Workday limits the use of its AI agents, insisting on adherence to corporate rules
Read more
TechCentral
techcentral.co.za

Workday limits the use of its AI agents, insisting on adherence to corporate rules

Workday aims to position itself as an AI company that strictly adheres to established rules. At the key event Workday Rising Emea in Barcelona on Wednesday, leaders of the software giant for human resources and finance emphasized one idea: artificial intelligence in systems responsible for payroll calculation and financial closing must remain within the limitations set by the company.

Co-founder and CEO Aneel Bhusri, who returned to this role in February, stated that it is impossible to perform book closing or calculate salaries based on probabilities. He noted that finance and human resources are inherently deterministic, and one cannot guarantee salary calculation accuracy at 90% because this percentage can change with every operation performed.

Bhusri argued that the rules embedded in Workday's processes make it safe to combine these processes with probabilistic AI. He explained that 'our AI is legitimate; it is based on business rules that you trust and that are built into your system.'

Critique of the autonomous enterprise concept

Workday President for Products and Technology, Gerrit Kazmaier, spoke out against broader concepts. He questioned whether the idea of an 'autonomous enterprise' sounds cold and soulless, referencing another company. Although he did not name this company, SAP made 'autonomous enterprise' a central theme of its Sapphire conference in May, presenting this strategy to partners in South Africa in Kyalami in July, and repeating it at the TechEd developer event this month.

Other advantages promoted by Workday include integration and control. The company integrates Sana, an AI interface acquired after purchasing a Swedish startup of the same name last year, into the tools of other companies. The Sana self-service agent is already functioning in Microsoft Copilot and Teams, Slack, and Google Gemini Enterprise, and is in early access in Anthropic's Claude Cowork. In Barcelona, Workday announced early access to a Copilot plugin that connects Sana to a wider range of company agents. Developers can create Workday agents in tools such as Claude Code, Cursor, and OpenAI's Codex.

Furthermore, Workday plans to acquire Honeydew, whose 'semantic layer' allows any AI tool to apply its own definitions of metrics, such as headcount, consistently.

Regarding control, Workday demonstrated identity management from C1.ai, formerly known as ConductorOne, which the company plans to acquire in October. Every agent had to request temporary data access, and every action was verified at the time of execution, with high-risk requests, such as extracting candidate profiles, being rejected. Workday also showcased its Flex Credit Estimator, allowing companies to forecast the cost of agent operations before making a final decision.

New 'autopilot' agents, such as a recruitment agent that monitors vacancies around the clock, are intended so that the final decision remains with humans. Alet Blanken, Workday Vice President of AI Engineering, stated: 'We decide what the agent can influence and what must remain under human control.' Bhusri added that 'AI should complement people, not replace them.'

Skeptical market

This message may resonate locally. According to the SA AI Index 2026, only 1% of large South African organizations actively use AI agents, and 71% do not plan to do so. Accuracy and reliability were cited as the main obstacles, noted by two-thirds of respondents.

Engagement among Workday's own customers is uneven. Kazmaier reported that over 6,000 customers use Workday agents in production, but in October, Workday informed analysts that only over 850 customers purchased agent AI products in the last 12 months leading up to July, and about 300 subscribed to Flex Credits, consumption-based pricing. These figures measure different things: the first counts all customers using any Workday agent, while the second counts those who additionally paid for its agent AI products, such as HiredScore, Paradox, and Sana.

Competitors are making similar promises. Oracle, which has a cloud region in Johannesburg since 2022, claims that its agent HR applications, launched in April, operate within existing security and approval rules. SAP presented new Joule assistants for SuccessFactors in May under the banner of 'autonomous HCM' (Human Capital Management), although the then-SuccessFactors Product Director Dan Beck wrote that 'people must remain firmly at the center.'

} , {

Similar stories

Startup Founder Predicts AI Will Be Used for Employee Termination Decisions
Read more
www.aajtak.in

Startup Founder Predicts AI Will Be Used for Employee Termination Decisions

Artificial intelligence (AI) is moving beyond routine office tasks and beginning to penetrate the upper management of companies. In the future, AI agents may assist CEOs in developing personnel strategies, providing recommendations for downsizing, and making major business decisions.

Michael Ayof, co-founder and CEO of the American AI startup Arist, believes that such agents will advise executives on all aspects, from reducing headcount to launching new products and implementing corporate changes. He emphasizes that the final decision always remains with company leadership, and AI serves only as a decision support tool.

At the Fortune 500 Innovation Forum held in Detroit, Ayof stated that over the next few years, AI agents will recommend most major transformations within companies. These decisions include the introduction of new products, staff reductions, and other structural changes.

Ayof clarified that while AI agents can offer advice on critical issues, ultimate responsibility for these decisions lies with company management. According to Fortune, Arist has partnered with over 15 percent of companies listed in the Fortune 500.

Arist's AI agents identify organizational problems by gathering employee feedback, analyzing internal company information, and determining trends affecting business efficiency. This technology is not limited to standard surveys or management reviews; the AI system independently selects employees for communication and determines what questions to ask, as well as what follow-up questions to ask based on the answers received.

Employee opinions are then compared with other metrics, such as their work experience, performance records, and engagement levels. By correlating personnel perspectives with internal company data, this platform aims to provide leadership with a clear picture of organizational issues and help identify necessary adjustments.

