Enterprise AI adoption needs a named executive who can fund the work, settle cross-team priorities and answer for the program's results. That executive might be the CIO, COO, CHRO or a business leader with a wider mandate. Choose by actual authority and the work being changed, rather than by title alone.
Give day-to-day coordination to an adoption lead with time to do it. Keep responsibility for each workflow's quality and business results with the leader who can change that workflow. In a smaller company, one person may fill more than one role, but each responsibility still needs an explicit home.
The difficult part is deciding what “owns AI” means. Buying software, helping employees learn, accepting risks and delivering better customer service involve different decisions. Assigning all of them to one enthusiastic person does not give that person the authority to make them.
Start with the decisions the owner must make
Before choosing an executive, identify the decisions that keep coming back unresolved. For an enterprise program, the sponsor should be able to act on three kinds of problem:
- Funding. Commit money and staff time to implementation, practice, support and ongoing maintenance.
- Cross-team priorities. Resolve a disagreement when one department must change its work for another to benefit.
- Results. Require workflow owners to show whether completed work has improved, then expand, repair or stop the investment.
These are practical selection criteria, not a claim that one reporting line guarantees success. A senior sponsor can delegate decisions within clear limits. They still need a route to resolve matters beyond those limits.
McKinsey's March 2025 report, based on a July 2024 survey of 1,491 respondents, found an association between CEO oversight of AI governance and higher self-reported financial impact from generative AI. It also found that governance was often shared and AI adoption resources were commonly partly centralized. The survey does not establish that giving the CEO a new responsibility causes a return. Read the findings and methodology.
Governance covers how AI is used responsibly. Adoption also requires changes to daily work, employee support and decisions about the benefits. A risk committee may be essential without owning all of that work.
Compare the mandate behind each title
Use the company's existing responsibilities as the starting point. The following comparison is a way to choose a workable arrangement, rather than a ranking of executives.
| Possible owner | When the mandate fits | What must be covered elsewhere |
|---|---|---|
CIO or technology leader | The role already leads business change as well as platforms, data and integration. | Business leaders must own workflow changes and benefits; HR must help with learning and workforce concerns. |
COO or operations leader | The priority is improving work across service, fulfillment or shared operations. | Technical, security and data decisions need capable owners and funded support. |
CHRO or people leader | The program primarily concerns workforce capability and changes to roles and working practices. | Operational leaders must commit to changing work and measuring results; IT retains technical responsibilities. |
Transformation office | It already has an executive mandate to coordinate funding, dependencies and business change. | A reporting office needs decision authority and permanent business owners for what it hands over. |
Business-unit leader | The scope is a defined business with its own outcomes and budget. | Shared platforms, enterprise standards and cross-unit conflicts need a central decision route. |
An IT-led program can be a sensible choice. In its April 2026 account of internal agent adoption, Microsoft describes its IT organization taking a leadership role through a cross-disciplinary initiative with separately accountable workstreams. Its guidance combines a dedicated adoption lead, a cross-functional team, executive sponsors and employee champions. That is a large technology vendor's own experience, not a staffing prescription for every company. Microsoft Digital's adoption account.
A CHRO-led skills program can also be appropriate within its scope. But if the intended result is faster order fulfillment, someone with authority over sales and warehouse operations must own those changes. Training completion cannot settle a disagreement about who checks an order.
For a company-wide program spanning several functions, give the sponsor a clear connection to the CEO. The CEO may sponsor it directly or authorize another executive to do so. The test is whether unresolved decisions get made, including when they cross that executive's usual boundaries.
Separate coordination from responsibility for results
In a September 10 LinkedIn post, Alex Lieberman, cofounder of enterprise AI partner Tenex, described discussing ownership with a strategy executive at a 5,000-person company. He argued for CEO sponsorship and a dedicated Head of AI with access across functions. That account describes a conversation and his advice, not a verified implementation outcome. Lieberman's post and discussion.
A reply from self-described chief AI officer Gregg Bayes-Brown adds a useful qualification. His strengths in strategy, communication and the human side of adoption still leave him needing engineering, operational and company-specific expertise. Appointing a capable coordinator does not make those other responsibilities disappear.
Keep the division of work visible:
- The sponsor decides enterprise priorities, resources and escalations within their mandate.
- The adoption lead coordinates the program, gathers evidence, arranges support and brings unresolved decisions to the right owner.
- The workflow owner changes the process, assigns checking and support duties, and answers for the quality and business result.
- Technical and risk owners maintain the relevant systems, data controls, review requirements and routes for handling problems.
- Managers and people teams make practice possible and address the effects on employees' roles and workload.
Use the AI adoption leader's charter to record the lead's delegated decisions, resource commitments and escalation limits.
These are responsibilities, not five mandatory new jobs. Name the people doing them and make sure they have capacity. A committee can bring them together, but it should leave each decision with someone authorized to act.
The adoption lead can report that employees are saving time. The workflow owner needs to establish what happens to that capacity. Use the AI value measurement guide to distinguish use, task improvement and business outcomes. Use the people and change guide when the unresolved issue is employees' working conditions or support.
Test the arrangement on one difficult handoff
Suppose a building-products distributor uses AI to prepare customer quotations from an approved price list and stock records. The sales lead wants to shorten the time from a customer's request to an accurate quote. Warehouse staff need to confirm availability, and finance must approve discounts outside the sales team's limit.
The adoption lead can arrange a trial and collect feedback. They cannot, simply by holding that title, change finance's approval limits or commit warehouse staff to a new checking routine.
In this example, the sales lead owns the quality and turnaround of the quotation process. The warehouse manager owns how stock confirmation is performed; finance owns discount policy. IT handles the approved connection to company records. A COO with the appropriate mandate can resolve competing demands on staff time and fund the agreed process. The adoption lead coordinates that decision and checks that the resulting support is in place.
If the COO lacks authority over one of those functions, the decision needs the CEO or another authorized executive. The same reasoning applies to a CIO-led program. An ownership label is useful only when the people involved know which decisions it permits.
Once that arrangement works, place it inside the broader enterprise AI adoption strategy. Revisit it when a pilot expands into another business unit or becomes an ongoing service. The person coordinating the experiment may not be the permanent owner of the resulting process.
Questions about AI adoption ownership
Should IT or HR own AI adoption?
Either can lead a suitably scoped program. IT is a plausible home when its mandate includes business change alongside systems and data. HR can lead workforce capability and changes to working practices. Neither department can deliver another function's business outcome without that function's participation and authority.
Choose the sponsor against the decisions the program must make, then name workflow owners and technical, risk and people responsibilities. The comparison of ownership mandates shows what each arrangement needs from the rest of the company.
Does the CEO need to run the AI adoption program?
The CEO does not need to coordinate every pilot, training session or technical decision. A company-wide program does need executive sponsorship with enough authority to fund work and settle cross-functional disputes. The CEO can provide that sponsorship directly or give another executive an explicit mandate and escalation route.
Keep daily coordination with a designated lead and business outcomes with workflow owners. Test the arrangement by taking the proposed sponsor one unresolved cross-team decision, rather than relying on an announcement of support.
Can AI adoption ownership be shared?
The work has to be shared, but responsibility for a particular decision should be clear. An adoption lead can coordinate a quotation trial while a sales leader owns quote quality, finance sets discount limits and IT manages access to records. A named sponsor resolves conflicts beyond those delegated responsibilities.
Write down who can decide, who must contribute and where unresolved questions go. The division of responsibilities helps you check for gaps without assuming that a new committee or executive title will fill them.
