Before your board asks: five Workday AI questions worth preparing for.

For most boards, the central question is not whether Workday AI is capable. It is whether the organisation is positioned to use it well, and whether the people in the room have done the thinking required to make a sound decision rather than an expensive one.

This article was accurate at the time of publication in June 2026 and is reviewed regularly to reflect changes to the Workday platform and commercial model.

01

THE CEO
Is this a real business decision, or are we being carried along by a vendor cycle?

Workday’s AI capability is being marketed heavily, and organisations that activate without the right foundations tend to end up with an additional layer of tooling, further complicating their system architecture, rather than an improved way of operating.

The capability has the potential to shift HR and finance systems from record-keeping platforms into governed execution platforms. That is a meaningful operational change. But it only materialises if activation decisions are tied to genuine process redesign.
 
The capability does not, on its own, change how work gets done. The decisions made around process make the change happen. The signal that this is a real business decision, rather than a vendor-driven one, is straightforward: KPIs tied to business outcomes, not usage metrics.
 
Adoption rates and active users are not the answer the CEO is looking for. Cost per resolved workflow, hours released in operations, and faster close cycles are.

02

THE CFO 
What have I already paid for, and what is this going to cost me on top?

This is the question most organisations cannot answer cleanly.

Workday’s commercial model is shifting in ways that are material to this conversation. Flex Credits are now the mechanism through which AI consumption is metered, covering eligible agents and platform innovations. The economic unit is work completed rather than user headcount, which sounds straightforward until you start modelling realistic consumption across a workforce.
 
What your organisation has already paid for depends on your subscription, contracted features, renewal timeline, and, in the future, it will be on whether you have signed the UMSA, which governs access to AI capabilities within the platform.
 
The real commercial risk is not the baseline cost. It is bill shock from unexpected usage, vague adoption plans, low-value usage, and continuing to fund the manual processes and legacy tools that AI was supposed to replace, while also paying for the new capability.
 
AI return is most commonly undermined not by any single line item but by the failure to retire what came before it or by optimising processes.
A board-level cost model for Workday AI should cover six areas: the Workday commercial package and Flex Credits; implementation and partner services; identity, security, and governance work; integration and data engineering; change management and training; and any temporary overlap with existing tools that AI is intended to replace.

The CFO should ask for a modelled Flex Credit consumption forecast before any commitment is made, with clear visibility on top-up mechanics and which capabilities are generally available versus early access. Without that, the commercial conversation is happening in the dark.

This is an area where independent advice, before the commercial conversation with Workday, tends to pay for itself. Understanding your entitlement position and modelling realistic consumption scenarios gives you a materially stronger starting point than going into that conversation without it.

03

THE CIO
Is Workday still the right platform, and what are we committing to if we go further with it?

The most important strategic insight for boards is that this is not a choice between Workday AI and general-purpose platforms like Gemini or Copilot. Workday is explicitly building a hybrid model.
 
Employees continue working inside Microsoft 365 or Google Workspace, but when an action touches HR or finance, such as a leave request, a payroll query, or a budget approval, it executes through Workday’s rules, permissions, and transaction layer. The experience lives in the tools people already use. The governance, the approvals, and the business logic stay in Workday.
 
The board question is therefore not which platform to back. It is where the organisation wants its business rules and governed execution to live. For HR and finance, the answer is usually Workday, and the hybrid model means that does not require abandoning anything else. Workday’s AI capability is strongest where the task is transaction-heavy, permission-sensitive, and tied to HR or finance workflows where the audit trail matters.
 
The real risk the CIO should name at board level is not classic vendor lock-in. It is fragmentation. Organisations that activate multiple AI tools without a clear architecture end up with employees managing disconnected systems, copying information between applications, and spending time on AI coordination rather than the work AI was supposed to simplify.
 
The answer is to designate Workday as the control plane for HR and finance actions, with clear boundaries around what executes there versus what sits in the broader productivity layer.
 
On the commitment question: once agents are embedded in processes and consumption is metered through Flex Credits, the cost of changing direction increases. That is true of any deep platform dependency and is worth naming rather than leaving unspoken.
 
The response is not to avoid activation. It is to go in with clear commercial terms, defined decision gates, and a realistic view of which capabilities are generally available now versus which remain in early access.

04

THE CTO
How does this fit what we are already building?

