## Background
Published on YouTube by Ai4 on August 24, 2025, this presentation features Alice Fournier, Chief Information Officer for the Americas at ISS Facility Services. The session examines how strategically implementing AI within corporate back-office functions eliminates administrative friction and unleashes human value, centered around a case study of frontline recruitment automation.
## TL;DR
Alice Fournier contends that while digital innovation traditionally focuses on customer-facing sales and marketing, the most transformative enterprise value lies in automating repetitive back-office functions like HR, legal, and finance. To solve acute hiring bottlenecks across a global workforce of over 320,000 physical facility workers, ISS Americas implemented Paradox AI's multilingual, voice-enabled recruitment platform. The system fully automates candidate intake, screening, and direct-to-calendar interview scheduling while preserving human recruiters for critical evaluation and hiring decisions. This targeted deployment reduced frontline hiring cycle times by 85%, demonstrating that disciplined, business-partnered AI automation generates immense bottom-line impact.
## Atmosphere & style
The presentation carries a pragmatic, energetic, and grounded tone, trading speculative AI hype for real-world operational execution and clear financial metrics. Fournier delivers her insights with conversational candor and warm humor, playfully contrasting IT's legacy reputation as the 'department of no' against the urgent, empathetic need to make digital systems accessible for non-technical frontline staff. Watching the talk feels refreshingly direct and business-focused, leaving the audience with an actionable blueprint for enterprise-scale functional transformation.
## Key takeaways
- [00:04](https://www.youtube.com/watch?v=89vpsHo6RsU&t=4s) ISS operates as a major global frontline enterprise with over 320,000 employees managing built environments, navigating critical return-to-office demands alongside technological shifts.
- [03:52](https://www.youtube.com/watch?v=89vpsHo6RsU&t=232s) While customer-facing functions adopted digital tools early, corporate back-office functions (HR, finance, legal, IT) hold the largest untapped economic opportunity for AI automation.
- [07:45](https://www.youtube.com/watch?v=89vpsHo6RsU&t=465s) Frontline service businesses rely heavily on time-to-hire for revenue and operational momentum, competing for hourly talent against fast-food giants equipped with frictionless hiring tech.
- [07:45](https://www.youtube.com/watch?v=89vpsHo6RsU&t=465s) Traditional digital application portals alienate frontline applicants facing language barriers or low digital literacy; conversational, voice-enabled interfaces make intake accessible and dignified.
- [12:20](https://www.youtube.com/watch?v=89vpsHo6RsU&t=740s) ISS deployed Paradox AI to automate candidate sourcing, multilingual screening, and text-based direct-to-calendar interview scheduling.
- [12:20](https://www.youtube.com/watch?v=89vpsHo6RsU&t=740s) Recruiting automation keeps humans in the loop by removing administrative coordination so recruiters can focus exclusively on conducting interviews and making hiring decisions.
- [14:59](https://www.youtube.com/watch?v=89vpsHo6RsU&t=899s) End-to-end frontline recruitment automation enabled ISS Americas to achieve an 85% reduction in overall time-to-hire.
- [16:14](https://www.youtube.com/watch?v=89vpsHo6RsU&t=974s) Future corporate AI roadmaps are prioritizing high-friction administrative functions, including contract review in Legal and automated processing in Finance.
- [16:14](https://www.youtube.com/watch?v=89vpsHo6RsU&t=974s) AI proofs of concept should avoid open-ended, non-committal testing and must be strictly time-boxed with clear exit criteria and committed business stakeholders.
- [17:34](https://www.youtube.com/watch?v=89vpsHo6RsU&t=1054s) Functional domain owners—not IT alone—must define the boundaries between AI agents and human workers to safeguard nuanced judgment and empathy.
- [17:34](https://www.youtube.com/watch?v=89vpsHo6RsU&t=1054s) Overcoming cultural resistance to AI requires a tight, collaborative partnership between IT and departmental business leaders.
- [19:09](https://www.youtube.com/watch?v=89vpsHo6RsU&t=1149s) Automated recruiting pipelines provide micro-level telemetry across every stage of the candidate lifecycle, enabling precise operational benchmarking and optimization.
## Detailed walkthrough
### [00:04](https://www.youtube.com/watch?v=89vpsHo6RsU&t=4s) Redefining the Built Environment: ISS and the Return-to-Office Challenge
Alice Fournier introduces ISS Facility Services as one of the largest corporate workforces in the world that most people have never heard of, operating with upwards of 320,000 employees across the globe. ISS is dedicated to making workplaces great by managing the physical built environment: cleaning hospitals, manufacturing facilities, and airport terminals—noting with a touch of humor how critical clean airport restrooms are to the traveler experience—as well as maintaining building management systems, HVAC, temperature control, and electrical infrastructure. In the United States, ISS also operates the corporate catering brand Guckenheimer, ensuring employees are well-fed on-site.
Fournier points out that while ISS is not part of conventionally "sexy" industries like consumer packaged goods (CPG), banking, or cool SaaS platforms, the company found itself at the epicenter of one of the century's defining societal transformations: the return-to-office movement. Alongside AI, return-to-office represents a major workplace disruption. Because ISS delivers human-powered physical services, the company played an unexpectedly vital role in helping enterprises redefine what work looks like inside the physical office space.
- ISS employs more than 320,000 frontline and technical staff worldwide.
- Core operational pillars span commercial janitorial services, technical facilities operations (HVAC, lighting, building controls), and corporate dining via Guckenheimer.
- The company navigates return-to-office demands as a frontline, human-delivered services provider rather than a software or consumer brand.
---
[03:52](https://www.youtube.com/watch?v=89vpsHo6RsU&t=232s) From the "Department of No" to AI in Back-Office Functions
Drawing from her background, Fournier shares how she spent the bulk of her career in CPG leading consumer-facing digital transformations across e-commerce and omnichannel retail. During that period, business units were forced to rapidly build data-driven analytical skills and adopt new tech stacks, often clashing directly with an IT organization she jokingly dubbed "the department of no." That frustration inspired her ambition to run an IT division and transform it into "the department of yes." Having served as CIO for three years, she quips that reality has taught her to be the "department of yes, and...", balancing corporate realities with technological innovation.
That CPG journey crystallized a fundamental insight: while customer-facing departments like sales and marketing adapted early to digital disruption, core corporate back-office functions—legal, finance, HR, and IT—often lagged behind. These operational backbones carry the burden of manual tasks, complex business processes, ticket requests, and approval dashboards. Fournier argues that while generative AI serves as a fine productivity booster for marketing and sales, its most transformative economic impact will happen inside corporate back-office functions. Organizations that successfully automate repetitive operational friction will outpace their competitors and lead their industries.
```mermaid
flowchart TD
accTitle: Corporate AI Transformation Gap
subgraph FrontOffice["Customer-Facing Functions (Early Adopters)"]
direction LR
M["Marketing & Sales"] --> E["E-Commerce & Omnichannel"]
E --> P["Productivity & Analytics Gains"]
end
subgraph CorporateFunctions["Enterprise Core Functions (High Untapped Impact)"]
direction LR
CF["Legal, Finance, HR & IT"] --> MP["Manual Processes, Approvals & Dashboards"]
MP --> AI["Targeted AI Automation Engine"]
AI --> CompAdv["Sustainable Market Leadership"]
end
```
- Sales and marketing teams historically embraced digital transformation early due to direct customer pressure.
- Corporate enterprise functions (finance, human resources, legal, IT) hold the vast majority of repetitive administrative steps, request workflows, and manual business overhead.
- Fournier's core thesis: properly applying AI to eliminate corporate operational drag creates far greater competitive advantage than standard front-office productivity gains.
---
[07:45](https://www.youtube.com/watch?v=89vpsHo6RsU&t=465s) Frontline Recruitment Hurdles and Deploying Paradox AI
In frontline service enterprises, operational momentum and revenue generation depend directly on time-to-hire: the faster an organization can recruit, screen, and place an employee on-site ready for day one, the healthier its financial performance. ISS faced a hiring workflow that was simply too long, especially while competing for the same hourly labor pool against quick-service restaurant (QSR) giants like Starbucks, McDonald's, and Chipotle, all of which deploy slick, frictionless recruitment technology. To eliminate this bottleneck, ISS implemented the conversational AI recruiting platform Paradox (Paradox.ai).
