting with small hedge funds. Kokko found that small funds are "excellent clients" — they face intense pressure to find an edge and are willing to pay for tools that prove their value. However, one surprise was that these clients often treated AlphaSense as a "secret weapon" and were reluctant to spread word-of-mouth.
Enterprise market breakthrough: Kokko described a "beachhead" strategy — starting with the Investor Relations (IR) department. IR departments need to understand how investors view the company, and AlphaSense helps them "replicate buy-side analysis" to tell a better story. From IR, the product expanded to corporate strategy, financial planning, competitive intelligence, corporate development, strategic marketing, product development, and even sales engineering. Each role has different titles across companies, requiring an "internal guide" to navigate.
Key insight: The product itself didn't need to change — the same search tool is equally valuable for business professionals. This proved that the addressable market is 10 times larger, covering "every type of business."
Kokko envisions that over the next five years, AI will act as "an increasingly smart team of super analysts," compressing research that once took weeks into minutes.
Current capabilities: Kokko gave an example: today, it is already possible to "build an introductory report on the XYZ industry" in 2–3 minutes. The system can ask the same five questions across 500 documents and then output "this is the most important set of documents, along with some preliminary conclusions."
Future directions:
1. Reasoning capabilities: Currently, AlphaSense's system requires building a multi-layer model sequence of "plan-execute-check-clean." If LLMs themselves possess stronger native reasoning abilities, this process could be simplified.
2. Larger context windows: While maintaining high speed and low cost.
3. Multimodal analysis: Processing video, audio, and more.
4. Simulation capabilities: Enabling machines to adopt different perspectives on an industry and simulate different paths — not predicting "where it will go," but showing multiple paths of "where it could go."
Impact on investors: Host Patrick summarized: "If you're an investor or analyst, you'd better not make 'finding information and synthesizing it' your core competency — these tools are making that better and better." Kokko agreed, noting that this will ultimately make markets more efficient and capital allocation better.
Technology planning approach: Kokko revealed that the company uses a "model-agnostic" architecture — it can plug into any LLM and quickly switch between different models for different tasks. Organizationally, the team stays agile, continuously tracking cutting-edge research papers, with "only 1 out of every 10 truly impactful papers making it to my desk."
| Position | Guest Sentiment | Key Data |
|---|---|---|
| Tegus | Acquired, highly bullish | Transaction size close to $1 billion; AlphaSense was previously a "distant second," Tegus innovated the business model and achieved scale |
| OpenAI (GPT Series) | Technology provider, neutral | Model capabilities change "almost every 1-2 months"; GPT-3/ChatGPT was a "truly major breakthrough" |
| Competitive reference | Previously launched an enterprise search product but later exited the market | |
| Perplexity | Competitive reference | Users can search financial documents such as SEC filings |
1. Jack Kokko believes LLMs are a "massive accelerator" for AlphaSense, not a disruptor — because the company has spent over a decade building a "truly intelligent search engine," and LLMs merely make the top-layer interaction more natural, rather than replacing the underlying search capability.
2. "Fragmentation" is the core pain point — Kokko draws from his own investment banking experience: needing to log into dozens of systems and keyword-search individual PDFs. AlphaSense's strategy is "one place, one search, done," making clients feel that "if it's not here, it doesn't exist."
3. Expert interviews are the most exciting dataset — because they provide a "no-stakeholder-bias" real perspective, especially filling the information vacuum in private markets, which are becoming increasingly important.
4. The hardest part of M&A is not deal negotiation, but cultural integration — Kokko emphasizes anticipating the fears of acquired employees, providing answers on day one, and painting a narrative of "growing together," making employees feel that "riding this wave can be cool."
5. The breakthrough into the enterprise market is AlphaSense's most defining moment — starting from the investor relations department, proving the product can serve all business professionals without modification, expanding the addressable market by 10x.
6. Small hedge funds are "excellent early customers" — they face intense pressure to find an edge and are willing to pay for tools that prove their value, but often treat these tools as "secret weapons" and are reluctant to spread word-of-mouth.
7. Kokko believes that in the future, AI will act as "an increasingly intelligent super-analyst team" — compressing research work that once took weeks into minutes, allowing the system to ask the same five questions across 500 documents, outputting the most important document set and preliminary conclusions.
8. The gap between European and U.S. tech ecosystems lies in a "chicken-and-egg problem" — Kokko points out that in Europe, it is harder to raise capital, harder to find experienced founders and employees, the culture is more risk-averse (failure in the U.S. is "a step toward success," in Europe it may be "the endgame"), and successful companies are often acquired by U.S. firms.