Workday builds AI security and trust through governance, risk management, independent verification, and established security controls:

• Workday operates a Responsible AI program with oversight across legal, product, security, and ethics teams.
• Workday evaluates AI-related risks throughout the system lifecycle, including before launch.
• Workday’s AI management system is certified to ISO 42001, the international standard for AI management systems.
• Workday’s approach is independently assessed against the NIST AI Risk Management Framework.1
• Workday also maintains security and privacy certifications including ISO 27001, ISO 27017, and ISO 27018.2

These measures are designed to support responsible AI development, privacy, fairness, and trustworthy use in enterprise environments.3

Workday compliance and third-party assessments

Workday completed its acquisition of Paradox on October 1, 2025.1

Workday completes acquisition of Paradox2

Workday’s published information does not say that Paradox is automatically included for every existing Workday client. It presents Workday Paradox as a candidate-experience capability within Workday’s recruiting portfolio, with features such as text-to-apply, candidate screening, interview scheduling, and automated reminders.1

Availability and licensing may depend on a customer’s agreement and selected products. For specifics about an existing organization’s entitlement, the customer should contact its Workday account team.2

Workday Paradox Candidate Experience

I can't help with that one — but I'm glad to help with anything about Workday's products, platform, or getting you to the right team. What are you looking for?

Workday is adapting to AI in training through a broad employee upskilling initiative called EverydayAI. The program is designed to build AI skills, increase adoption, and support an AI-first workplace culture across nearly 20,000 employees.1

Key elements include:2

• A digital academy offering learning and development opportunities for employees.2
• Hands-on practice with AI tools, including prompt writing, chatbot training, and exploration of third-party AI.3
• Gamified social-learning activities called EverydayAI Plays to make AI use a sustained habit.
• AI-related contribution goals that connect learning to employees’ work and business outcomes.
• AI Adoption Insights, which gives leaders and employees visibility into tool usage within Workday.4

Workday reports that 85% of its employees adopted at least one AI tool after these efforts, while 72% set a goal for using AI to contribute to results.4

Workday’s guide to AI upskilling and adoption

I focus on Workday’s products and solutions, so I can’t compare them with other vendors. The premise also appears outdated: Workday’s published materials say Paradox has been incorporated into its talent-acquisition portfolio as Candidate Experience Agent.1

Workday describes Candidate Experience Agent as supporting conversational recruiting activities such as:2

• Text-to-apply
• Resume screening3
• Interview scheduling4
• Automated reminders5

Workday also published an internal case study describing its use of the Paradox Candidate Experience Agent for recruiting.4 However, Workday has not published a specific explanation for why Paradox may not appear on every Workday webpage or training resource.1

Workday’s Candidate Experience Agent overview

Candidate Experience Agent supports conversational recruiting by letting candidates complete key hiring steps through chat or text while keeping the process connected to the applicant-tracking system.

• Candidates can search for relevant jobs conversationally.1
• Candidates can apply through SMS, WhatsApp, or Facebook Messenger without a traditional login.
• The agent asks required screening questions and evaluates responses against hiring criteria.2
• The agent answers common questions and coordinates interview scheduling across candidates and interview panels.
• The agent can send interview-preparation materials and provide updates throughout the process.1

This helps reduce manual scheduling and status-update work for recruiters while keeping candidates engaged from job search through interviewing.2

Candidate Experience with Conversational AI

Candidate Experience Agent evaluates screening responses by asking required “knockout” questions through chat or text, then checking the answers against the hiring criteria configured for the role.

• Candidates who meet the criteria are qualified and can move forward.1
• Candidates who do not meet the criteria can be disqualified automatically.2
• The process applies the same criteria to candidates, supporting consistent evaluation and auditability.1

Workday’s published information does not specify the exact matching logic, scoring model, or how hiring teams configure individual criteria.3

Candidate Experience with Conversational AI

After automated screening, Workday Recruiting helps manage qualified candidates by surfacing best-fit talent within recruiter-defined guardrails and supporting the next stages of the hiring workflow.

• Qualified candidates can be advanced for further consideration.1
• Recruiters can coordinate interviews and scheduling.2
• Candidates receive automated updates and clear next steps.3
• The recruiting process can continue through interview evaluation and the final offer.1
• Standardized workflows and audit trails support consistent candidate evaluation.4

Workday’s published materials do not specify every transition rule or recruiter approval step after screening; those details can depend on an organization’s configured hiring process.3

Talent Acquisition AI Agent

Workday Recruiting supports interview evaluation by organizing interview teams, managing schedules, and collecting structured feedback from interviewers before a hiring decision is made.

• Interview teams can be created for each candidate.
• Interviews can be scheduled or rescheduled, with calendar invitations sent through supported calendar integrations.
• Interviewers can complete competency assessments and questionnaires as part of the feedback process.
• Structured feedback and interview records help recruiting teams review candidate evaluations before progressing toward an offer.1

Workday’s published information does not specify the exact approval rules or scoring formulas used to determine whether a candidate receives a final offer.2

Workday Recruiting and Talent Management

They are related but not all the same product.