Fournier highlights the human element behind this technological choice. Many frontline candidates—such as cleaners, line cooks, plumbers, and electricians—are not technologically savvy, may have limited formal education or literacy, and frequently navigate English as a second language. Traditional corporate application portals can create an acute feeling of shame and intimidation when candidates struggle to complete basic digital forms. To solve this, Paradox provides an accessible, mobile-first, conversational interface that is fully multilingual and voice-enabled. Applicants who are uncomfortable typing can simply speak to the system, which transcribes, translates, and guides them through the screening process. Simultaneously, the platform automates downstream recruiter burdens, addressing the tedious manual labor of job posting, candidate sourcing, and interview scheduling.
```mermaid
flowchart LR
accTitle: Frontline Candidate Experience via Paradox AI
subgraph CandidateFacing["Frontline Candidate Experience"]
A["Applicant Inquiry"] --> B["Multilingual Voice & Chat Interface"]
B --> C["Automated Screening & Pre-Qualification"]
end
subgraph HRBackOffice["Automated Back-Office Operations"]
C --> D["Automated Interview Scheduling"]
D --> E["ATS Tracking & Recruiter Hand-off"]
E --> F["Day-One Deployment"]
end
```
- Candidate competition centers around large QSR brands (Starbucks, McDonald's, Chipotle) that set high expectations for candidate speed and convenience.
- ISS serves a frontline demographic that often contends with low digital literacy, language barriers, and complicated application interfaces.
- The Paradox implementation provides a voice-activated, multilingual intake mechanism that automates applicant tracking and screening while eliminating manual scheduling tasks for HR staff.
For organizations hiring only a handful of people a year, manual interview scheduling and rescheduling might seem manageable. However, for a massive frontline organization that hires thousands of workers annually, the minutes wasted on routine coordination represent substantial financial losses. Automating these basic logistical steps directly improves the enterprise's bottom line. The overarching goal of the initiative is never to eliminate the human element from the process, but rather to eliminate low-value, repetitive "busy work" that does not require human creativity or nuanced judgment.
The recruiting platform automates the intake of candidate submissions, application parsing, job posting regeneration, and candidate communications. When an applicant is qualified, an AI assistant coordinates directly with them via text messaging and books time directly onto the recruiter's calendar without requiring manual intervention. Recruiter energy is purposefully concentrated on the high-value moments: conducting the actual interviews and making the hiring decisions. Once a human recruiter reaches a verdict, automated workflows immediately trigger the appropriate subsequent steps in the hiring pipeline. This tangible time-to-value is precisely why executive leadership embraces the technology: "This for us makes my CEO and my CFO love AI, because the dollars that are built into saving these minutes is really, really critical."
```mermaid
flowchart TD
accTitle: ISS Americas Automated Frontline Recruiting Workflow
subgraph Intake["Automated Top-of-Funnel"]
A["Job Application Received"] --> B["Automated Job & Candidate Processing"]
B --> C["Automated Text-Based Candidate Outreach"]
end
subgraph Scheduling["Automated Coordination"]
C --> D["Direct-to-Calendar Interview Scheduling"]
end
subgraph HumanDecisions["High-Value Human Touchpoints"]
D --> E["Recruiter Conducts Interview"]
E --> F["Human Hiring Decision"]
end
subgraph PostDecision["Automated Pipeline Execution"]
F --> G["Automated Workflow Execution & Next Steps"]
end
```
- **Candidate Scheduling:** Fully automated and text-based, allowing applicants to book directly into recruiter calendar slots without back-and-forth emails.
- **Strategic Focus:** Preserves human intervention strictly for candidate interviews, evaluation, and cultural onboarding from day one.
---
[14:59](https://www.youtube.com/watch?v=89vpsHo6RsU&t=899s) Impact Metrics: Reducing Frontline Hiring Time by 85%
Transforming unglamorous back-office corporate functions produces dramatic operational improvements, proving that optimizing everyday processes can deliver outsized value. By implementing end-to-end automated workflows for frontline recruitment, ISS Americas achieved an 85% reduction in overall frontline hiring time, significantly outperforming broader industry benchmarks for time-to-hire across candidate lifecycles.
Modernizing these workflows revitalizes functions that have historically been overlooked or sidelined during enterprise digital transformation programs. Eliminating administrative friction creates a streamlined, modern experience for prospective employees, reduces fatigue for HR personnel, and infuses digital-first thinking into core operational divisions.
- **85% reduction** in total frontline hiring turnaround time from application to hiring decision.
- **Competitive benchmark:** Outperforms standard industry averages for candidate cycle times across frontline enterprise roles.
---
[16:14](https://www.youtube.com/watch?v=89vpsHo6RsU&t=974s) Q&A: Next Functional Frontiers and the Flaws of Open-Ended POCs
When asked about the roadmap for rolling out generative and agentic AI to other corporate functions, internal corporate demand has surged with multiple departments seeking implementations. Top priorities include Legal—specifically focusing on contract strategy and automated legal documentation review—as well as Finance.
Addressing agentic AI initiatives and proofs of concept (POCs), Fournier expresses skepticism toward open-ended, non-committal testing, comparing poorly structured pilots to "dating without commitment" where organizations fail to decide whether they are actually "getting married." For enterprise POCs to deliver real value, they must be strictly time-constrained with clear exit criteria and backed by deeply committed operational stakeholders.
- **Target Corporate Expansions:** Legal (contract strategy, documentation workflows) and Finance operations.
- **POC Governance:** Avoid open-ended exploration; enforce time-boxed evaluation periods with committed functional business sponsors.
---
[17:34](https://www.youtube.com/watch?v=89vpsHo6RsU&t=1054s) Q&A: Human-Agent Boundaries and C-Suite Cultural Buy-In
Determining the operational boundaries between automated AI agents and human employees relies directly on functional domain expertise rather than technical mandates. HR leaders best understand the nuances of human interaction required during recruitment, and system workflows are iteratively tailored around where HR staff should focus their time and energy.
Navigating corporate culture and securing executive-level buy-in in traditional, non-IT organizations requires establishing deep alliances with departmental leadership. The recruiting initiative had previously languished on the back burner until a new Head of HR partnered directly with IT. Being "attached at the hip" with a dedicated business stakeholder—who understands the operational objectives regardless of their personal technical depth—is the single most critical factor in overcoming organizational resistance to AI transformation.
- **Task Allocation:** Governed by functional domain owners (e.g., HR determines which candidate evaluation stages require empathetic human judgment).
- **Executive Alignment:** Project viability relies on joint ownership between IT and functional department heads rather than top-down technical mandates.
Addressing how the organization quantified its 85% frontline hiring reduction, Fournier explains that the company's core business model already required tracking detailed operational hiring metrics prior to automation. This baseline made it straightforward to benchmark pre-AI turnaround times against post-deployment performance.
With end-to-end automation now in place, telemetry captures highly granular data at every stage of the funnel. The business tracks precise durations across every micro-step—from the moment an open job requisitions goes live, to the time it takes an applicant to complete a mobile application, to scheduling latency, through to final offer generation.
## Notable quotes
> [12:20](https://www.youtube.com/watch?v=89vpsHo6RsU&t=740s) This for us makes my CEO and my CFO love AI, because the dollars that are built into saving these minutes is really, really critical. — Alice Fournier
## Mentioned
Alice Fournier, ISS Americas, ISS Facility Services, Guckenheimer, Starbucks, McDonald's, Chipotle, Paradox, Paradox.ai
## Video description
Alice Fournier, Chief Information Officer, Americas, ISS Americas
Abstracts
AI properly applied to back-office corporate/back-office functions offers tremendous opportunities for forward-thinking organizations to beat the market. It is within these functions that the manual/repetitive tasks will be truly transformed, therefore unleashing greater creativity and human value-add.
Attendees will hear of a case study focused on AI in recruiting for a large front-line organization.
