您的浏览器禁用了JavaScript(一种计算机语言,用以实现您与网页的交互),请解除该禁用,或者联系我们。 [伯恩斯坦]:商业服务领域的人工智能:从就业担忧到生产力提升 - 发现报告

商业服务领域的人工智能:从就业担忧到生产力提升

信息技术 2026-05-15 伯恩斯坦 玉苑金山
报告封面

AI for Business Services: From job fears to productivity on a plate This is part of a series of reports on AI disruption. Business services and catering stocks havebeen hard hit by fears of significant job destruction. While we acknowledge the importance ofthe shift induced by AI, we are convinced that its impact will not necessarily be negative for thetwo sectors. A long-term decline in employment seems unlikely, in our view, as we see clearsigns of these companies leveraging AI to increase productivity and efficiency. In this report,we address two key questions: i) What will these two industries look like when AI has been fullyintegrated? ii) How might estimates change to reflect the impact of AI? Sabrina Blanc+33 1 42 13 47 32sabrina.blanc@bernsteinsg.com Will Kirkness+44 20 7676 8355will.kirkness@bernsteinsg.com Filippo Giardini+44 20 7762 4723filippo.giardini@bernsteinsg.com Market sentiment has cooled recently on AI-exposed names,amid either white-collarjob loss fears at catering or structurally lower barriers to entry for some business servicesnames (EXPN, TEP, staffers). Some argue that job destruction could outpace job creation,particularly for white-collar workers, as AI automates an increasing share of cognitive tasks.This reinforces the view of AI as primarily a major economic disruptor in its own right, ratherthan merely an efficiency tailwind. History suggests reallocation, not collapse.Past technological shocks (electrification,IT, the internet) resulted in reallocation of jobs and lifted employment over the long term,as noted by Adecco. Generative AI could accelerate the automation of certain white-collartasks, and thus render some roles obsolete, but it is also likely to drive demand for newcapabilities, such as development of AI systems to manage migration complexity, retrainusers, re-implement integrations, and mitigate switching risks. Disrupted or supercharged?Mapping AI risks in Services and Software. McKinsey’s 2030 scenario projects that c.30% ofwork hours will be automated across Europe and the US, reducing headcount in canteens andrestaurants (non-cooking roles) and in back-offices, which are core exposures in the Business& Industry segment, while boosting demand in healthcare and STEM fields. In European business services we differ from consensus opinions on key stocks:1) forExperian (U), AI is a contributory factor to our lower medium term organic growth estimates,2) likewise at TP (OP) we see lower organic growth vs medium term expectations as higherautomated volumes = lower price, 3) for staffers, we are more optimistic than consensus,we think the incumbents are primed to deliver efficiencies from AI tooling and the nature ofrevenues - blue collar temp focused placements - are well-placed. We estimate that for catering operators, AI-linked operational gains couldultimately offset most of the demand risk. We analyze the impact of AI as a meansfor catering companies to optimize processes, accelerate task automation and gainefficiencies across business models. For catering companies, we assume a maximumnegative impact on revenue of around 1-2%, partially mitigated by additional costefficiencies. This would imply that the sector will be disrupted but only to a limited extent. Our top picks within the two sectors remains Compass (Outperform, PT $38.7) andRentokil (Outperform, PT 576p). BERNSTEIN TICKER TABLE INVESTMENT IMPLICATIONS Our top picks within the two sectors remains Compass (Outperform, PT $38.7) and Rentokil (Outperform, PT 576p). DETAILS BUSINESS SERVICES SECTORS FACING THE AI ROLL-OUT IS THE STOCK MARKET OVERREACTING TO THE RISKS OF AI? Over the past year and a half, market concerns have been building about the potential impact of AI on companies in the businessservices sector — whether firms involved in task automation or database management (Experian, Teleperformance) or in thetemporary work segment (Adecco, Randstad), or more broadly suppliers to other businesses that could be affected — namelycaterers (Compass, Sodexo, Elior). There have been fewer concerns raised with regard to TIC companies. Overall, the stockperformance of business services stocks has been negative over the past 16 months: Teleperformance has lost 25%, whileAdecco and Experian are down 20%. Even though there may be other factors that could explain this performance, particularlyfor Sodexo, Sunbelt and Bunzl, sentiment on the sector is clearly negative. The business services sector offers disparate exposure to the rise of AI. A collection of stocks that are expected to havenegative implications, albeit debated to varying degrees and stocks that offer a degree of immunity due to the physicality ofthe work.Our central thesis is that AI is a force for productivity which changes the nature of work,changes the shapeof markets and increases the speed that new products and services can be brought to market. It doesn't necessarilychange who owns the revenue streams but will have implicatio