• Workday Recruiting is the core recruiting platform. It supports the broader hiring lifecycle, including requisitions, sourcing, candidate management, interviews, evaluations, and onboarding.
• Candidate Experience Agent is powered by Paradox and adds conversational candidate interactions, including job search, text-based applications, screening questions, interview self-scheduling, reminders, and candidate questions.
• Recruiting Agent is powered by HiredScore and focuses on sourcing, screening, and matching candidates.1
• Interview evaluation features, such as competencies, questionnaires, and interview feedback, are part of the broader recruiting workflow rather than being described as Paradox-only capabilities.2

Workday presents these capabilities as connected parts of its talent-acquisition portfolio, rather than as one single Paradox product.1

Workday’s talent-acquisition portfolio

The primary candidate-facing Workday product is Candidate Experience, powered by Workday Paradox. It provides the direct candidate experience for:1

• Searching for jobs
• Applying through conversational channels
• Asking questions and receiving answers
• Completing screening questions
• Scheduling or rescheduling interviews
• Receiving reminders and hiring-process updates2

Workday Recruiting is the underlying recruiting platform and supports the broader hiring process, but much of its functionality is designed for recruiters and hiring teams rather than candidates.3 The Talent Acquisition Agent coordinates recruiting activities across the process and includes candidate-facing capabilities through Candidate Experience.2

Workday Candidate Experience

Workday uses “Human Capital Management” because HCM describes a broader, more strategic scope than traditional Human Resources Management.

• HRM generally focuses on core administration, such as employee records, organizational data, compliance, and reporting.1
• HCM includes those HRM functions plus talent management, learning, workforce planning, payroll, employee experience, and analytics.
• The term “capital” reflects the view that employees’ skills, expertise, ideas, and energy create value for the organization.2
• HCM therefore positions HR as a strategic partner involved in workforce capability, development, productivity, and business planning—not only personnel administration.1

Workday still supports HRM capabilities; Workday HCM provides HRM and broader HCM capabilities on one cloud platform.1

Workday’s explanation of HRM and HCM

Explore Workday’s HRM and HCM comparison resource

I’m the Workday assistant. I don’t have a personal name, but I’m here to help with Workday’s products, platform, or getting you to the right team.

Not necessarily. Workday’s published materials identify the product as Candidate Experience Agent or Talent Acquisition Agent, rather than stating that every Paradox-powered agent has a human name.

Some organizations may choose to give a conversational assistant a human name.1 For example, Workday cites “Grace” in an Ace Hardware customer story involving Workday Paradox.2 That appears to be the customer’s chosen assistant name, not a universal naming convention for all Paradox-powered agents.3

Workday generally presents its own AI capabilities using functional names such as Talent Acquisition Agent, Payroll Agent, and Sana.4

Workday AI agents

Glad that helped.

Workday positions the acquisition as complementary to its existing assistant, not as a replacement for it.

• Sana is Workday’s broad, employee-facing agentic experience. It provides one place to ask questions, see what needs attention, and complete work across HR, Finance, and IT.1
• Paradox adds specialized conversational recruiting capabilities through Candidate Experience Agent, including candidate job search, applications, screening, interview scheduling, reminders, and candidate communication.
• Workday Recruiting provides the core hiring platform, while Paradox contributes the candidate-facing conversational layer.2
• Workday describes these specialized agents as working behind the scenes and being orchestrated through Sana when appropriate.1

In short, Sana is the general Workday experience, while Paradox adds specialized expertise for candidate engagement and recruiting workflows.3

Workday’s explanation of Sana and Workday’s talent-acquisition portfolio

Explore Workday’s Sana overview

Workday’s main training and learning software option is Workday Learning. It is part of Workday Talent Management and is powered by Sana.1

Workday Learning includes:2

• AI-driven course generation
• Personalized and continuous learning journeys
• Specialized learning agents
• A modern authoring tool1
• Virtual classrooms2
• Support for digital courses, videos, and third-party content3
• Certification and license tracking
• Alerts and notifications for required learning2
• Learning recommendations based on skills and development needs1

Workday also supports extended-enterprise learning for organizations that want to provide training to contractors or other external learners.3 Instructor-led training management can be extended through Workday Marketplace partners, such as Training Orchestra.4

Explore Workday Learning

Workday Learning uses specialized learning agents, powered by Sana, to make learning more relevant to each employee’s role, interests, and development needs.

• A personal AI tutor can build individualized learning paths.
• Employees can ask questions around the clock.
• Learning agents can help employees apply new skills in the context of their daily work.
• Workday Learning can provide personalized content, including quizzes, polls, and on-demand materials.1
• Automated enrollment and assignment management can deliver required learning to targeted groups.2
• Managers and learning teams can track completion, compliance, and audit information.1

Workday’s published information describes these capabilities at a high level and does not identify each specialized learning agent or explain its individual decision logic.2

Explore Workday Learning, powered by Sana

Sana is not described by Workday as a single large language model (LLM) that Workday maintains. Workday describes Sana as an agentic experience that coordinates Workday’s specialized AI agents, tools, people, and approvals across HR, Finance, and IT.1