With the current focus on the consumer-facing impact of AI we may miss the greatest opportunity - the functions that keep the enterprise running.
ISS美洲借助招聘AI将一线招聘时间缩短85%
企业后勤工作流的自动化比面向客户的工具能产生更大的企业价值，因为消除日常行政阻力可直接保护营业利润率。
Alice Fournier认为，虽然数字化创新传统上侧重于面向客户的销售与营销，但最具变革性的企业价值实际上存在于人力资源、法务和财务等重复性后勤职能的自动化中。为解决遍布全球超32万名实体设施员工面临的严峻招聘瓶颈，ISS Americas引入了Paradox AI的多语言语音招聘平台。该系统实现了候选人信息录入、初筛以及直接同步至日历的面试排程全自动化，同时保留人工招聘人员负责关键评估和录用决策。这一针对性部署将一线招聘周期缩短了85%，表明严谨且与业务紧密结合的AI自动化能对企业底线利润产生巨大影响。
ISS Americas 首席信息官 Alice Fournier 指出，相较于早早拥抱数字化的前台部门，法务、财务、HR 等企业后台职能由于充斥着大量繁琐审批与手工流程，反而是 AI 自动化最具潜力的商业价值落地点。 (3:52)
针对一线岗位面临的数字化门槛与多语言痛点，ISS 引入 Paradox 对话式 AI 招聘平台，通过多语言语音交互与短信自动排程，不仅提升求职尊严，更大幅释放了招聘人员的行政琐事压力。 (7:45)
在人机协同模式下，AI 负责候选人预审与日历直接预约，HR 仅专注于实际面试与最终录用决策，从而推动 ISS Americas 一线招聘整体耗时降低了 85%。 (14:59)
Fournier 表示未来 AI 将进一步拓展至法务合同审核与财务流程，并强调企业 POC 必须设置明确时限与考核退出标准，依赖业务部门负责人的深度结盟而非盲目测试。 (16:14)
Reading summary
Alice Fournier, CIO at ISS Americas, argues that while generative AI is often marketed for customer-facing operations, its most significant financial impact lies in automating friction inside enterprise back-office functions.
Targeting Operational Drag in Corporate Functions (3:52)
While sales and marketing departments adopted digital transformations early due to direct consumer pressures, corporate back-office units such as human resources, finance, legal, and IT continue to shoulder manual tasks, ticketing systems, and complex approval workflows. Fournier argues that eliminating this administrative friction through targeted AI yields a stronger competitive advantage for enterprises than marginal productivity boosts in customer-facing divisions.
Resolving Frontline Recruitment Bottlenecks With Conversational AI (7:45)
Employing more than 320,000 workers globally across cleaning, facilities engineering, and dining operations, ISS Americas faces direct hiring competition for hourly workers against fast-food chains like McDonald's, Starbucks, and Chipotle. Traditional application portals presented significant hurdles for frontline candidates, many of whom navigate English as a second language, limited literacy, or lower digital familiarity.
To simplify access, ISS deployed Paradox, a mobile-first, voice-enabled, and multilingual recruiting tool. Applicants can speak directly to the interface, which translates and transcribes responses for automated pre-screening. Downstream, the platform automates job postings, candidate parsing, and text-based interview scheduling straight onto recruiter calendars, reducing the manual logistical overhead associated with hiring thousands of workers annually.
Preserving Human Judgment While Automating Pipeline Tasks (12:20)
Rather than attempting to eliminate human involvement, ISS configured the platform so that recruiters focus strictly on high-value interactions: interviewing candidates and making hiring decisions. Once a human recruiter issues a decision, automated workflows resume to trigger downstream onboarding steps. By eliminating manual scheduling friction, the organization achieved an 85% reduction in frontline hiring turnaround time from initial application to hiring decision, verified against pre-existing operational baseline metrics.
Governance, Departmental Alliances, and Functional Expansion (16:14)
Following the HR deployment, ISS is evaluating AI expansions into legal document review and finance workflows. Fournier advises against open-ended proofs of concept, warning that unstructured pilots often stall without clear business commitment. Instead, she advocates for strictly time-boxed evaluations with defined exit criteria.
Fournier also stresses that boundaries between automated agents and personnel should be determined by functional domain owners rather than IT mandates. At ISS, the recruiting deployment succeeded only after securing an active partnership with a dedicated Head of HR, highlighting the necessity of joint functional and technical ownership.
While the deployment established measurable cycle-time reductions in frontline staffing, extending agentic systems into complex domains like legal contracting and finance will depend on maintaining tight pilot governance and securing committed functional sponsorship.
精炼阅读摘要
ISS Americas 首席信息官 Alice Fournier 指出，生成式与企业级人工智能最具变革性的经济价值并非来自营销等前台职能，而在于消除法务、财务与人力资源等后台核心流程中的重复性操作摩擦，从而为传统重资产企业构建可持续的竞争壁垒。
企业后台职能的自动化潜能与竞争壁垒 (3:52)
