What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Workday UK and Ireland general manager Dan Pell argues that organisations will need a system of record for AI agents as those agents gain the ability to act autonomously. In a Computer Weekly interview published 18 June 2025, he described Workday’s plan to give agents identities, roles, skills and permissions, drawing on the company’s experience managing employee information. The interview presents a strategy, not evidence that a finished, generally available product has been independently assessed.
Why Pell says AI agents need an identity record
An AI agent that can take actions inside an organisation is more than a tool that generates text. It may need access to company information or workflows, so an organisation must be able to define what the agent is, what it can do and how its activity is governed.
Pell frames agents as non-human identities that can act autonomously, making identity and access management a central concern. He says Workday is building an “agent system of record” and imagines organisations managing, controlling and understanding agents through a Workday instance. These are Workday’s stated intentions in the interview, not a universal technical requirement or a confirmed product specification. Computer Weekly’s interview with Dan Pell was published on 18 June 2025.
What an agent system of record could contain
Pell compares the proposed agent record with employee administration. In his account, an agent profile could capture:
#1 Best Overall
- Identity and role: what the agent is and the function it serves.
- Hierarchy and permissions: where it sits in an organisational structure and which actions or resources it is permitted to access.
- Skills and context: what the agent can do and the organisational information relevant to its work.
- Activity or behaviour: what the agent is doing or how it behaves.
The interview does not set out an access-control architecture, implementation standard, audit mechanism or technical specification. It also supplies no customer case study, availability confirmation, pricing or independent test of the proposed system. A record of identity and permissions is Pell’s proposed approach; the interview does not show how Workday would implement or validate it.
How Pell illustrates agents moving from text to action
Pell uses job-description creation as an example of how an agent might go beyond drafting text. With budget and headcount context, an agent could help identify internal candidates and potentially message people. In his illustration, organisational knowledge and thresholds would constrain the agent’s actions.
Rank #2
That example describes a possible workflow, not a verified Workday deployment. It also highlights why governance matters: generating a draft and taking an action that affects employees are different levels of responsibility. The interview does not explain what approval gates, monitoring or limits such a workflow would use.
Why Pell says people still need to apply judgment
Pell presents generative AI as a starting point rather than a substitute for human judgment: “I think AI, particularly generative AI, gives you a good starting point. But it’s definitely not the final outcome.” He argues that people must read, process and apply judgment to AI-generated work.
He also forecasts that leaders will manage agents alongside people and discusses using agents to augment workforce skills, including communication and problem solving. “We are the last generation of leaders who will only manage physical people,” he says. That is Pell’s prediction, not a measured finding about employment, productivity or adoption; the interview reports no statistics on those outcomes.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What the interview does—and does not—establish
The Computer Weekly interview records Workday’s 2025 vision for managing AI agents. It does not establish whether the proposed system was later released, what features it has today, whether organisations have deployed it, or whether a possible collaboration with Salesforce became a formal partnership. Pell mentions that Workday might collaborate with Salesforce and manage different agents on its platform, but the interview describes a possibility rather than an announcement.
For an organisation evaluating agent governance, the issues raised by Pell’s proposal are useful questions, not proof that any particular system answers them:
Quick Recap
- How are agent identities, roles and permissions assigned and reviewed?
- Can an organisation see agents’ actions and behaviour?
- Which actions require human approval, and how is that approval enforced?
- How does organisational context constrain an agent’s work?
- How do agent records relate to existing workforce systems?
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
