There is no shortage of enthusiasm about what AI can do. There is considerably less clarity about what organizations need to do differently because of it.
That gap matters. Buying an AI tool is a purchasing decision. Helping people work effectively alongside it requires decisions about skills, job design, development, accountability and performance. Those decisions reach well beyond the technology team.
Cornerstone’s latest analyst town hall put that challenge at the center of its Workforce AI strategy. The team outlined a broader approach to workforce readiness, including Cornerstone Workforce Planning and Cornerstone Agent Ready, which addresses agent conduct assessment.
The interesting part is how these pieces connect. Understanding the workforce, deciding how work should change, developing people and governing AI behavior are becoming parts of the same business problem. Learning and talent leaders have an opportunity to help solve it.
Start With the Work People Actually Do
Not only is AI changing work, but it is also doing it at the task level. That should influence how organizations approach almost everything that follows.
A job title tells us relatively little about which activities AI can support, which responsibilities require human judgment or what someone will need to learn next. Even people with identical titles may spend their time doing very different things.
Consider a customer service role. AI might help retrieve information, summarize an interaction or recommend a response. The employee still needs to recognize when the situation requires empathy, an exception or escalation. As responsibilities shift, the development priorities shift with them.
This is where learning can become more strategic. Start by understanding the tasks that create value, how those tasks are changing and the capabilities people need to perform them. Then build development around those requirements.
Otherwise, organizations risk producing a ton of AI training with no guidance as to how the employee’s own work is supposed to change.
For employees, this also needs to be understandable. If a role is changing, people should be able to see what is expected, which capabilities they already have and how to develop the rest. A sophisticated workforce model has limited value if the person affected cannot clearly understand the next step.
Workforce Planning Has to Keep Moving
Cornerstone’s workforce planning proposition addresses a familiar disconnect. Business conditions change continuously, while workforce plans often move through an annual budgeting cycle.
A plan can be carefully constructed and still lose relevance when demand shifts, an acquisition changes priorities or automation alters the work. Headcount and cost remain essential inputs, but they cannot explain the full capability picture.
Cornerstone describes their Workforce Planning solution as connecting workforce and skills intelligence with scenario modeling and execution. Its proposed process moves from understanding current and future needs to evaluating options and assigning actions.
Those options include building capabilities internally, hiring, borrowing talent and using automation: Build, Buy, Borrow, Bot. The value comes from examining the tradeoffs. How quickly is the capability needed? What will each option cost? What knowledge should remain inside the organization? What would the change mean for employees?
For learning leaders, this creates a clearer role in business planning. A reskilling recommendation needs to include a credible timeline, evidence of readiness and a connection to the work that needs doing. That gives the business something useful to compare with hiring or automation.
Insight Needs an Owner and a Deadline
A skills gap on a dashboard is simply a news report. Closing it requires someone to make a decision and follow through.
Cornerstone’s briefing emphasized this connection between planning and action, including commitments with owners and dates, delivery through workplace tools such as Slack and Teams, and monitoring that feeds results back into the plan.
That feedback matters. Suppose an organization chooses to develop internal talent for an emerging role. Assigning learning is only the beginning. Leaders need to know whether employees can demonstrate the required capabilities, whether managers provide opportunities to apply them and whether the business can fill the roles when needed.
If the answer is no, the plan needs to change. Perhaps the development takes longer than expected. Perhaps employees need practice, coaching or access to different assignments.
Managers will be central to making that happen. They translate plans into assignments, create room for practice and observe performance in context. Any readiness strategy should make those responsibilities clear and give managers the time and support to carry them out consistently across teams.
This is the practical meaning of continuous readiness: a repeated cycle of decisions, action and adjustment, with enough evidence to understand what is working.
Your AI Agents Need Behavioral Assessment, Too
The Agent Ready proposition raises an issue that deserves more attention: an agent can retrieve the correct information and still handle a conversation poorly.
An employee asking about leave, confidentiality or a sensitive workplace situation needs more than a fluent answer. The agent needs to stay within organizational policy, recognize its limits and respond appropriately when someone pressures it to make an exception.
The briefing illustrated this through conversations in which a user challenges an answer or claims a manager has already approved something. These are useful scenarios because workplace conversations rarely follow a perfect script.
Cornerstone described Agent Ready as assessing agent responses against its conduct curriculum and an organization’s selected policies. The proposed outputs include conversation records, expected behaviors, assessment verdicts and remediation guidance, with reassessment when policies or models change.