Fournier 基于其消费品（CPG）与全渠道零售数字化转型背景指出，市场与销售等面向客户的前台部门早已因市场压力完成数字化，而法务、财务、人力资源和 IT 等后台部门长期承担着繁重的审批流程、工单流转和人工操作负担。作为在全球拥有超过 32 万名员工的设施管理巨头，ISS 需要在医院保洁、机场设施运维、暖通控制及企业餐饮（Guckenheimer）等劳动密集型场景下应对重返办公室趋势带来的空间运营重塑。
Fournier 认为，尽管生成式 AI 在营销创作端展现出增效作用，但企业真正的长期竞争力取决于后台运营效率。如果组织能够针对性地运用 AI 自动化解决后台行政拖累，消除重复性流程壁垒，其所释放的经济效益与敏捷度将远超常规的前台生产力改善，帮助企业在行业竞争中确立领先优势。
一线蓝领招聘瓶颈与多语言会话式工具 (7:45)
在设施服务这类依赖一线人力履约的企业中，从招聘、筛选到入职部署的周转周期直接决定了业务营收与履约交付能力。ISS Americas 不仅需要与星巴克、麦当劳、Chipotle 等部署了无摩擦招聘技术的快餐连锁（QSR）品牌争夺小时工劳动力，还面临着一线求职群体（如保洁员、厨师、电工等）数字化技能较弱、受教育程度有限以及英语非母语等实际障碍。传统的复杂在线招聘门户常让求职者产生挫败感甚至排斥心理。
为突破这一瓶颈，ISS 引入了会话式 AI 招聘平台 Paradox.ai。该方案采用移动优先架构，提供支持多语言的语音交互界面，允许不习惯打字的应聘者通过语音对话完成信息登记与初步资质预审。系统同时在上游接管了岗位发布与候选人寻访等机械化事务，大幅降低了求职门槛并消除了招聘人员的前期手动录入负担。
人机协同分工与端到端周期缩减 85% (12:20)
对于年招聘规模达数千人的大型企业，协调面试日程等琐碎事务会导致严重的工时浪费。ISS Americas 实施的 AI 方案并非替代人类评估，而是剔除低价值的行政事务：系统通过短信等渠道全自动协调合格候选人的日程，并将面试时间直接锁定在招聘人员的日历中。招聘团队得以将全部精力集中于核心面试交流与文化契合度判断等高价值环节；一旦招聘官做出录用决定，系统便自动触发后续流转环节，从而赢得管理层对投资回报的认可。
在成效方面，端到端自动化流程使 ISS Americas 的一线员工整体招聘周转时间缩减了 85%，表现大幅超越行业标准周期基准。由于公司此前已建立完善的运营指标监控体系，团队能够细致追踪从岗位发布、移动端填报耗时、排程延迟到最终发出 Offer 的每个微观节点，确保运营提速具有充分的数据佐证。
概念验证治理要求与业务主导的人机边界 (16:14)
在将 AI 进一步扩展至法务（如合同审查与法律文档处理）和财务等关键职能时，Fournier 对无明确期限的概念验证（POC）提出质疑。她将缺乏承诺的开放式测试比作“不以结婚为目的的恋爱”，强调企业级 POC 必须设定严格的时间限制、清晰的退出判定标准，并获得具备实质承诺的业务部门支持。
此外，AI 代理与人类员工之间的协作边界必须由业务领域专家定义，而非由 IT 技术条线单方面主导。例如，人力资源领导层最清楚候选人评估中何处需要同理心与专业判断，IT 则据此迭代工作流。该项目的成功关键在于 IT 与新任人力资源主管建立深度绑定合作，表明由业务部门主导目标、与技术团队协同共建，是克服传统组织转型阻力的核心保障。
ISS Americas 证明了将生成式与会话式 AI 嵌入非技术性后台业务所能释放的巨大效率，但这种成功高度依赖于精细的过程遥测、严苛的时间盒验证治理以及业务部门对人机权责边界的绝对主导；随着自动化向法务与财务等专业决策领域延伸，如何在非结构化流程中持续平衡风险控制与敏捷落地仍是企业需要验证的课题。
Detailed notes — zh
## Background
本内容由 Ai4 于 2025 年 8 月 24 日发布在 YouTube 平台上。视频记录了 ISS 美洲区首席信息官 Alice Fournier 的主题演讲，主要探讨将 AI 技术深度应用于容易被忽视的企业中后台职能（以一线员工招聘为例），以此消除繁琐的手动流程并释放组织核心竞争力的实践案例。
## TL;DR
ISS 作为拥有超过 32 万名员工的全球设施管理巨头，面临着庞大一线劳动力招聘的效率瓶颈以及与餐饮连锁品牌争夺候选人的激烈竞争。为此，ISS 美洲区团队引入了支持多语言与语音交互的对话式 AI 工具 Paradox.ai，将简历筛选、短信沟通与面试排期等机械琐事完全交由自动化接管，而让人类招聘官专注于面对面面试与最终录用决策。这一中后台变革成功将一线岗位的全流程招聘时间大幅缩短了 85%，不仅有力证明了后台职能自动化的巨大商业价值，也为法务、财务等其他中后台部门的 AI 落地提供了可复制的范式。
## Atmosphere & style
演讲风格务实、干练且极具同理心，Alice 巧妙地将个人职业背景的转型反思与一线蓝领求职者的真实困境结合起来，既有对“降低技术门槛、守护应聘者尊严”的人文关怀，又有以数据指标说话的商业魄力。在随后的问答环节中，她的回答直率幽默，例如将无休止的 POC 比喻为“不以结婚为目的的恋爱”，整场分享展现出一种深谙企业政治与组织变革逻辑的成熟领导者气场，令人倍感信服与启发。
## Key takeaways
- [00:04](https://www.youtube.com/watch?v=89vpsHo6RsU&t=4s) ISS 作为全球拥有超 32 万名一线员工的设施管理巨头，正身处后疫情时代“重返办公室”与重塑物理办公空间的变革核心。
- [03:53](https://www.youtube.com/watch?v=89vpsHo6RsU&t=233s) 传统数字化转型过度集中于前台销售与营销，而法务、财务、IT、HR 等中后台职能充斥着繁琐的手动审批与流程，是 AI 释放组织加速度的核心阵地。
- [03:53](https://www.youtube.com/watch?v=89vpsHo6RsU&t=233s) 现代企业 CIO 的核心使命是将传统被称为“拒绝部门”（Department of No）的 IT 角色转变为业务赋能与支持的部门。
- [07:26](https://www.youtube.com/watch?v=89vpsHo6RsU&t=446s) 一线员工到岗周期直接决定业务履约能力与营收，ISS 必须与星巴克、麦当劳等拥有极简应聘流程的连锁品牌在劳动力市场上激烈竞争。
- [07:26](https://www.youtube.com/watch?v=89vpsHo6RsU&t=446s) 针对一线求职者数字化素养有限及语言不通的痛点，招聘流程必须消除复杂技术交互带来的心理门槛与挫败感，保障求职者的尊严。
- [10:31](https://www.youtube.com/watch?v=89vpsHo6RsU&t=631s) 引入 Paradox.ai 对话式 AI 工具，通过原生多语言与全语音交互支持，让候选人直接语音对话即可完成申请与排期，大幅提升应聘完成率。
- [12:20](https://www.youtube.com/watch?v=89vpsHo6RsU&t=740s) 在海量招聘的一线企业中，面试协调与日程改期等微小环节节约的时间，能够直接规模化转化为企业底线（Bottom line）的真实利润。
- [12:20](https://www.youtube.com/watch?v=89vpsHo6RsU&t=740s) 坚持人机协同的分工设计：AI 负责线索抓取、岗位匹配与短信排期等机械琐事，人类招聘官专注于深度面试沟通与核心录用决策。
- [15:02](https://www.youtube.com/watch?v=89vpsHo6RsU&t=902s) 通过全流程自动化改造，ISS 美洲区一线员工端到端招聘时间大幅缩短了 85%，显著跑赢行业平均招聘周期。
- [16:13](https://www.youtube.com/watch?v=89vpsHo6RsU&t=973s) 在招聘场景跑通后，生成式 AI 与 Agentic AI 正加速向法务（合同解析与合同策略）及财务分析等企业职能部门拓展。
- [16:13](https://www.youtube.com/watch?v=89vpsHo6RsU&t=973s) 在企业技术验证上拒绝无实质承诺的漫长 POC，坚持设定严格时间盒（Time-constrained）并要求业务利益相关者做出实质推进承诺。
- [17:34](https://www.youtube.com/watch?v=89vpsHo6RsU&t=1054s) AI 与人类员工的权责边界划分应由精通业务的职能专家主导定义，技术团队负责灵活适配与敏捷迭代。
- [17:34](https://www.youtube.com/watch?v=89vpsHo6RsU&t=1054s) 成功推动传统企业转型的关键在于 IT 领导者与业务一把手深度结盟，并通过全链路细颗粒度的数据监控量化业务回报。
## Detailed walkthrough
### [00:04](https://www.youtube.com/watch?v=89vpsHo6RsU&t=4s) 走近 ISS：全球庞大一线劳动力的隐形巨头与“重返办公室”浪潮
Alice Fournier 开场向与会观众致谢，并表示将在接下来的十余分钟内分享 ISS 的智能化转型历程，探讨为什么在一家大型一线劳动力企业中实施招聘 AI 是水到渠成的必然选择。ISS（ISS Facility Services）是全球拥有最庞大员工群体之一的隐形巨头，在全球拥有超过 32 万名员工，专注于设施管理（Facilities Management）以打造优质的工作场所。从医院、制造工厂到机场卫生间的专业清洁，再到楼宇管理系统维护、HVAC（暖通空调）温控调节及灯泡更换等技术保障，甚至包含在美国运营的高端企业餐饮品牌 Guckenheimer，ISS 的业务渗透在实体建筑环境的方方面面。
在行业属性上，Alice 幽默地指出，设施服务业既不像快消品（CPG）或银行业那样光鲜亮丽，也不是酷炫的 SaaS 平台，而是纯粹“以人为核心的服务”（Services through people）。然而，正是这样一家传统企业，却意外站在了本世纪除 AI 之外最重大的社会与组织变革浪潮的中心——“重返办公室”（Return to Office）。如何通过重构实体物理环境来重新定义未来的工作模式，成为 ISS 身处其中的独特机遇与挑战。