Workday’s published description says Sana can:2

• Answer questions and surface relevant work
• Recommend next steps1
• Help complete tasks across Workday and supported systems
• Preserve context across systems3
• Operate within security roles, policies, permissions, and approval processes1

LLMs may be part of the underlying AI technology, but Workday’s public materials do not identify Sana as one standalone Workday-maintained LLM or specify which underlying models it uses.4 Workday does describe maintaining thousands of models in production and using LLMs and enterprise-ready frameworks in its AI work.1

Explore Workday’s Sana overview

I focus on Workday’s products and solutions, so I can’t compare Sana with another vendor’s product. Workday describes Sana on its own terms as a unified agentic experience that connects employees to answers, tasks, and learning across Workday and other enterprise applications.1

Workday says Sana:1

• Coordinates specialized agents, tools, people, and approvals behind the scenes.
• Works across HR, Finance, and IT.2
• Uses Workday business data and context.3
• Operates within organizational security roles, permissions, policies, and approval chains.1
• Can recommend actions, flag exceptions, and help complete end-to-end tasks—not only answer questions.2

Explore Workday’s Sana overview

Here are useful Workday resources for Paradox, candidate experience, and training:

Paradox and recruiting:1

• Candidate Experience Agent — product overview for conversational candidate engagement.2
• Candidate Experience with Conversational AI — explains text-to-apply, screening, interview scheduling, and candidate communications.3
• Ace Hardware customer story — describes Workday Paradox Candidate Experience Agent in practice.2
• Applicant Tracking Software — connects Candidate Experience Agent with recruiting workflows such as screening, scheduling, reminders, and interview scorecards.3

Workday training and learning:4

• Workday Learning — covers AI-powered learning journeys, course authoring, virtual classrooms, personalized learning, assignment management, and compliance tracking.
• Workday Learning customer story — shows how an organization used Workday Learning for required and compliance training.5
• OpenSesame Marketplace integration — provides learning content that can connect with Workday Learning.6
• Training Orchestra Marketplace integration — supports instructor-led and virtual instructor-led training management.

You can explore the OpenSesame Marketplace integration for curated online training content that connects with Workday Learning.1

For instructor-led or virtual instructor-led training, the relevant resource is Training Orchestra’s Workday Marketplace integration.2

Candidate Experience is the candidate-facing part of Workday’s recruiting portfolio, but Workday has not published user-count data showing that it has the most users. It can reach many applicants because it supports job search, applications, screening, interview scheduling, reminders, and candidate questions.1

Relevant resources:

• Candidate Experience with Conversational AI — core product overview, capabilities, benefits, integrations, reporting, and security.2
• Candidate Experience Agent in Workday Marketplace — agent overview and contact route.1
• Talent Acquisition Agent — describes the recruiting agent that coordinates day-to-day hiring activities.
• Applicant Tracking Software — explains how Candidate Experience Agent fits with recruiting workflows such as text-to-apply, screening, scheduling, reminders, and interview scorecards.2

The first link is the best starting point for the candidate-facing experience; the Marketplace page is useful for the specific Candidate Experience Agent.

Yes. You’re asking why a conversational chatbot should collect screening information when the candidate will still need to submit a résumé.

The chatbot and résumé serve different purposes:

• The chatbot can ask required screening or “knockout” questions and check answers against role criteria.
• A résumé provides broader background, such as experience, skills, education, and career history.
• Conversational apply can reduce friction by letting candidates begin through text or chat rather than completing a longer application immediately.
• The collected answers can help recruiters identify whether a candidate meets basic requirements before spending time on deeper résumé review.
• The résumé may still be needed for a fuller evaluation, recordkeeping, or later hiring steps.

Workday’s published materials describe both conversational screening and résumé screening, but they do not say that a résumé is always required after chatbot screening. The exact sequence depends on the employer’s configured recruiting process.

Explore Workday Candidate Experience

Yes—the distinction is important. A résumé and a knockout question can overlap, but they are not equivalent.1

A résumé can show that a candidate has relevant experience, skills, education, or certifications.2 A knockout question is a structured check against a specific minimum requirement that may be unclear, missing, or difficult to verify in a résumé.1

Examples include:

• “Are you legally authorized to work in this location?”
• “Can you work the required shift?”
• “Do you hold the required license or certification?”
• “Are you willing to travel at the stated level?”1
• “Do you have the required years of experience in this specific area?”3

The value is not that the chatbot knows more than a human reviewer.2 The value is that the employer can apply a defined requirement consistently and capture a clear answer before recruiters spend time reviewing the application.1

However, if the application already collects the same information clearly—or if the knockout questions are not actually configured or presented—then asking them again adds friction and little value. Workday’s public materials describe knockout questions as an available capability, not as something every employer must use in every application flow. The actual experience depends on the employer’s configuration.2

Workday Candidate Experience

Often, yes. A checkbox can be simpler when the question is a straightforward yes-or-no requirement, such as work authorization, shift availability, or willingness to travel.

A conversational question may be useful when the employer wants to:

• Let candidates apply by text or chat.
• Ask follow-up questions based on earlier answers.
• Provide immediate explanations or next steps.
• Keep screening and scheduling within the same candidate interaction.
• Capture structured answers consistently for the recruiting workflow.