The CTO is the person who has to live with this decision after the board meeting ends. Their question is about architecture, interoperability, and what happens when the organisation’s broader technology strategy meets a platform-native capability that operates on its own terms.
 
Workday AI runs within the Workday environment. That is its strength. Governed access to structured HR and finance data in a controlled, auditable setting, and it defines its boundaries. The CTO needs a clear picture of where Workday AI sits within the broader AI estate.
 
Workday’s direction on interoperability is increasingly coherent. Its Agent System of Record and Agent Gateway are designed to register and monitor third-party agents operating within or alongside the platform.
 
The principle is that Workday retains ownership of policies, approvals, permissions, and transaction execution, while other tools handle the experience layer. The board should approve an architecture only when identity, permissions, audit logging, and human approval boundaries are explicit and documented.
 
The questions worth having clear answers to before this conversation:
  • What data leaves the Workday environment, and under what conditions
  • How do Workday agents interact with processes that span multiple systems?
  • What does the integration model look like for use cases that are not fully contained within Workday?
  • And if the organisation later wants to extend or replace elements of this, what does that involve?

These are not objections to Workday AI. They are the right questions for anyone responsible for a coherent technology strategy. The CTO who comes into this conversation with clear answers to them is a different proposition from the one who comes with reassurances.

05

THE CHRO
What do I tell my people, and am I about to be handed a headcount case I did not sign up for?

AI agents will reduce the time certain tasks take. In some organisations, that will mean headcount reduction. In others, it will mean the same headcount handling higher volume, or redeployment into areas the business has been under-resourced in.
 
The outcome is not determined by the technology. It is determined by what leadership chooses to do with the capacity that is released, and whether that decision has been made consciously before activation, rather than reactively after it.
 
That means understanding, before activation, which roles carry significant concentrations of the tasks agents are most likely to handle first. It means having a clear organisational position on what happens to that capacity. And it means leading the workforce conversation before the technology forces it, rather than managing the fallout after a go-live date has been set.
 
Workday positions its agents as augmentation rather than replacement, and customer evidence supports that framing in organisations that have approached activation with genuine process redesign and change management behind it. The organisations where that framing does not hold are the ones that activated without either.
 
The answer the CHRO needs to bring into this room is not a guarantee about headcount. It is a clear position on workforce impact, a change programme that precedes activation rather than follows it, and the confidence to lead that conversation rather than wait for it to arrive.

The big question behind all five: are we governing this properly?

Each of the five questions above has a governance dimension that the board should address explicitly rather than assuming it is covered.

The most important distinction is between AI use cases that support information and workflow, and AI use cases that affect employment decisions. Absence handling, policy queries, and finance process automation sit in a different risk tier than recruiting, performance evaluation, promotion, and compensation.

The latter category carries legal exposure under existing privacy, discrimination, and employment law that does not wait for future regulation to apply. The board should require separate approval paths for these categories, not a single AI programme that treats them the same.

The regulatory picture varies by geography, but the direction is consistent. Whether the relevant framework is the EU AI Act’s high-risk employment timeline, ICO guidance on fairness and transparency, or US anti-discrimination law applied to AI-assisted decisions, the underlying obligation is the same: document, test, and keep humans in the loop where people are materially affected. None of this is a reason to delay activation. It is a reason to build the governance structure before the capability is live.

In practice that means:

  • Impact assessments for employment-affecting use cases
  • Defined human review requirements for high-risk decisions
  • A clear decommissioning plan for the manual processes and tools AI is replacing.

That last point connects directly back to the CFO question. Governance and commercial discipline are the same conversation.

The questions in this piece are answerable. What varies is how much groundwork has been done before the meeting.

We help Workday customers do that groundwork across entitlement, configuration, commercial positioning, and workforce impact, so the board conversation is built on substance rather than assumption. If you would like an independent view of where you stand, get in touch with the Preos team.

Get in touch

Further reading

GLOSSARY

The Workday AI glossary: what the terms you hear mean

Sana, Flex Credits, service agents, uMSA and more: what each term means and why it matters for your organisation.

8 min read     Read resource 

ARTICLE

What is Workday AI?

Workday has used ML for years. The latest capability is a meaningful shift. A plain explanation of Workday AI, how agents work, and what is available right now.

4 min read     Read resource 

ARTICLE

What is Sana, and what does it mean for Workday customers?

What Workday’s acquisition of Sana means in practice: what changed, how the three tiers work, and what to do before your next renewal.

5 min read     Read resource