2025年国家商业与金融

WHAT’SINSIDE Executive Summary1 CHAPTER1The AI Productivity Paradox4 CHAPTER2The Specialized Intelligence Imperative7 CHAPTER3The Expert Renaissance Powered by AI9 CHAPTER4Solving the Internal Knowledge Challenge11 CHAPTER5The Rise of Automated andProactive Intelligence13 CHAPTER6Trust, Security, and RegulatoryReadiness as Differentiators15 CONCLUSIONFrom Adoption to Integration: The Road Ahead17 Methodology & Firmographics20 Executive Summary In 2025, a year marked by hype and hesitation, the teams seeingthe biggest gains aren’t chasing shiny tools, they are quietlytransforming how work gets done. Organizations adopting workflow-integrated, expert-driven AI are seeing measurable gains in efficiencyand insight generation. By embedding AI into real workflows, they’re unlocking new levelsof productivity, sharper insights, and faster decisions. Thesefindings are based on a survey conducted in May 2025, reflectingthe perspectives of 300 full-time working professionals in theUnited States, the majority of whom are decision-makers ongenAI, across areas like consulting, corporate strategy, competitiveintelligence, asset management, and investment banking. See complete methodology and firmographic information at the end of this report. Key Insights Trust and accuracy are keydifferentiators: Professionals achieve Expert synthesisand genAIcapabilities amplify,not replace, humanjudgment 7+ 74%trust specialized,industry-specific AI hours/week productivitygains from AI use While generative AI has been around for a number of years,organizations are still shaping how they will use genAI tools in theirworkflows. In the past two years, genAI has gone from hypotheticalto reality. Our latest study shows that firms are reclaiming 6-10hours per professional each week — time that converts directlyinto faster decisions, measurable revenue gains, and lower riskexposure. In market research, genAI is mainly used for reportgeneration, trend analysis, and profiling, while in investmentresearch, it’s primarily used for risk management, predictiveanalysis, and market monitoring. These use cases suggest thatgenAI’s value is beginning to extend beyond efficiency, supportingmore informed analysis and timely decision-making. Furthermore, domain-trained insights beat generic answers: 74%of professionals trust industry-specific AI, which is significantlyhigher than their trust in consumer grade tools (62%). And manyare pairing that confidence with expert-level synthesis to amplifyhuman judgment. Looking ahead, most organizations expect continued stabilityand growth in their AI capabilities — expanding both internal andcustomer-facing use cases — with a shift toward hybrid models thatblend in-house development with trusted external solutions. This report presents the data behind this shift, offering a clearpicture of where organizations are finding the greatest returns andwhere challenges still remain. CHAPTER1 The AI ProductivityParadox More AI isn’t solving persistentknowledge challenges. Instead, many organizations are layering tools without integratingthem, leading to fragmented workflows and diminishing returns.To unlock true productivity gains, AI must be embedded intocore workflows, connecting internal knowledge with externalintelligence in a seamless, usable way. PREDICTION Organizations thatdon’t integrateand consolidatetheir AI tools willsee decliningproductivity. The SpecializedIntelligence Imperative Generic AI tools are rapidlylosing ground. Organizations still struggle to verify AI-generated outputs, andtrust is becoming a strategic priority, prompting a decisive shifttoward specialized, domain-trained systems. To stay competitive,companies are investing in high-quality data sources andtailored AI tools built for their specific markets, use cases, andcompliance needs. PREDICTION Specializeddata sets —expert transcripts,regulatory filings,niche marketinsights — willbecome essentialcompetitive assets. CHAPTER3 The Expert RenaissancePowered by AIHuman expertise becomesexponentially more valuable. As access to traditional expert networks remains costly, slow, andinconsistent, organizations are increasingly using genAI to scaleand accelerate how they capture and apply expert insights. Thenext phase will be integration: embedding AI-driven expert insightsdeeply into market research, competitive intelligence, and duediligence workflows. PROBLEM INSIGHT PREDICTION Primary research is essential(87% of organizations rely onit for their market intelligence),but expert access remainscostly, slow, and inconsistent. AI-driven expertsynthesis willbecome the norm. 80%of firms now leverage genAItools to enhance expertresearch, creating scale andspeed advantages. Solving the InternalKnowledge Challenge Your most valuablecompetitive advantagealready exists — internally. Yet for many organizations, critical insights remain locked insilos, disconnected from daily decision-making. To change that,compani