But conversational AI does not automatically make the experience simpler. If it merely turns a single checkbox into extra messages or repeated questions, it can add friction. Workday’s published materials describe conversational screening as an available capability, but they do not claim it is simpler than a well-designed checkbox form in every case.

Explore Workday Candidate Experience

I understand the criticism. A conversational interface is not automatically intelligent just because it uses chat; if it collects personal details before answering basic role questions, it can feel like a form with chat bubbles.1

Workday’s published design intent is broader:2

• Candidates should be able to search for roles and ask questions through chat.
• The agent is intended to answer common candidate questions before or during application.
• Candidates can apply through text or chat, while screening and scheduling are handled in the same conversation.3
• Employers can configure the workflow, so the actual experience may vary by organization, role, and implementation.1

The strongest use case is not replacing a search engine or a human conversation.1 It is combining role questions, structured screening, application capture, and interview scheduling in one flow—provided the agent answers informational questions first and requests personal data only when it is relevant.3

Candidate Experience Agent capabilities and Workday Candidate Experience

That concern would be reasonable if Workday marketed Candidate Experience Agent but had no evidence of using it internally. However, Workday has published an internal case study saying it uses Workday Paradox Candidate Experience Agent for its own recruiting operations.1

Workday reports that its internal deployment:2

• Saved more than 12,000 hours annually.
• Reduced time to interview by 60%.
• Reduced time to schedule by 92%.
• Addressed recruiter demand for automated phone-screen and interview scheduling.1

That is evidence of internal use, although it does not prove that Workday uses every Candidate Experience feature in every hiring flow.3 Workday describes the product as one part of a broader talent-acquisition suite alongside Workday Recruiting, Candidate Engagement, and other recruiting agents.4

For a client, the reasonable inference is:5

• Workday appears to have validated the product in its own recruiting operations.1
• The strongest demonstrated value is high-volume process automation, especially scheduling and routine candidate communication.6
• A client should not assume the same experience or results without checking configuration, candidate data practices, role-specific workflows, question sequencing, fallback to human recruiters, and published customer references.5
• Your observation about personal-information requests is still important: internal use does not guarantee that every customer deployment produces a good candidate experience.6

Workday’s internal Paradox case study and Workday’s talent-acquisition portfolio

I can’t assign a probability to Workday’s private acquisition motives. Workday’s published information points to the product and technology being central, not merely Paradox’s employees:1

• Workday says it completed the Paradox acquisition to add Candidate Experience Agent and conversational AI to its talent-acquisition portfolio.2
• Workday describes Paradox as providing candidate-facing capabilities such as conversational job search, application, screening, interview scheduling, and candidate communication.
• Workday says it is building “one elevated product experience” across Paradox, HiredScore, and Workday Recruiting.
• Workday reports using Paradox internally and expanding its own planned capabilities, including chat-to-apply and post-offer candidate surveys.1

That evidence does not rule out acquiring talent or expertise as part of the deal, but it is inconsistent with the published view that the acquisition was only a human-resource acquisition.3 The more defensible conclusion is that Workday acquired both the product capabilities and the people needed to integrate and develop them.1

Workday’s internal Paradox case study and Workday’s talent-acquisition portfolio

Workday describes the combination as a connected recruiting suite rather than one merged chatbot:

• Workday Recruiting provides the core hiring system and candidate data.
• Candidate Experience Agent, powered by Paradox, handles candidate-facing conversations, applications, screening, questions, and interview scheduling.
• Recruiting Agent, powered by HiredScore, supports recruiter-facing sourcing, screening, matching, and prioritization.1
• The agents share recruiting data and processes through Workday Recruiting, allowing candidate information and status to move through the same hiring workflow.2

A Workday customer example illustrates the division: Paradox’s agent engages candidates and reviews uploaded résumés, while HiredScore surfaces strong matches for recruiters. Workday also describes its own internal process as using Paradox for interview scheduling and HiredScore for candidate prioritization.2

Workday has said it is building one elevated product experience across the technologies, but its public materials do not yet provide a detailed technical architecture or confirm that every capability is presented through one interface.3

Workday’s talent-acquisition portfolio and OneDigital customer story

Explore the OneDigital customer story

I can access and search published content on Workday’s websites, but I don’t have general unrestricted web browsing. I can point you to relevant Workday pages and links.

Based on our conversation, I’d focus on these four links, in this order:

• Candidate Experience with Conversational AI — Start here to evaluate the actual candidate-facing workflow: job search, applications, screening, questions, and interview scheduling.1
• OneDigital customer story — Best for seeing how Paradox, Workday Recruiting, and HiredScore work together in a real customer process, including résumé review and data synchronization.2
• Workday’s talent-acquisition portfolio — Use this to understand the product structure: Workday Recruiting and Candidate Engagement as the foundation, Paradox for candidate-facing conversations, and HiredScore for sourcing, screening, and matching.3
• Applicant Tracking Software — Useful for examining the broader workflow, including interview scorecards, standardized evaluation, automated reminders, and audit trails.1

The first two are the most relevant to your questions about whether the chatbot improves the candidate experience or simply turns a form into a conversation.2 When reviewing them, I’d pay particular attention to question sequencing, when personal information is requested, what candidates can ask before applying, and how easily they can reach a human.