- 全球员工总数超过 320,000 人，覆盖保洁、技术运维及餐饮等多元一线工种。
- 旗下餐饮品牌 Guckenheimer 在美国市场为众多企业客户提供高品质餐饮服务。
- 深度参与后疫情时代“重返办公室”的场景重塑与工作空间变革。
---
### [03:53](https://www.youtube.com/watch?v=89vpsHo6RsU&t=233s) 从“拒绝部门”到“赋能部门”：AI 在企业中后台职能的战略机遇
Alice 回顾了自己的职业生涯演进：此前她长期深耕于 CPG 领域的全渠道电商与数字化转型。在当时推动新技术与数据驱动分析工具上线的过程中，业务团队经常撞上被称为“拒绝部门”（The Department of No）的传统 IT 部门。这段经历促使她暗下决心，未来要亲自执掌 IT 并将其重塑为“提供支持的部门”（The Department of Yes）。而在出任 ISS 美洲区首席信息官（CIO）这三年来，她更深刻地认识到大型企业中后台职能的复杂现实与巨大潜力。
Alice 提出了一个核心战略假设：过去由消费者与客户需求倒逼的数字化浪潮，主要集中在销售和营销等前台早期采纳部门；而真正维系企业日常运转的核心中后台职能（包括法务、IT、财务和人力资源），往往充斥着大量繁琐的手动审批、流程流转和仪表板维护，尚未经历深度的技术跃迁。AI 在销售和通用员工生产力上固然表现出色，但一旦将其深度应用到承载着大量繁重手动流程的企业中后台职能中，企业便能获得无可比拟的组织加速能力，从而在各自行业中确立绝对的领先地位。
```mermaid
flowchart TD
accTitle: AI 在企业中后台职能中的赋能逻辑
subgraph FrontOffice["前台业务驱动 (早期采纳)"]
A["销售 / 市场 / 电商"] --> B["前台数字化与体验创新"]
end
subgraph BackOffice["中后台职能转型 (核心潜力)"]
C["法务 / 财务 / IT / 人力"] --> D["繁琐流程 / 审批 / 手动任务"]
D --> E["引入深度 AI 自动化流程"]
end
B --> F["提升前台转化"]
E --> G["释放组织加速度与核心竞争力"]
```
- 传统 IT 常被称为 "Department of No"，CIO 的使命是转向 "Department of Yes, and..." 的赋能定位。
- 核心假设：企业中后台（Corporate Functions）蕴含着最密集的流程与手动工作，是 AI 释放真正商业价值的核心阵地。
---
### [07:26](https://www.youtube.com/watch?v=89vpsHo6RsU&t=446s) 一线招聘的痛点：与连锁巨头竞争与消除求职者的技术门槛
对于像 ISS 这样依赖一线员工的大型组织而言，从发布招聘、面试筛选到新员工第一天顺利到岗，整个周期缩短得越快，对公司业务运营和直接收入的保障就越强。然而在现实劳动力市场中，ISS 必须与星巴克（Starbucks）、麦当劳（McDonald's）、奇波雷（Chipotle）等拥有强大技术支持和极速招聘流程的 QSR（快速简餐）巨头争夺一线候选人，候选人对极简应聘体验的期望已经被这些品牌拉得非常高。
更为严峻的挑战在于一线劳动者的群体特征。无论是在一线保洁员、厨师、管道工还是电工群体中，不少人数字化素养较低，甚至面临语言障碍或文化水平有限的问题。Alice 动情地提到，如果求职者因看不懂复杂的应聘流程或不懂英文而不知所措，会陷入一种深层的“羞耻与挫败感”（Position of shame）。因此，ISS 绝不能沿用传统复杂的企业级招聘系统，而是必须打造一个具备极高包容度、能够消除技术与语言门槛的前端应聘体验。
- 招聘到岗周期的长短直接与一线业务的营收及服务履约能力挂钩。
- 与快速简餐品牌（Starbucks、McDonald's 等）在劳动力市场上存在直接竞争。
- 重点关注技术弱势群体与移民群体的求职尊严，力求消除复杂技术交互带来的心理障碍。
---
### [10:31](https://www.youtube.com/watch?v=89vpsHo6RsU&t=631s) 引入 Paradox.ai：多语言与语音驱动的无门槛应聘体验
为了从根本上重塑招聘体验，ISS 引入了对话式 AI 工具 Paradox.ai。该工具彻底颠覆了传统的应聘流程：前端通过类似智能 ATS 的 AI 自动化追踪系统引导求职者，整个交互过程被简化到了极致。系统原生支持全流程多语言切换以及全语音交互（Voice-enabled）。这意味着不习惯键盘打字或不擅长英文书写的候选人，只需直接对着手机讲话，AI 就能自动完成语音识别、实时翻译并推进应聘步骤。
这种前端体验的升级极大改善了候选人对 ISS 的第一印象与申请完成率。而在 HR 后端，该系统开始全面接管传统招聘中繁琐的手动节点。HR 人员不再需要耗费大量时间进行职位手动分发、来回沟通调整面试日程或机械跟踪流程，AI 正在将人力资源团队从机械的行政劳动中彻底解放出来。
- 采用 Paradox.ai 构建全自动化前端应聘追踪流程。
- 系统具备全语音交互和多语言实时翻译能力，彻底消除打字与语言门槛。
- 推动 HR 部门告别职位发布、面试排期等海量手动操作，实现招聘流程的后端提速。
### [12:20](https://www.youtube.com/watch?v=89vpsHo6RsU&t=740s) 招聘全流程自动化与“人机协同”的分工设计
在规模庞大的一线员工组织（Frontline Organization）中，招聘往往涉及每年成千上万名员工的入职。如果一家公司每年只招 20 个人，那么协调面试、改期等零碎事务的影响并不明显；但对于需要海量招聘的企业而言，在安排面试、沟通时间等看似简单的环节上省下的每一分钟，都会直接转化为企业底线的真实利润（Bottom line）。这种真金白银的效率提升，也是推动 CEO 和 CFO 真正爱上 AI 技术的核心驱动力。
在自动化流程的设计哲学上，核心目标绝不是“完全替代人类”，而是将招聘人员从枯燥、重复、不产生实际人类价值且不需要创意的“忙碌琐事”（Busy work）中彻底解放出来。系统将候选人信息捕获、简历接收、职位描述重复生成、面试日历调度等环节完全交给 AI 自动化处理。招聘助手通过文本（短信）与候选人全天候互动，并直接与招聘官的日历实时联动完成排期，无需招聘专员手动介入。人类招聘官的精力则被高度集中在真正需要人性的环节——实际面试沟通、综合评估与最终录用决策；一旦录用决策完成，后续的工作流又会无缝切回自动化流转。
```mermaid
flowchart TD
accTitle: ISS 自动化招聘工作流与人机协同设计
subgraph S1["自动化阶段（AI 驱动）"]
A["候选人捕获与简历接收"] --> B["岗位发布与自动匹配"]
B --> C["AI 短信交互与日历自动预约"]
end
subgraph S2["人类核心价值阶段"]
C --> D["招聘官进行深度面试"]
D --> E["做出最终录用/淘汰决策"]
end
subgraph S3["后续自动化阶段"]
E --> F["Offer 发放与入职工作流触发"]
end
```
- **非人工干预环节**：包括候选人线索抓取、职位申请接收、职位发布信息复用、基于文本短信的面试排期与改期。
- **人类专属环节**：面对面/视频面试深度交流、候选人文化与能力终审决策。
- **财务与管理价值**：将零散琐碎的时间损耗转化为规模化节省，直接降低招聘单人边际成本。
### [15:02](https://www.youtube.com/watch?v=89vpsHo6RsU&t=902s) 一线招聘提速 85% 与后台数字化变革的商业回报
展示了一张小狗 Luna 在泥地里玩耍的轻松照片后，演讲者指出“看似枯燥的后台流程其实充满乐趣”，并公布了这套自动化招聘体系带来的惊人业务成果：企业的一线员工招聘时间（Frontline Hiring Time）大幅缩短了 85%。这一降幅彻底击败了行业平均的招聘周期（Time to hire），实现了端到端流程的实质性飞跃。
这项实践向企业管理层强有力地证明：当把目光投向企业内部长期被忽视、在以往数字化转型中被边缘化的后台与职能部门时，通过深度自动化改造，不仅能显著提升企业的整体竞争力与盈利能力，更能彻底革新求职者与 HR 团队的交互体验，真正将数字化技能和数字化思维深植于核心职能之中。
- **核心量化指标**：一线岗位从发布到录用的全流程招聘周期缩短 85%。
- **行业竞争力**：在候选人端到端转化速度上大幅超越同行业平均水平。
- **组织效益**：打破了过去数字化转型过度聚焦前台/消费端的偏差，使后台职能成为效率和竞争力的发动机。
### [16:13](https://www.youtube.com/watch?v=89vpsHo6RsU&t=973s) 问答环节：从法务、财务到 Agentic AI 与对“POC”的务实态度
在现场问答环节，参会者首先提问关于生成式 AI（GenAI）与智能体（Agentic AI）在其他企业职能部门的拓展计划。演讲者表示，随着招聘场景的成功，公司其他职能部门现在已经开始“排队”寻求 AI 落地。对于 ISS 而言，法务部门（Legal）是当前最受关注的领域，尤其是法律文档解析与合同策略（Contract Strategy），市面上已有许多出色的初创产品；紧随其后的是财务职能（Finance）。
针对企业内部是否开展智能体 AI（Agentic AI）的概念验证（POC），演讲者表达了对传统 POC 模式的鲜明态度。她将无休止的 POC 比喻为“不以结婚为目的的约会”（"dating without commitment"），直言自己更喜欢明确目标——“我们要么结婚，要么不结”。因此，尽管团队会开展验证，但必须设定严格的时间限制，且必须要求业务利益相关者（Business Stakeholders）展现出坚定的推进承诺，而非停留在浅层尝试。
- **重点拓展职能**：法务（合同审查、文档生成与合同策略）与财务分析领域。
- **智能体与 POC 策略**：拒绝漫长无结果的概念验证，坚持时间盒约束（Time-constrained）与业务端深度承诺。
