Workday has agreed to buy C1.ai, the identity governance company formerly known as ConductorOne, in a deal announced on October 13, 2026 at Workday Rising. Workday's newsroom announcement describes C1.ai as an agentic identity platform for people and agents. The price was not disclosed, and the deal is expected to close in Workday's fiscal fourth quarter, which ends January 31, 2027, subject to customary conditions, according to coverage of the announcement.
The short version: the system that already decides what a new hire can touch on day one is trying to decide what that hire's AI agents can touch too. That is a sensible idea and an early one. This post separates what was announced from what is still a roadmap, and lists what C1 customers should ask before the deal closes.
TL;DR: the deal in questions
| Question | Answer |
|---|---|
| Who is buying whom? | Workday (Nasdaq: WDAY) is acquiring C1.ai, formerly ConductorOne |
| When was it announced? | October 13, 2026, at Workday Rising |
| How much? | Not disclosed |
| When does it close? | Expected in Workday's fiscal Q4, ending January 31, 2027, subject to conditions |
| What does C1 do? | Identity governance: who has access to what, for people and AI agents |
| Is the HR-triggered agent access live? | No. It is the stated aim, not a shipped feature |
| Does this affect C1 customers who use another HR system? | Not addressed in the announcement |
What C1.ai actually is
C1 is a Portland-based identity governance and administration (IGA) company. It renamed itself from ConductorOne to C1 on April 7, 2026, and one analysis puts its last raise at $79 million in fall 2025. Portland Business Journal's report, relayed by Techmeme, put total funding at $111 million and noted that the company's headquarters has moved between San Francisco and Portland.
IGA software answers a boring but critical question: who can access which system, who approved it, and when does it get removed. According to an analysis of the deal, C1 ships 400+ connectors, including Okta, Entra ID, Workday, Oracle Fusion HCM and SAP SuccessFactors. Those connectors matter because they are what let a governance tool actually grant and revoke access across a messy real-world stack.
Identity verification ID card check representing AI agent identity credentials in Workday C1.ai agent governance
Why Workday wants this: HR as the source of truth
Workday's pitch, as reported, is that HR events should drive access. When someone is hired, changes role or leaves, Workday already records it. The plan is for those same events to grant and remove access for employees and for the AI agents tied to them.
Agents have no manager to notice that they are still running. A contractor whose badge is revoked is obvious; an agent with a long-lived API token and a forgotten schedule is not. Tying an agent's lifecycle to a human or team record in a system of record is the same fix the industry applied to service accounts a decade ago, now applied to a population that is growing much faster.
This connects to Workday's earlier move. The company introduced its Agent System of Record in February 2025 to register and manage agents, and C1 would supply the enforcement layer: the part that actually opens and closes doors in other systems. If you want the background on why this layer is suddenly in demand, our posts on know-your-agent identity and how AWS frames agent autonomy and scope cover the same problem from other angles.
What is announced versus what is roadmap
Read the announcement carefully and three things stand out.
1. The HR-triggered agent flow is a plan. The analysis cited above notes that C1's Workday connector documentation does not list provisioning as a capability today. C1 reads Workday data and governs access elsewhere. The two-way flow, where a Workday event automatically provisions or de-provisions an agent, is what the combined company says it will build.
2. The headline numbers come from interested parties. The analysis flags a 95% figure for companies running autonomous agents as coming from C1's own research, and a forecast of 50 to 100 agents per human as coming from Workday's product president. Both are sales-oriented claims, useful as direction and not as measurement. Treat them the way we treated vendor figures in our AI verification tax roundup, which also quoted Workday survey data: as the vendor's own number.
3. Customer terms are unaddressed. Neither the announcement nor the coverage we found says what happens to C1 customers whose HR source is Oracle or SAP, whether C1 stays available as a standalone product, or whether the current pricing meter survives. C1 reportedly bills by managed identity and by usage, and agents are exactly the identities most likely to multiply under that meter.
Precedent: when an independent identity vendor gets absorbed
The same analysis points to ServiceNow's purchase of Veza as a pattern: the announcement centered on the acquirer's platform, and customer terms were left for renewal conversations. Governance tools are valued partly for neutrality, because the auditor should not be owned by the thing it audits. Neutrality is harder to keep once the tool sits inside a larger platform.
That does not make the Workday deal bad. It means buyers who chose C1 for being independent of their HR vendor have a legitimate reason to ask what independence looks like after January.
The agent identity market is getting crowded
C1 is not alone in chasing this layer. The past month alone brought several adjacent pieces:
- Gumloop's AgentMail and AgentID gave agents their own sign-in identities.
- Amex's intent framework tackled which purchases an agent is authorized to make.
- Shopify's agent checkout terms set budgets for agent buyers.
- Our overview of AI agent security platforms and the MDM-style agent management guide map the vendors and approaches.
Two separate keyhole gates a green key must pass through, illustrating zero-trust access for AI agents in agent identity governance
The common thread: once agents act on their own, every question that identity teams ask about people (who approved this, what can it reach, when does it expire) needs an answer for agents. Acquisition is one way for a platform vendor to get there fast.
What this means for what you build or pay
If you run AI agents inside a company, the practical lessons do not depend on whether the deal closes:
- Give every agent an owner of record. A named human or team, stored somewhere that changes when the person leaves.
- Issue short-lived credentials. An agent token that expires in hours fails safe when nobody remembers it.
- Review agent access on the same cadence as human access. Access reviews that skip service and agent accounts leave the largest blind spot.
- Log what an agent did, not just that it signed in. Identity tells you who held the key; runtime controls tell you what they did with it. To block risky agent actions before they happen rather than review them afterward, see AgentBeam.
What C1 customers should do now
The advice that holds up from the analysis, adapted:
- Now: review your order form for renewal date, price-increase caps, termination rights and change-of-control language. Note which HR system your C1 tenant reads from.
- Before close: ask in writing for commitments on standalone availability, connector support and pricing for at least one renewal term. Export access-review and approval logs in a format you control.
- Before renewal: price an alternative so you have leverage. The analysis names Okta, Saviynt and ServiceNow as possible comparisons.
What we still do not know
- The purchase price and whether any of it is paid in Workday stock.
- Whether C1 will keep its current name and brand after close.
- Whether Oracle and SAP HR sources stay first-class connectors.
- When, or whether, the HR-to-agent provisioning flow ships, and in what form.
- How regulators will view the deal; no review was mentioned in the coverage we read.
We will update this post when Workday publishes closing details or customer terms.
Why it matters beyond Workday
The most useful reading of this deal is as a signal about where enterprise AI spending is going. A year ago the budget conversation was about models and copilots. Now a large HR and finance vendor is buying a company because it believes the number of non-human identities will dwarf the number of employees, and that someone must be accountable for each of them. Whether the specific bet pays off, the problem is real: agents already hold credentials, call APIs and, as our coverage of CERT-In's advisory on agents accessing live systems showed, sometimes touch systems their owners never intended.
Related reading
- Know-your-agent identity (KYA) explained
- AWS trust in depth: agent autonomy, scope, and policy
- AI agent security platforms compared
- Managing AI agents like devices: the MDM approach
- Gumloop AgentID and agent sign-in
- The AI verification tax
- CERT-In advisory on AI agents in live systems
Details reflect public announcements and reporting as of October 17, 2026. The acquisition has not closed and terms may change.