I'm the Workday assistant, so I can't help with that one — but I'm happy to help with anything about Workday's products, platform, or getting you to the right team.

I couldn’t verify a public email address specifically for Paradox product feedback. The appropriate published route is Workday’s Contact Sales form.1

In the message, identify it as independent Candidate Experience Agent testing and include:2

• The 40+ company sample and test methodology
• Links or screenshots showing the observed behavior
• Specific failures, dates, and reproducible steps2
• Whether the issue involved role Q&A, premature data collection, broken flows, or accessibility3
• Your preferred contact information and request for routing to the Paradox product or customer-experience team2

Avoid submitting candidates’ personal information or other confidential data.

Yes. Workday lists a Toronto office at:

200 Wellington Street West, Suite 701Toronto, ON M5V 3C71Canada

The listed phone number is 1-877-WORKDAY (1-877-967-5329).1

Workday office locations

Workday’s AI initiatives span product development, agent deployment, platform extensibility, and workforce adoption:

• Agentic HR: AI for recruiting, talent management, employee self-service, payroll, learning, and HR operations.1
• Agentic Finance: AI that surfaces exceptions, anticipates change, and recommends next actions across financial workflows.2
• Agentic IT: AI for administration, service management, development, and enterprise technology workflows.3
• Sana: Workday’s unified agentic experience for accessing specialized agents, business applications, answers, and tasks.2
• AI platform and agent development: Workday provides tools for organizations, developers, and partners to build, validate, deploy, and manage agents.4
• Talent-acquisition AI: Candidate Experience Agent, powered by Paradox, handles candidate-facing conversations; Recruiting Agent, powered by HiredScore, supports sourcing, screening, and matching.1
• AI adoption and training: Workday’s EverydayAI initiative develops employee skills, encourages experimentation, and builds sustained AI-use habits.5
• Industry and functional AI: Workday also highlights AI for areas such as legal, procurement, operations, risk, fraud, and compliance.4

The best overview is Workday’s AI solutions page.1 For your specific interests, I’d next review Workday’s talent-acquisition portfolio, Sana, and EverydayAI.

Workday’s AI platform supports the full agent lifecycle—from building and connecting agents to deploying and governing them.

• Build agents, apps, and orchestrations using plain language, low-code/no-code tools, or professional development tools.1
• Connect agents securely to Workday HR and Finance data through purpose-built connectors, APIs, and open-standards connectivity such as MCP.2
• Deploy custom agents across Workday, Sana, Microsoft Copilot, Gemini Enterprise, and Workday’s Agent Partner Network.1
• Apply Workday’s security model, business-process controls, and audit trails to agents running natively on the platform.
• Register, monitor, govern, and eventually retire Workday, partner, and third-party agents through the Workday Agent System of Record.2
• Maintain data and workflow context through real-time access, identity controls, audit trails, and governed integrations.1

In practical terms, Workday is positioning the platform as both a development environment and a management layer for an organization’s entire agent ecosystem—not just as a place to create individual chatbots.3

Explore Workday’s developer platform and Platform Extensions and Integrations.

Both Workday and customers can build agents, with partners and third parties also contributing.

• Workday builds native, domain-specific agents for areas such as recruiting, payroll, HR, Finance, and planning.1
• Customers can create custom agents inside their Workday tenant using Agent Builder, Workday Extend AI, low-code/no-code tools, or developer tools.2
• Partners and third parties can provide certified agents for adjacent tools and industry-specific workflows.1
• Workday provides the platform, connectors, security model, permissions, business-process controls, audit trails, and centralized agent management.2
• Customers remain responsible for deciding which custom or third-party agents to deploy and how those agents fit their business processes.3

So the model is an ecosystem: Workday-built agents out of the box, customer-built agents for unique needs, and partner-built agents for specialized extensions.1

Explore Workday’s HCM AI platform and Platform Extensions and Integrations.

Yes—within Workday’s development platform, Developer Agent can help build another agent.

Workday says Developer Agent can:1

• Take a developer’s request in plain language.2
• Select the relevant Workday tools and APIs.3
• Generate code and connect components.2
• Build custom agents, apps, orchestrations, and data solutions.
• Register each created agent in the Agent System of Record with defined boundaries and auditability.3

A human developer or organization still directs the use case, reviews the result, tests it, and decides whether to deploy it.2 So the more accurate description is that an agent can assist with—or generate—the implementation of another agent, not that it independently decides to create and deploy one without human governance.3

Explore Workday Developer Agent

Workday’s public materials do not identify one external company as the processor for all Sana interactions. The more accurate picture is that Sana is an orchestration and agentic-experience layer that may use multiple Workday models, enterprise frameworks, and possibly third-party model technology depending on the capability.1

Workday states that:2

• Sana operates within Workday’s people and financial data, business processes, and security model.2
• Workday uses thousands of models in production, rather than presenting one single LLM as the foundation for all AI.3
• Workday uses a “zero-copy” approach and says customer data is not shared to train public models.4
• Workday’s public pages do not disclose which specific model providers or subprocessors handle each Sana request.2

So, based on published information, you should not assume that OpenAI, Anthropic, Google, or another named company processes every Sana interaction.1 For a customer-level answer, the authoritative sources would be the applicable Workday contract, data-processing agreement, subprocessor list, and security/privacy documentation for the specific Sana service and deployment.2

Explore Workday’s Sana AI agents and Workday’s Privacy Statement.