### [17:34](https://www.youtube.com/watch?v=89vpsHo6RsU&t=1054s) 问答环节：人机边界划分、高管认同攻坚与全流程指标监控
随后，参会者就“如何界定 AI 与人类的分工边界”、“如何在传统非 IT 组织中克服文化阻力获取高管支持”以及“指标衡量的颗粒度”进行了深入探讨。
关于 AI 智能体与人类员工的边界划分，演讲者强调技术团队并不越俎代庖，而是完全依赖业务专家来定义规则。HR 最清楚哪些事情应当由人类完成，人机分工是业务团队与技术团队共同探讨演进的结果，并在初期上线后根据实际反馈进行了敏捷迭代。
```mermaid
flowchart LR
accTitle: 推动企业级 AI 落地的关键要素
A["业务与IT紧密绑定<br/>(与业务主管紧密协作)"] --> B["清晰的人机分工边界<br/>(业务专家主导定义)"]
B --> C["高管层支持与文化认同<br/>(聚焦核心业务价值)"]
C --> D["端到端细颗粒度指标度量<br/>(全链路追踪提效数据)"]
```
在谈及克服老牌企业文化阻力、获取 C-Suite（高管层）认同时，演讲者分享了背后的曲折经历。该项目此前曾被搁置在冷板凳上，原因在于当时的 HR 负责人缺乏兴趣；直到新任 HR 负责人上任，双方“紧密绑定、协同推进”（"attached at the hip"），项目才得以迅速破局。她强调，推动 AI 落地必须绑定一个强大的业务负责人，对方不需要精通技术，但必须深刻理解业务想要达成的目标。在指标衡量方面，得益于公司的商业模式，团队从岗位发布、申请提交到发放 Offer 的每一个微观环节均有详尽的耗时追踪，从而精确量化出自动化前后端到端 85% 效率提升的具体表现。
- **人机边界界定机制**：由精通业务的职能团队主导决策，技术系统提供灵活适配，并在版本迭代中持续微调。
- **高管认同与组织变革**：IT 领导者必须与业务部门一把手（如 Head of HR）深度绑定结盟，共同向管理层输出价值。
- **指标度量机制**：依托成熟的业务数据基线，端到端精细化监控“申请耗时”、“排期间隔”、“发 Offer 周期”等各环节指标。
## Notable quotes
> [00:04](https://www.youtube.com/watch?v=89vpsHo6RsU&t=4s) Services through people — Alice Fournier
> [03:53](https://www.youtube.com/watch?v=89vpsHo6RsU&t=233s) The Department of No — Alice Fournier
> [03:53](https://www.youtube.com/watch?v=89vpsHo6RsU&t=233s) The Department of Yes — Alice Fournier
> [07:26](https://www.youtube.com/watch?v=89vpsHo6RsU&t=446s) Position of shame — Alice Fournier
> [12:20](https://www.youtube.com/watch?v=89vpsHo6RsU&t=740s) Busy work — Alice Fournier
> [16:13](https://www.youtube.com/watch?v=89vpsHo6RsU&t=973s) dating without commitment — Alice Fournier
> [16:13](https://www.youtube.com/watch?v=89vpsHo6RsU&t=973s) 我们要么结婚，要么不结 — Alice Fournier
> [17:34](https://www.youtube.com/watch?v=89vpsHo6RsU&t=1054s) attached at the hip — Alice Fournier
## Mentioned
Alice Fournier, ISS, ISS Americas, ISS Facility Services, Guckenheimer, Paradox.ai, Starbucks, McDonald's, Chipotle, Luna
## Video description
Alice Fournier, Chief Information Officer, Americas, ISS Americas
Abstracts
AI properly applied to back-office corporate/back-office functions offers tremendous opportunities for forward-thinking organizations to beat the market. It is within these functions that the manual/repetitive tasks will be truly transformed, therefore unleashing greater creativity and human value-add.
Attendees will hear of a case study focused on AI in recruiting for a large front-line organization.
With the current focus on the consumer-facing impact of AI we may miss the greatest opportunity - the functions that keep the enterprise running.
[Source](https://www.youtube.com/watch?v=89vpsHo6RsU)
Source quotes
[0:04] Alice Fournier: Services through people
[3:53] Alice Fournier: The Department of No
[3:53] Alice Fournier: The Department of Yes
[7:26] Alice Fournier: Position of shame
[12:20] Alice Fournier: Busy work
[16:13] Alice Fournier: dating without commitment
[16:13] Alice Fournier: 我们要么结婚，要么不结
[17:34] Alice Fournier: attached at the hip
Source quotes
[12:20] Alice Fournier: This for us makes my CEO and my CFO love AI, because the dollars that are built into saving these minutes is really, really critical.
Original transcript
[0:04] Well, good afternoon. Thank you so much
[0:06] uh for coming to spend what will
[0:08] hopefully be um an interesting next
[0:12] couple of minutes, 15 20 minutes. I do
[0:15] intend on keeping a little time for Q&A.
[0:17] So, as I as I go through our journey and
[0:20] I um go through the presentation, please
[0:24] make sure that you um keep any
[0:26] questions. hopefully there's time to ask
[0:28] them right now and if not I'm certainly
[0:31] um absolutely happy to answer them
[0:34] afterwards. So the way we're going to
[0:38] run the next 18 minutes or so is that
[0:40] I'll tell you a little bit about who we
[0:41] are as a company um and why the
[0:44] implementation of AI in recruitment
[0:47] actually makes a ton of sense for a
[0:49] company like ours. Um I'll tell you also
[0:51] a little bit about myself and then we'll
[0:53] get into the how um and the case study.
[0:55] So um we are ISS facility services. Now
[1:00] ISS has is one of the companies with the
[1:03] largest workforces uh in the world that
[1:06] you have most likely never heard of. So
[1:10] what we do is facilities management. We
[1:13] have upwards of 320,000
[1:16] employees across the world. It's a huge
[1:19] workforce and we take care of we make
[1:22] workplaces great. So if you think about
[1:24] it, what does a great workplace require?
