2025 has already been an incredibly exciting year for AI development. From the widespread rollout of multimodal AI systems that seamlessly understand and generate text, images, and audio in real-time, to the first glimpses of embodied AI agents performing complex tasks in the physical world, this year is proving to be a defining period. The theoretical is rapidly becoming practical, moving beyond research labs and into the hands of everyday users.
Yet, despite the proliferation of AI tools, their impact at work has, in many ways, been more talk than action. According to Gartner®, “only 12% of individual contributors have integrated generative AI (GenAI) into a critical work task, defined as an essential activity or deliverable at the core of their role.”₁ We believe this means organizations are pouring billions of dollars into building and buying AI tools without seeing the expected ROI.
This has led many to draw the conclusion that this technology is simply overhyped. I prefer to look at the way these tools are built as a core contributor to this fracture between expectation and reality — and why we’re seeing a lack of consistent, meaningful adoption across organizations.
Our Blueprint for High-Impact AI
According to Gartner®, “Daily use of AI unlocks outsized productivity. Workers who use everyday AI daily are 3.2 times more likely to report major productivity gains; however, fewer than half of workers do so.”₁ Regular usage comes down to one thing: how successful these tools are at solving real, tangible, time-consuming problems.
For over a decade, AlphaSense has empowered teams by centralizing the world's most critical business information in one place. This addressed one of the biggest challenges our clients faced every single day: getting the right information they needed to make multi-million dollar decisions.
The next step in this evolution is building tools to mirror our clients’ workflows — creating an on-demand analyst that can do initial diligence, monitor emerging markets, analyze earnings trends, and perform sophisticated scenario analyses. Instead of spending hours finding and synthesizing disparate insights, AlphaSense empowers you to focus on strategy; beginning your research with a cohesive narrative, rather than a list of documents.
Our strategy is built on a simple premise: To deliver transformative value, AI must be built on a foundation of specialized content, intelligent architecture, and deep workflow integration. These three pillars are the backbone of our AI strategy:
1. Differentiated Content: AI is only as good as the information it learns from. While many generative and agentic AI tools leverage data from what’s publicly available on the open web, our platform aggregates premium and proprietary business content, creating a robust knowledge base you can’t find anywhere else. This includes exclusive expert insights from Tegus and in-depth company models from Canalyst, alongside millions of top-tier equity research reports, company filings, and news sources. This allows our AI to reason over data that is vastly more reliable, specific, and valuable for answering sophisticated market intelligence questions.
2. A Purpose-Built AI Platform: We’ve built market-leading AI tools like Generative Search, Generative Grid, and Deep Research on a proprietary platform engineered to excel at financial and business analysis. Rather than relying on a single large language model (LLM), our platform acts as an intelligent orchestrator. It is trained to understand the nuances of our domain and dynamically leverages and combines the world’s leading reasoning models — like Anthropic’s Sonnet 4 family, Google’s Gemini 2.5 and OpenAI’s o3 — with our high-value content. This ensures the best possible model is used for each specific task, from summarization to deep-dive analysis, delivering superior accuracy and insight. And, because these models are continuously upgraded, we have the benefit of being on the cutting edge of each new innovation in language model development.
3. Workflow-Driven Development: At AlphaSense, we’ve taken a workflow-driven approach to the integration of new technology, like generative and agentic AI, which has allowed us to create a strong foundation for our entire suite as the problems they solve get more complex. As we look toward the future, that means building intelligent agents designed to fully automate end-to-end market intelligence workflows. Instead of just augmenting a single step, these agents handle entire processes that consume days or weeks of an analyst’s time. Imagine an agent that automatically prepares a comprehensive pre-meeting brief on a target company by synthesizing recent earnings calls, expert interviews, and market news, or an agent that continuously monitors your portfolio for thematic risks discussed in broker research. This is the future we are building — moving from simple answers to fully automated analytical work.
And of course, for high-stakes business decisions, trust is non-negotiable. Every insight generated by our AI is grounded in our high-quality data and designed for complete auditability. We provide in-line citations for every key point, allowing users to instantly click back to the source document and verify the information for themselves. This transparency eliminates the “black box” problem of generic AI and gives our users the confidence to rely on our platform for their most critical decisions.
Our AI Strategy in Practice
Our recent release of Deep Research in AlphaSense is a prime example of our AI approach. When Gemini and OpenAI released their Deep Research models in early 2025, our team rallied to develop our own tool that could leverage this new technology to serve our clients specifically.
In order for a Deep Research tool to really meet our users’ needs, it needed to be trained on how to best break down strategic tasks like market landscaping, target identification, opportunity sizing, and more; it needed to fundamentally understand the types of premium content it had access to, like broker research and expert call transcripts. It had to be trustworthy, reliable, and produce results like an analyst. And that kind of specialization takes time, as well as significant investment.
In the first month after launch, Deep Research automated over 50,000 research tasks. We’re thrilled with the adoption, and it’s easy to see why — this tool is augmenting days or even weeks of clients' work, while offering unmatched precision and reliability. Users are leveraging Deep Research to create industry deep dives, prep for meetings, and craft pitch recommendations, all with the confidence that they can trust AI to do the heavy lifting.
“It’s rare to get both precision and speed. Deep Research gave us insights we would’ve needed multiple interviews and days of digging to uncover — and did it with citations we could immediately trust and explore.”
– Sudheer Yerabati, CEO at MarketStrat
"The most valuable thing AlphaSense Deep Research gives me back is time. Because it's trained exclusively on reliable sources, I've nearly eliminated the need to double-check for accuracy. For the first time, I can trust the AI's output, which has fundamentally accelerated my entire research process.”
– Marc Nussbaumer, Investment Manager at ZIMA Holding AG
“Most generative AI tools are helpful for ideation — but they rarely cite reliable sources, and the content often lacks the context and accuracy required for high-stakes decisions. AlphaSense’s Deep Research changed that. In minutes, I had access to content that genuinely felt like it came from a dedicated team of analysts after a full week of deep-dive research.”
– Alvin Jogasuria, Head of Marketing, Americas & Europe, GenScript PROBIO
More broadly, over the past six months, we’ve seen adoption of generative AI tools in AlphaSense explode, illustrating just how ingrained they are becoming in clients’ daily lives.
Part of that is leveraging emerging technologies; a lot of it is building tools that actually matter and delivering insights that our clients need, from the most critical sources.
Our vision for the future of AlphaSense is centered on enabling each customer to be 10x or even 100x more efficient at what they do. Everything from initial idea generation, to deep diligence and analysis, to intelligent monitoring will be automated end to end, auditable, and ultimately guided by the users. Our AI will perform complex workflows alongside the user, giving back time, going deeper at each step, and expanding the scope and pace to a scale that hasn’t been possible before.
15 Unexpected Trends in Generative AI Value Realization You Can’t Afford to Ignore, Gartner, Tori Paulman, 8 July 2025
GARTNER is a registered trademark and service mark of Gartner, Inc. and/or its affiliates in the U.S. and internationally and is used herein with permission. All rights reserved.