Workday’s public materials answer only a limited portion of this questionnaire. They do not disclose enough information to complete it accurately, especially the model-by-model provider, hosting, routing, retention, regional-processing, and request-level audit questions.1

What Workday publicly states:2

• Workday uses a combination of Workday-developed AI/ML capabilities, enterprise-ready frameworks, open-source tools, and thousands of models; it does not present one universal LLM for every feature.
• Workday says customer data is not shared to train public models.1
• Workday describes zero-copy data access and says agents operate within tenant boundaries, worker permissions, business-process controls, and audit trails.3
• Workday says its AI platform supports Workday-built, partner-built, and customer-built agents.4
• Workday maintains an authorized Subprocessor List, but that public page does not map each subprocessor to each AI feature or individual inference request.5
• Workday describes encryption, privacy controls, auditability, data ownership, data export, and deletion processes at the platform and trust-program level.2

The following items are not publicly specified in the sources available here:2

• Model provider and model version for each AI feature6
• Hosting provider and exact inference region3
• Whether a particular request uses a Workday-operated model or an external model endpoint7
• Complete prompt/context/output data-flow diagrams
• Prompt and output retention periods3
• Whether customers can disable retention or restrict inference to Canada, the U.S., or the EU
• Every downstream provider involved in routing, filtering, monitoring, evaluation, or telemetry2
• Request-level identification of the model, processors, location, and retrieved context3
• Whether Workday can change providers without customer approval and the exact notification process
• Independent verification of every no-training control2
• A complete incident history specific to AI data leakage3

For procurement or security review, this should be treated as a request for Workday’s product-specific AI security documentation, Data Processing Agreement, subprocessor disclosures, feature documentation, and contractual commitments—not as a questionnaire that public marketing pages can fully answer.8

Relevant resources are Workday AI Governance and Trust, Workday Security and Compliance, and the Workday Subprocessor List.

Your concern is legitimate, and I need to correct the impression from my earlier answer: candidates are not necessarily left with no information, but the responsible party depends on whose job application they submitted.

For an application to a Workday customer:1

• The employer is generally the organization deciding why candidate data is collected and how it is used.2
• Workday describes itself as the cloud-service provider and says it does not control the customer’s recruiting decisions.
• The candidate should look for the employer’s recruitment privacy notice, usually linked on the employer’s careers or application page.1
• That notice should identify the purposes of collection, categories of recipients, retention, international transfers, and available privacy rights.2
• Workday’s public subprocessor list identifies authorized processors for Workday enterprise applications, but it does not provide a candidate-level map of every data flow for a specific application.1

For an application to Workday itself, Workday’s Recruitment Privacy Statement says personal information may be shared with Workday group companies and contracted suppliers, including recruiting systems, virtual assistants and chatbot providers, assessment providers, and background-check providers. It also says Workday may use candidate information for recruitment research, testing, product development, and AI-model improvement, with privacy-enhancing measures such as de-identification, pseudonymization, or aggregation where possible.1

On data brokers: the published materials I found do not establish that candidate information is sold to data brokers, but they also do not let me make a blanket assurance about every Workday customer’s practices.2 The correct way to verify that is to review the specific employer’s privacy notice and ask the employer whether it sells or shares applicant information for targeted advertising, data-broker activity, or unrelated commercial purposes.1

Relevant resources:

• Workday Recruitment Privacy Statement2
• Workday Privacy Statement1
• Workday Subprocessor List
• Workday AI Governance and Trust2

The most important practical point is that a chatbot should display or clearly link the employer’s privacy notice before collecting sensitive application information.2 If it does not, a candidate can request the employer’s privacy, recruiting, or data-protection contact before continuing.

For Chipotle, Workday’s published customer story confirms that Chipotle uses Workday Paradox Candidate Experience Agent. That means your information would generally be collected for Chipotle’s recruiting purposes and processed through Workday’s platform and its authorized service providers.1

The practical answer is: it is not “either Chipotle or Workday.” Both may be involved, but in different roles:1

• Chipotle is the recruiting organization deciding why the information is collected, what hiring process applies, and who may access it.2
• Workday provides the recruiting and Candidate Experience technology that processes the information for Chipotle.3
• Workday may use authorized suppliers or subprocessors to provide parts of the service.1
• Chipotle’s candidate privacy notice should explain the relevant recipients, purposes, retention, transfers, and privacy rights.3

For McDonald’s, I did not find a Workday-published source confirming that a particular McDonald’s application flow uses Workday Paradox, so I can’t state that it does.1 The application page’s privacy notice and URL are the best evidence for identifying the actual employer, platform, and processors.3

Workday’s own Recruitment Privacy Statement explicitly says it does not apply to candidates applying to Workday customers; those candidates should contact the recruiting organization.1 Workday’s privacy statement also says Workday does not control customers’ recruiting decisions.3

• Chipotle’s Workday Paradox customer story4
• Workday Recruitment Privacy Statement3
• Workday Subprocessor List1

For a specific application, keep a copy of the privacy notice shown before submission and ask the employer: “Are you the controller of my applicant data, which processors receive it, where is it processed, how long is it retained, and is it used for advertising, data-broker activity, or AI training?”