[1:27] First of all, it's got to be clean. So
[1:30] we clean in the built environment all
[1:32] over the world. We clean hospital,
[1:35] manufacturing plants, airports. Think
[1:37] about the airport's toilet. It's
[1:40] important that they are clean. Um it
[1:42] really adds or detracts from the
[1:44] experience. Um we also have a large
[1:48] technical workforce. So, think of a
[1:50] building like this one. Changing light
[1:52] bulbs, making sure that the building
[1:54] management system functions, HVAC,
[1:56] managing temperature, of course, and
[1:59] most important in many ways, not most
[2:01] important, but certainly as important,
[2:03] we also have a great corporate catering
[2:06] brand named Guggenheimr here in the US
[2:08] because of course, great workplaces, uh,
[2:11] you don't want to be left hungry. So, we
[2:13] serve great food. Now, one of the
[2:15] interesting things about ISS is that we
[2:18] were thrown in the middle of the
[2:21] redefinition, probably the biggest
[2:24] crisis or opportunity that this century
[2:27] has to offer next to AI and and in many
[2:30] ways connected and that's return to
[2:33] office. Um, we are not and I as I was
[2:35] sharing this morning as in an earlier
[2:37] presentation, we are not the sexy
[2:39] industry. We're not CPG. We're not
[2:41] banking. We're certainly not cool, you
[2:44] know, SAS platforms. We do services
[2:48] through people. We do genitorial,
[2:50] technical. And so to think that a
[2:52] company such as ours was really thrown
[2:54] into being part of how we were going to
[2:58] redefine work partly through the built
[3:00] environment for the way forward was
[3:03] actually and has been really really cool
[3:05] and kind of an unexpected societal
[3:08] crisis to be a part of. Um so let me run
[3:11] a quick video uh two minute not even 30
[3:14] second which will give you an overview
[3:16] of more on what ISS does um and then
[3:19] we'll jump into the next bit
[3:22] [Music]
[3:52] All right, great. So, it gives you a
[3:53] quick overview. This is what we do and
[3:55] then a little bit about myself because
[3:57] this is a a bit of the why we're having
[3:59] this conversation right now. Um so I've
[4:02] spent the bulk of my career more on the
[4:04] CPG side in digital transformation that
[4:07] was driven by consumer requests by
[4:09] customer requests really moving through
[4:12] e-commerce omni channel and that entire
[4:14] disruption. Now in that journey of many
[4:18] years and as e-commer the e-commerce
[4:21] world was building up and growing and
[4:23] changing and increasingly what we were
[4:26] looking for was to onboard all these new
[4:28] applications on board new technologies.
[4:31] Very quickly we moved into a world that
[4:33] required datadriven analysis. So all
[4:36] kinds of new skills that we were all
[4:38] very or generally uncomfortable
[4:41] unnowledgeable about. Um but we were
[4:45] doing quite well and I was sitting more
[4:46] in the sales and marketing and digital
[4:48] transformation and yet kept hitting my
[4:52] head against the wall of it. The
[4:55] department of it which I lovingly
[4:58] referred to as the department of no.
[5:02] And so very quickly I told myself well
[5:06] one day I'm going to go lead it and turn
[5:09] it into the department of yes. And what
[5:11] I say is that now three years in my role
[5:14] as chief information officer heading IT
[5:17] I'm the department of yes and uh so not
[5:20] quite yes uh because I've learned better
[5:23] and all of course all of the challenges
[5:26] of of heading it in large um large
[5:30] corporate environments but in that
[5:33] journey um what I have learned through
[5:35] the digital transformation consumer
[5:38] customerf facing is that that has really
[5:42] pushed these functions to learn new
[5:45] skills and really one of my big
[5:48] recognition now heading it is that many
[5:50] of the corporate functions and of course
[5:52] that will depend on the the organization
[5:54] you're part of but many of our corporate
[5:57] functions the functions that really
[6:00] power up your enterprise and if you take
[6:03] one second no matter which organization
[6:05] you're in you can think of legal you can
[6:08] think of IT you can think of finance
[6:10] finance. These corporate functions are
[6:14] core to your business moving at speed,
[6:17] moving well. But in many cases, with
[6:20] some exception, but in many
[6:22] organization, these departments have not
[6:24] yet gone through the accelerated digital
[6:28] transformation that you might have seen
[6:30] on the sales and marketing side. And my
[6:33] hypothesis is actually that once AI gets
[6:37] implemented really really successfully
[6:40] in the corporate functions, we are going
[6:43] to see an acceleration of of
[6:45] organizations. And so my hypothesis is
[6:48] really AI is a phenomenal tool for
[6:52] employee productivity in general for
[6:55] sales for marketing still who are
[6:57] generally early adopters. But once you
[7:00] bring it in and you properly deploy it
[7:02] in these corporate functions where a lot
[7:04] of the business process, a lot of the
[7:07] manual work resides, do it well in those
[7:10] functions and you will defined your you
[7:13] will be the leader in your industry no
[7:16] matter the industry and uh my belief
[7:20] actually led me as the CIO to do that
[7:23] with one of our functions. So if we move
[7:25] into the next slide um oh this is the
[7:29] boring stuff which is my favorite photo
[7:31] to show the boring stuff but um so
[7:34] really thinking about again processes
[7:37] requests approval dashboards this is why
[7:39] it makes sense in corporate function so
[7:40] let's move to the next slide so at ISS
[7:45] um as I mentioned before we are a large
[7:48] frontline organization so if you are in
[7:50] frontline organizations you know one
[7:53] thing and one thing for sure and that's
[7:56] that the faster you can hire recruit all
[8:00] the way to have an employee ready to
[8:03] start on day one the faster you can do
[8:07] that journey the faster and more
[8:10] successful you will be in your business
[8:13] and in our cases there's direct revenue
[8:16] implications so when I joined the
[8:19] organization that journey was a really
[8:21] long one now you have to think that we
[8:23] are competing comping in a world where
[8:26] we're competing against Starbucks,
[8:28] McDonald's, Chipotle, all these
[8:30] phenomenal QSR, quick service restaurant
[8:34] brands who have a lot of technology were
[8:38] moving really quickly and our employees
[8:40] have those or candidates really have
[8:42] those expectations. So, we were looking
[8:45] for a way to really shrink down that
[8:48] journey and make it a journey that they
[8:50] could relate to. So what we ended up
[8:52] deploying and I'm not I prefer to say
[8:55] I'm not paid by this company to talk
[8:57] about them but this is genuinely
[9:00] something that has impacted our
[9:02] business. Um and so we deployed this
[9:05] tool called paradox.ai AI and I'll tell
[9:08] you a little bit about our why and how
[9:10] we did it uh because it was so material
[9:12] and it's a real success story of
[9:16] implementing AI in a way that transform
[9:19] a corporate function and accelerates
[9:22] business success. So think about our
[9:26] people. So in your mind, think of a
[9:28] janitor, right? A chef, um a plumber, an
[9:33] electrician. In many cases, especially
[9:36] when you think about janitorial
[9:38] workforces, you have people who are much
[9:41] less technologically savvy, right? In
[9:44] some cases, we work with people who are
[9:47] who have very low level of literacy, let
[9:49] alone technological literacy. But if you
[9:52] think about I don't know if in your
[9:54] families um maybe you have immigrants in
[9:57] your family uh maybe you have people who
[9:59] are frontline workers less education if
[10:02] you think about the position of shame
[10:05] that you can be put into when you apply
[10:08] for a job and you're struggling to even
[10:11] understand the basics right because you
[10:13] don't understand the language because
[10:15] the technology the experience is too
[10:18] complicated that's important and we
[10:20] really really value how we present
[10:23] ourselves as a company to our candidates
[10:26] and potential employees. And so for
[10:29] those of us who are comfortable with
[10:31] technology, we don't generally although
[10:33] as mo things move quickly, we we have to
[10:36] learn quickly and keep the pace. But
[10:38] this sentiment of being completely
[10:41] uncomfortable with a technology
[10:42] experience is not something that we
[10:44] necessarily get to experience much, but
[10:47] our people certainly do. And so to bring
[10:50] to them a technology that would really
[10:54] create a front-end experience that was
[10:56] comfortable for somebody who may be less
[10:59] comfortable with technology or may not
[11:01] speak English. This was one of the core
[11:04] decisions why we chose this from a
[11:06] candidate experience. So basically what
[11:09] the system does is that it automates the
[11:13] application process. It has it really
[11:16] follows the applicant um through AI
[11:19] tracking the the applicant like a
[11:21] traditional ATS and it it really makes
[11:24] the whole process super simple. It's
[11:27] also uh completely multilingual,
[11:30] completely voice enabled. So if typing
[11:32] is not your thing, it's not comfortable.
[11:34] When you speak into the system, it will
[11:37] recognize, it will translate, it will
[11:39] adjust. So from a candidate experience,
[11:43] deploying this AI tool really made a
[11:46] difference to how people responded to
[11:48] us. Now in the back end is even more
[11:52] interesting because
[11:54] think of your HR department, right?