That is a meaningful extension of the readiness discussion. Organizations have long expected employees to demonstrate appropriate conduct. As agents take on more interactions, leaders need evidence of how those systems behave under realistic conditions.
An assessment record can support that work. Leaders still need to decide who reviews failures, who approves fixes and when an agent should hand the conversation to a person.
Learning and Talent Bring Essential Context
It would be easy to treat agent assessment as something for IT to manage. Technology teams clearly have an essential role, but the behavioral expectations require broader expertise.
Learning professionals understand scenario design, practice, feedback and assessment. HR teams understand employee processes and organizational policies. Business leaders understand the consequences when a conversation or decision goes wrong.
Bringing those perspectives together can improve both employee development and agent evaluation. A difficult customer interaction, for example, can reveal what a person needs to practice and what an agent needs to escalate.
The briefing’s broader Workforce AI architecture connects learning and talent development with skills intelligence, workforce context and agent experiences. Its significance will depend on whether those connections help organizations make better decisions and support people more effectively.
More recommendations alone will not establish that value. The useful question is whether the system helps someone become ready for a responsibility the business actually needs them to perform.
Competitive Landscape
Cornerstone is part of a broader shift. Workday, SAP SuccessFactors, Oracle and Degreed compete with different parts of its learning and talent portfolio, and each is connecting AI with workforce capability. The overlap is useful context, even though their approaches and product scope differ.
Workday connects skills intelligence with development and workforce decisions through Skills Cloud. It identifies skill gaps and supports personalized learning and internal mobility. Its Agent System of Record adds a related governance layer for registering, configuring and monitoring AI agents, including access controls and audit records. The parallel is a growing need to understand what people can do and oversee the agents working alongside them.
SAP SuccessFactors is also bringing planning, development and agentic AI closer together. SAP has announced workforce planning capabilities and a Workforce Upskilling Assistant that connect to this theme. The planning direction includes scenarios involving people and agents, while the assistant is designed to identify skill gaps and deliver personalized learning through the tools employees already use. As with Cornerstone, buyers need to establish which announced capabilities are available in their environment.
Oracle uses Dynamic Skills to connect a shared skills foundation with hiring, learning, career mobility and planning. It describes learning recommendations that adapt as roles and business needs change, alongside visibility into role requirements and employee skills. That addresses the same disconnect discussed here: knowing the workforce is useful only if that knowledge informs how people develop and where they can contribute.
Degreed approaches the issue through learning and skills development. Degreed Maestro combines skills and learning history with personalized pathways, AI coaching, roleplay and skill reviews. Its relevance is the emphasis on practice, feedback and evidence of proficiency. That fits the blog’s argument that readiness requires more than assigning content or counting completions.
These examples suggest that the competitive question is becoming how effectively each provider connects insight to action. They do not establish equivalent agent conduct assessment capabilities. Agent registration and monitoring, employee skill assessment, and testing an agent against organizational policies serve different purposes. Leaders should ask vendors to demonstrate each capability they need and explain how the evidence feeds back into workforce decisions.
What Leaders Should Ask Next
Cornerstone organized its approach around four steps: Know, Plan, Act and Govern. That provides a useful starting point for evaluating the strategy, provided each step connects to evidence.
Before expanding an AI initiative, leaders should ask:
- Which tasks are changing, and what capabilities will people need as a result?
- How current and reliable is our understanding of workforce skills?
- What would make development a better option than hiring, borrowing or automation?
- Who owns each action, and when should we expect results?
- How will we assess agent behavior and respond when it falls short?
These questions also provide a practical basis for a pilot. Choose a defined business need, establish the current performance level and agree on what improvement would look like. Then evaluate readiness, execution and outcomes together.
Cornerstone’s presentation outlines product direction, so implementation, availability and demonstrated results still need evaluation. Buyers should establish which capabilities they can use today, how those capabilities connect to existing systems and what effort is required to keep workforce information current.
The broader direction deserves attention. AI adoption is creating decisions that cross traditional boundaries between workforce planning, learning, talent management and governance. Treating those decisions separately makes it harder to understand their combined effect on performance.
For learning and talent leaders, the opportunity is to bring those conversations together around the work. Understand what needs to happen, help people build the capability to do it and establish how both people and agents will be held accountable.
That is where workforce readiness becomes a business discipline. The next step is to choose a business problem and start building it.