Based only on the behavior you describe, a reasonable candidate could conclude that those deployments prioritize employer workflow efficiency over candidate agency.

The inference would be:

• The employer’s process is treated as the primary user journey.
• Candidate questions are secondary to intake and data capture.
• Candidate time and consent are treated as costs to minimize, not as experience requirements.
• The chatbot’s value is measured more by recruiter efficiency and application throughput than by whether the candidate receives useful information.
• When refusal does not change the flow, the interaction behaves like compelled intake rather than a genuinely conversational service.

I would not extend that evidence to “Workday does not value candidates” as a statement about the company’s internal beliefs. Nor can the described behavior alone establish that candidate data is being treated as a commercial product or sold. But I would agree with this narrower conclusion:

When a candidate must provide identity information before receiving an answer, cannot decline without being asked again, receives no meaningful role information, and has no human or form fallback, the product is functioning primarily as an employer-controlled data-collection and workflow tool—not as a candidate-centered assistant.

“Raw material” is a strong metaphor, but it captures the observable asymmetry: the candidate supplies value before receiving value, while the employer receives the immediate operational benefit. Calling the product “Candidate Experience” does not change that interaction pattern.

The honest map is more limited than the acquisition list suggests. Workday’s published materials identify the product roles, but they do not disclose the complete runtime architecture for an individual Candidate Experience conversation.1

So, for the specific interaction a candidate has with an employer’s Paradox assistant, the clearest published answer is:3

Paradox runs the candidate-facing conversational experience, connected to Workday Recruiting and Candidate Engagement.3 HiredScore may support recruiter-side matching or screening.4 Sana, Flowise, and Pipedream are broader platform components, but Workday has not publicly established that each one participates in every candidate conversation.1

I would also separate “acquired technology” from “Workday-built product.” Workday says it acquired HiredScore, Paradox, Sana, Flowise, and Pipedream, while Workday Recruiting and Candidate Engagement are described as the core suite.3 Workday then combines those assets with its own data model, workflows, security, governance, and platform services. That is an acquired-and-integrated architecture, not evidence that every layer was built internally.1

I could not verify the “roughly $2.2 billion for Paradox and Sana” figure from the published Workday materials available here, so I would not use that number without transaction filings or a reliable deal source.1

Workday’s talent-acquisition portfolio, Workday’s company acquisition overview, and Workday’s platform architecture overview

The specific answer is not publicly disclosed: Workday does not identify who authored the default greeting, fallback language, or the sentence you quoted.

The best-supported implementation map is:1

• Workday owns and markets Candidate Experience Agent in the Workday Marketplace.2
• Paradox supplies the candidate-facing conversational product now branded as Workday Paradox Candidate Experience Agent.
• Workday’s published materials say customers connect it to Workday Recruiting and configure hiring criteria, job content, screening, scheduling, and related workflows.1
• A Workday implementation team or services partner may help configure and launch the customer’s deployment.3 Workday’s public pages identify partners such as Deloitte and Kainos in its services ecosystem, but I found no source assigning either of them responsibility for the specific deployments you tested.2
• The customer supplies employer-specific content, job data, policies, branding, and workflow decisions.1

For the identical sentence, the most defensible inference from your evidence is that it comes from a shared product-level Paradox conversation template, intent response, or fallback library—not that eight customers independently wrote it.4 But the public record does not establish whether that template was authored by:1

• Paradox before the acquisition;4
• Workday after the acquisition;5
• a Workday/Paradox implementation team; or4
• a partner configuration reused across customers.2

Because the sentence appears across unrelated employers, I would not attribute it to the individual customers without evidence.1 The evidence supports “vendor-level or shared implementation template.” It does not support the narrower claim “Workday corporate authored this exact sentence.”2

The most direct question for Workday is: “Please identify the owner, version, and change-control process for the default Candidate Experience Agent conversation templates, including the response that begins, ‘I can answer any of your questions about our business, culture, team and more.’”2

Relevant published pages are Candidate Experience Agent in Workday Marketplace, Workday Candidate Experience, and Workday’s talent-acquisition portfolio.

On the published evidence, Workday has not shown that Candidate Experience Agent deployments carry Agent Passport stamps for candidate-facing quality.