Arist claims that its system can conduct such a comprehensive investigation in just 72 hours, whereas traditional consulting firms may require months to complete the same task.

The company believes that with AI assistance, enterprises can discover internal problems and make decisions faster without lengthy consulting processes. Ayof is convinced that artificial intelligence can change how companies approach identifying and solving internal problems. He notes that traditional human-led checks have limitations, especially regarding carefully listening to employees and objectively understanding their concerns.

Ayof also questioned the effectiveness of traditional consulting. According to him, companies spend significant time and resources checking organizational problems, but this does not guarantee the right outcome. In an interview with Fortune, Ayof suggested that consultants are wrong in diagnosing the problem about 90 percent of the time, although this is his personal assessment and not independently verified industry statistics.

Ayof's main goal is to help large corporations operate as quickly as small businesses. His argument is that timely access to reliable information allows companies with tens of thousands of employees to respond to challenges more rapidly.

The growing use of AI agents in business opens a new discussion about the role of artificial intelligence in corporate decision-making. Beyond performing daily tasks and data analysis, these systems are now being tested to assist in studying complex business situations and strategic planning.

This shift could significantly impact human resources management. If companies use AI-derived information to assess employee contributions, identify unnecessary tasks, or determine the need for departmental changes, AI could play a key role in employment-related decisions.

However, concerns also arise regarding accountability, transparency, and potential algorithmic bias. It is important to remember that an employee's contribution cannot be measured solely by numbers, and AI assessments may overlook an employee's personal circumstances or the scale of the consequences of personnel decisions. Despite this, Ayof insists that ultimate authority remains with company leadership, yet the growing potential of AI agents could transform management methods and strategic planning in corporations. The challenge for companies will be to avoid a trade-off between the speed and thoroughness of data-driven decisions and human understanding, fairness, and accountability.

Cognizant to hire 20% more freshers instead of layoffs; company's view on the impact of AI
Read more
www.aajtak.in

Cognizant to hire 20% more freshers instead of layoffs; company's view on the impact of AI

Despite the growing impact of Artificial Intelligence (AI) and concerns about its effect on jobs, Cognizant has announced that it will increase the number of freshers by 20% instead of reducing hiring. The company believes that AI is not eliminating entry-level jobs but rather changing their nature. This strategy indicates that in an environment of declining campus recruitment, the importance of new roles, AI training, and human decision-making skills is increasing.

AI has started handling many traditional tasks and roles globally, and 37% of entry-level tasks in large Indian companies are currently being performed by AI. Amid concerns about the potential negative impact on jobs, Cognizant has decided to focus on freshers.

Rajesh Warrior, Chairman and Managing Director of Cognizant India and Global Operations Head, stated that AI is not ending entry-level jobs, but they are merely transforming. The company continues to recruit freshers while also providing AI training to its existing employees and creating specialized AI roles.

In an interview with Business Today, Warrior referred to the June 2026 report by Cognizant and Pershing, 'The AI Workforce Pulse: The Adaptability Imperative.' According to this report, Indian organizations estimate that 37% of entry-level tasks have already been handled by AI, which is higher than the global average of 33%. In the survey, 80% of participants also stated that AI provides employees with the opportunity to focus on more critical tasks.

The report also indicated another shift in entry-level jobs. 94% of HR heads included in the survey expect AI to create new roles that do not currently exist.

According to NASSCOM data, the Indian IT sector added approximately 135,000 people in FY26, compared to around 133,000 in FY25. A large portion of these new jobs came from global delivery centers. However, there has been a sharp decline in campus recruitment, making it difficult for students to secure entry-level positions. Rajesh Nambiar, Chairman of NASSCOM, said in a recent discussion with Nikkei Asia that 'from a fresher's perspective, this is a matter of concern, and I will not sugarcoat it in any other way.'

When asked about limited job opportunities, Warrior acknowledged that the change is real, but it should not be seen as the end of opportunities. He clarified that 'AI is automating some routine and repetitive tasks that were traditionally part of entry-level jobs. However, we see this as a transformation of roles rather than the disappearance of entry-level opportunities.'

According to Warrior, changes are also visible in daily tasks. He mentioned that 'roles focused primarily on task completion are now shifting towards roles that involve checking AI outputs, understanding their meaning, handling exceptions, and using human judgment.'

He further stated that young professionals may now spend less time on routine tasks and need to focus more on managing AI-driven workflows. This includes interpreting information, solving business problems, and identifying situations where human judgment is necessary. This suggests that in the age of AI, basic coding and software testing alone may not be sufficient to secure a job.

Warrior expressed confidence in India's digitally adept workforce, but also noted that the shortage of AI skills remains a challenge. This could pose a major hurdle in preparing the workforce for the changing job market. He said, 'The question is not whether India has the talent to succeed. The question is whether we can develop the capacity for AI at the speed and scale required across various industries.'

He emphasized that employers require individuals who possess digital acumen along with critical thinking, communication skills, creativity, and industry knowledge. Cognizant's previous report, 'New Work, New World 2026,' mentioned that 93% of jobs could be affected by AI in some way. Warrior concluded that 'AI can generate results, but people must provide context, ask the right questions, verify the results, and take responsibility for decisions.'

Popular