[11:57] Think of the number of steps that go
[12:00] into hiring somebody, right? I have to
[12:04] find candidates. I have to recruit them.
[12:06] I have to post the job. I have to make
[12:08] sure that I get the candidates in. I
[12:10] have to organize um all kinds of
[12:13] interviews. I have to change interviews.
[12:15] All these things are manual steps and
[12:18] they're manual steps that actually cost
[12:20] businesses money. So if you think about
[12:23] all the time, right, all the time spent
[12:26] just looking to schedule and reschedu an
[12:29] interview. Now if you're hiring 20
[12:32] employees a year, that's one thing. If
[12:34] you are a large frontline organization
[12:37] having to to hire thousands of people
[12:40] every year, the minutes that you save on
[12:44] automating things as simple as um
[12:47] organizing interviews, planning
[12:48] interviews are m is money that goes
[12:51] right to the bottom line. So let's look
[12:54] at the next step, the next slide. And so
[12:57] this will give you an idea of how the
[13:00] automation, how the process works
[13:02] because the idea here is not to
[13:04] completely remove the human, right? We
[13:07] are not looking to remove the human. Nor
[13:10] should we. We are really looking to
[13:12] accelerate the tasks that are busy work,
[13:15] that are not material, that are not
[13:18] human-driven, that don't require
[13:20] creativity. And so what we have built is
[13:23] a system that automates in blue all the
[13:26] things that I really don't need a person
[13:29] doing. So I don't need a person really
[13:32] capturing all the candidates, receiving
[13:35] all the job applications. I don't need
[13:37] somebody having to recreate a job
[13:40] posting and application every time.
[13:43] Interview scheduling. In this case, what
[13:45] this enables us to do is to go schedule
[13:48] automatically right in the Oh, we can go
[13:51] back one. U this enables us um to
[13:54] schedule the the employees um
[13:57] automatically right in um uh recruiter's
[14:01] calendar. So the recruiter really
[14:03] organizes connects the assistant to
[14:05] their calendar. The interviews get
[14:08] scheduled automatically and this is all
[14:10] textbased with the candidate. So that
[14:12] the recruiter really does not need to
[14:15] invest any time in that. Now where do we
[14:17] want the recruiter to invest their
[14:19] energy of course on interviews hiring
[14:22] decisions the rest of the process is
[14:26] also automated once I've interviewed
[14:28] once I've made a decision on a candidate
[14:31] do I want to hire them do I not want to
[14:33] hire them then in my workflows I can
[14:36] determine what are the next steps all of
[14:38] it automated. So this for us makes my
[14:43] CEO and my CFO love AI. Uh because the
[14:47] dollars that are built into saving these
[14:50] minutes is really really critical. And
[14:53] of course I want a human to be a great
[14:56] human on day one. So moving to the next
[14:59] slide. Um I often talk about right why
[15:02] is this important and the boring stuff
[15:04] can be fun. Uh so this is Luna having
[15:07] fun in the dirt. Um but what does that
[15:12] impact? We decreased frontline hiring
[15:14] time by 85%.
[15:17] For us this is absolutely material. We
[15:20] beat the a industry average on time to
[15:23] hire the time it takes us to take a
[15:26] candidate from beginning to end. Um and
[15:28] of course you know material efficiencies
[15:31] you saw everything we automated and that
[15:34] is critical to the business. So with
[15:37] that um and I'll take two minutes for
[15:39] questions but really for us it has been
[15:43] absolutely convincing that when we take
[15:46] all these back office corporate
[15:48] processes that sit in many of your
[15:51] functions we automate them we have made
[15:54] our company a lot more successful while
[15:57] transforming the candidates experience
[15:59] HR's experience and really bringing
[16:02] digital skills and digital thinking to a
[16:04] critical function
[16:06] but one that's been largely forgotten or
[16:09] put aside in the overall digital
[16:10] transformation. So with that, are there
[16:13] any questions? Oh yes, yes, I saw your
[16:16] hand first.
[16:18] >> Wonderful.
[16:19] >> Thank you.
[16:21] Which
[16:22] >> which other corporate functions are you
[16:24] rolling out geni to next? And are you
[16:28] doing aentici's
[16:30] proof of concepts in your organization?
[16:32] Thank you. Great presentation.
[16:33] >> Yeah, thank you. Um so um the the truth
[16:38] is uh amazingly enough I I now have the
[16:41] functions a little bit lining up. U the
[16:43] piece that and and again for us you have
[16:45] to think about what makes most sense for
[16:47] your business. Um legal is something I'm
[16:50] very very interested in. Uh there's so
[16:53] much great work out there and great
[16:55] startups on legal documentation. Uh
[16:57] contract strategy is a big thing for us.
[16:59] Um, and of course finance. Um, and yes,
[17:02] we are doing um, agent thick. And on the
[17:04] PC, I'm actually not a huge fan of PC's.
[17:07] I was referring earlier to it's a little
[17:10] bit like dating, right? It's like dating
[17:12] without commitment. Um, I like to just
[17:15] figure out are we getting married or are
[17:16] we not? Uh, and so yeah, so and I I
[17:20] prefer I mean we run them but they have
[17:22] to be time constraint. I need business
[17:24] holder business stakeholders really
[17:26] committed. Um, so but we are running
[17:28] them. Yes.
[17:30] Other questions? Yes. And then two more.
[17:32] Yeah.
[17:34] >> Hello. Yeah. How do you draw a line
[17:36] between what an agent should do versus
[17:38] what a human should do?
[17:39] >> I rely on business stakeholders um who
[17:42] know their function. So, um HR knows
[17:46] what humans should do best than HR. Um
[17:49] and so that's how we work together. It's
[17:52] a discussion. We could have built it
[17:54] every any which way. Um the system
[17:57] allows for that. But we were really
[17:58] focused on where do we want our people
[18:01] to spend their time. Um and and actually
[18:04] the first version we we've modified some
[18:07] things.
[18:08] >> Good. Good afternoon. Um some some of
[18:10] the problems that we face in getting AI
[18:13] installed in a an older organization
[18:15] especially nonIT organizations
[18:18] >> is getting buy in at the seuite level.
[18:20] Yeah. And when you were showing this
[18:22] demonstration of how you're effectively
[18:24] reducing headcount in the HR department,
[18:27] >> how did you overcome those cultural
[18:29] obstacles in in selling AI to seuite?
[18:32] >> Um attached at the hip with the head of
[18:35] HR. Um so this could not have happened
[18:38] and in fact was sitting in the back
[18:41] burner. I was very excited about it.
[18:45] head of HR in that moment was was not
[18:49] into it and when the new head of HR came
[18:53] it was a joint effort um and and that's
[18:57] absolutely critical. I need and this is
[18:59] backed I need a strong business
[19:01] stakeholder who doesn't need to be the
[19:03] most technically savvy but needs to
[19:06] understand what is it we're trying to
[19:07] achieve. Yeah, I saw that you had a lot
[19:09] of steps in your workflow
[19:11] >> and you said you got 85% uh you know
[19:14] front line decreased and uh these
[19:17] metrics. So are you looking at each and
[19:18] every step or just an aggreate like how
[19:20] are you doing it from the when the human
[19:22] was doing it to when the AI was doing
[19:23] it? How do you measuring that?
[19:25] >> Yeah, for sure. So we know exactly how
[19:27] much time it takes to hire somebody
[19:29] right from from the moment you put a job
[19:32] out till the moment. Um, and part of it
[19:34] is because of our business model where
[19:37] we track these metrics. Um, so it was
[19:40] rather easy to to track how that would
[19:43] play out. Um, and so we can track and
[19:46] now of course everything automated. We
[19:48] know exactly how long uh we have very
[19:51] detailed uh metrics that I can't share,
[19:54] but we know exactly, you know, how long
[19:57] it takes to apply for a job um how long
[20:00] it takes to have an offer. Um and the
[20:02] numbers are just really really
[20:05] something. Um so we do track it partly
[20:07] because of our our business model.