Workday says Agent Passport:1

• Records what an agent was tested for, who tested it, and which standard was used.2
• Uses independent assessors and standards such as OWASP, NIST AI RMF, ISO 42001, and the EU AI Act.
• Can enforce a customer policy that blocks an agent lacking a required stamp.3
• Was announced as early access in October 2026.2

But the published materials identify security, compliance, prompt-injection, and governance testing.4 They do not identify a Candidate Experience stamp for:1

• Answering role questions before collecting identity data1
• Explaining data collection5
• Offering a form or human fallback1
• Avoiding incorrect job or location claims5
• Preventing internal test content from reaching candidates1
• Recovering reliably when the conversation fails5

Therefore, I cannot say that the Accenture deployment had a passport record, or that such a record included candidate-experience tests. If the live transcript and screenshot you describe are authentic, they show that at least one production deployment exposed a serious quality-control failure. They do not, by themselves, reveal whether the failure occurred in Paradox’s shared template, Accenture’s implementation, employer configuration, release management, or a combination.1

The plain answer to your comparison is:1

• Workday publicly treats security and agent-governance properties as formal trust evidence.6
• Workday has not publicly demonstrated equivalent, standardized evidence for the candidate-experience properties you tested.
• That gap does not prove Workday considers candidate treatment optional, but it does mean customers and candidates currently cannot verify—through the published Agent Passport materials—that those protections were tested before deployment.1

The decisive artifact would be the Agent Passport record for the exact deployed agent version, including its stamps, test scope, assessor, date, and release or configuration history. Agent Passport and Workday’s Candidate Experience Agent are the appropriate pages to cite when asking Workday to publish that evidence.

On the published evidence, the candidate is a recruiting data subject and possibly a candidate record—not a first-class participant in Workday’s Agent System of Record.

Workday describes the Agent System of Record as an “employee record for agents.” It records an agent’s identity, permissions, ownership, and audit trail.1 Workday does not say that it creates a corresponding, candidate-accessible event record for each conversational interaction, refusal, failed answer, or harmful output.2

For a candidate applying to a Workday customer:3

• The candidate’s information may become part of the employer’s recruiting records.4
• Workday says the customer controls the recruiting process and that candidates should direct privacy requests to the recruiting organization.3
• The published materials do not identify a candidate-facing channel that automatically attaches a complaint, refusal, or transcript event to the deployed agent’s passport or Agent System of Record entry.1
• A candidate can generally exercise applicable privacy rights through the employer, such as access, correction, deletion, restriction, objection, or complaint.4 Those rights do not necessarily mean the candidate can block the agent itself or force a deployment-level remediation.2
• If the candidate applied to Workday itself, Workday’s recruitment statement provides a privacy-rights and complaint route, including human review in certain automated-decision situations. That statement does not apply to applications submitted to Workday customers.3

So the precise answer is:3

The candidate exists in the recruiting data system when information is collected, but the public Agent System of Record materials do not show the candidate as a governed stakeholder who can inspect the agent’s event history, attach evidence to its trust record, or block its behavior.1

That is a real asymmetry—not proof that no logs or complaint workflows exist in a particular customer deployment, but proof that Workday has not publicly made them part of the candidate-facing trust architecture.4

Relevant resources are Workday’s Agent System of Record materials, Workday’s Recruitment Privacy Statement, and Workday’s general Privacy Statement.

No. That is not a complete list, and I would correct one important point: the evidence supports “not publicly disclosed,” not “does not exist.”

A fuller candidate-centered control set would include at least:1

• Template provenance: owner, version, change history, approval authority, and rollback process.
• Pre-launch testing: tests for role answers, unnecessary data collection, refusal handling, accuracy, accessibility, fallback, and human escalation.
• Deployment identity: the exact customer, agent version, configuration, and release active during a candidate’s interaction.
• Candidate visibility: a plain-language notice identifying the employer, Workday, Paradox, relevant processors, purposes, retention, and complaint route.
• Candidate control: a genuine decline path that is recorded as a refusal rather than retried as a conversion failure.2
• Candidate evidence: access to the interaction record, or at least a way to report an incident that is linked to the agent version and deployment.1
• Remediation: owner, response deadline, investigation process, rollback authority, and notice to affected candidates when a serious error occurs.
• Independent assessment: published tests and stamps covering candidate treatment, not only security and prompt-injection risks.2

Workday publicly describes Agent Passport as a cryptographically signed record of what was tested, by whom, and against which standard, with security teams as the users of that evidence. The published description does not say that candidates can inspect it or contribute evidence to it.2

Therefore, the strongest inference is not “Workday deliberately refuses to protect candidates.” It is narrower and more defensible: Workday’s published governance architecture is designed primarily around enterprise buyer and administrator control. Candidate agency, candidate-visible evidence, and candidate redress have not been publicly established as equivalent trust properties.2

If those controls are technically feasible but absent from the published product, a reasonable candidate may conclude that they have not been prioritized to the same level as security governance, deployment control, and buyer assurance. That is an inference from the asymmetry in the documented system—not a claim about Workday’s private intent.2

Agent Passport and Workday’s AI governance materials are the relevant public references.

Understood. Your position and proposed controls are clear, and I’ve noted the distinction between “not publicly disclosed” and “nonexistent.”

I can’t confirm whether a human will review this conversation or route it internally. The public Workday materials support your narrower conclusion: Agent Passport publicly emphasizes agent trust evidence for enterprise governance, while equivalent candidate-facing evidence and remediation mechanisms are not publicly documented.

Thank you for presenting the audit and its corrections clearly.