企业在AI竞赛中的进与退

Data Science& AI ReportHow Companies Are Moving Ahead— Or Not—in the AI Race The Big Picture The AI Race is On, but AI DevelopersSay Companies Are Still Learningto Tie Their Shoes Only 22%of respondents say they would in a messy middle between experimentation,iteration, ROI, and scalable value. There’senthusiasm around leveraging AI andincorporating it more into workflows.However, that eagerness often isn’t matchedwith a cohesive strategy, which can limitinnovation and adoption while slowing downand impeding the value and impact of AI forbusinesses and their customers.To better understand how companies aredeploying and implementing AI and where productivity of AI initiatives. Over 57% report thatit takes over one month to move from developmentto production of AI projects. This delay indicatesa lack of cohesion among teams. A unified andclear strategy results in faster time to value andinnovation, reducing model-to-production cycletimes from months to days.That translates into long-term adoption, as well:75% of respondents say employees and customers describe their organization’s AIdeployment as strategic. don’t begin using AI tools until 1-6 months afterthey’re deployed. Additionally, over 30% say trustis either still developing or is non-existentafter 3 months or longer.As AI-assisted workflows and production usecases increase, companies are at a critical common roadblocks may lie, we surveyedsoftware and AI engineers, data scientists,and others in similar roles. Their responseshighlight some encouraging trends movingforward, as well as areas ripe for improvement.One of those areas is the lack of clarity withAI usage: only 22% of respondents say Over 57%report that it takes over one monthto move from development to junction. The difference between those whofind success and those who spin their wheelsis in strategic approaches to AI usage, and inconfidence and organizational support. This reportwill explore where the greatest value in AI toolslies and what’s holding users back. It will alsodive into overlooked opportunities and ways forcompanies to move forward faster. Consider ityour personal training guide for the AI marathon. production of AI projects. they would describe their organization’s AIdeployment as strategic. That’s fewer than onein four companies that have a targeted planaround their AI usage. The rest of the classmight need to take a step back and re-evaluatewhat they’re really hoping to achieve. How We’re Getting ValueOut of Our AI Work How does your organiation demonstrate ROI from its initiatives?Select all that apply. Not Everyone Sees Eye to Eyeon ROI—And That’s Okay error reduction or risk mitigation (25%),but that’s another important considerationfor companies. According to McKinsey,executives spend about 40% of theirtime making decisions. That’s time thatcould be better spent on driving morevaluable business outcomes. Testingand experimentation are okay. However,routinely making the wrong decisions isn’tmerely a waste of time, energy, and money;it can cause fractures across the entireorganization. More action around errorreduction and risk mitigation can help steadyorganizations from too much tinkering orveering down a troublesome path. focusing on efficiency gains. Wideningthe ROI lens to reveal AI as a strategicgrowth and resilience driver, not just a costreduction, opens up new opportunitiesand insights across teams.Productivity improvements (58%) and cost savings (47%) are the most commonways people showcase ROI from AIinitiatives. These are welcome benefitsthat organizations can point to from aneffectiveness standpoint. Companiesshould certainly keep track of them forlarger projects or more significant rolloutsto showcase their results and reinforcetheir value to the business, especially whenthey’re demonstrating a quicker timeto value. HOW WE’RE GETTING VALUE OUT OF OUR AI WORK Could the Best ROI Come FromNot Measuring It At All? There’s a line betweencollecting metrics andmeasuring ROI with thosemetrics. Understand whereAI makes you better. Youshould always be measuringand looking at the outcomes,just perhaps not evaluatingAI initiatives strictly on whatmetrics move. Now is thetime to hone these tools.” isn’t measuring ROI at all or is unsurehow their company is tracking it. About26% of respondents also noted difficultydemonstrating ROI as a top concern aboutAI risk in their organization.At first glance, that seems overwhelming. Nearly 1 in 7companies isn’t measuringROI at all/are unsure howtheir company tracks it. Company leaders may wonder if theseadvancements can be properly measuredif their teams aren’t working toward certainbenchmarks. But what if not strictlymeasuring ROI is actually a strategic moveto advance AI innovation? Companies aregenerally leaving space for their teamsto experiment internally with these typesof tools. At the same time, there’s oftenpressure to showcase outcomes from aninvestment, especially one as grandiose asAI. But AI an