Ask retail decision-makers whether AI agents will shape how their customers buy, and the argument is largely settled. Looking out ten years, more than nine in ten expect agentic commerce to leave a lasting mark on retail. 86.1% say it is likely that AI agents will influence their customers' purchase decisions.
Then ask who owns it.
Everyone expects it. Nobody owns it.
In a study SoftServe commissioned from the EHI Retail Institute, 72.4% of retailers from the DACH region report no permanent ownership for agentic commerce. 37.6% handle it project by project. 34.8% have no organizational ownership at all. A shift that almost everyone expects to reshape their business sits, for most retailers, on nobody's desk.
Figure 1. Only 27.6% of retailers have a permanent owner for agentic commerce.
Source: EHI Retail Institute study, commissioned by SoftServe. The 27.6% figure is derived from the reported 72.4%.
The gap isn't technical
The agent technology is not what is holding retailers back. Their own commerce data is. Only 39.2% rate their data foundation as good or very good, and data is the precondition for everything an agent does. An agent that cannot read consistent product attributes, live stock, current pricing and promotions, and the content that describes them cannot recommend, compare, or buy correctly. It will do all three badly and confidently.
Your customer won't read that as an agent problem. They'll read it as your brand being wrong about its own products. Staffing is a separate constraint, and a roadmap does not close it. Only about one in five retailers feel adequately staffed for this work. Even where the strategy is clear, the people to execute it are often not in the building.
The basic build question is unsettled too. Nearly half of retailers have not yet decided whether they will build in-house, buy externally, or combine both. However, a retailer who builds in-house and a retailer who buys can land in exactly the same place. What separates them is whether the system learns. An agent that recommends the same way in month six as it did in week one is a static catalogue with a conversation on top.
Figure 2. Expectation is settled. Readiness is not.
Source: EHI Retail Institute study, commissioned by SoftServe. Permanent-owner figure derived from the reported 72.4%.
That’s why learning is a critical design decision, and it has to be made early. Conversion results, A/B tests, returns, service contacts and campaign performance need to feed back into the models behind recommendation, pricing and content, without anyone scheduling a re-tuning exercise. Retrofitting that later means rebuilding.
None of these are fundamentally engineering problems. They are decisions about ownership, data, and where intelligence sits. All of them can be made before the technology question is settled.
Retailers want help. They don't want a landlord.
More than half of retailers are willing to work with a service provider on implementation: 23.6% already do and 32.4% plan to. More than half also do not intend to integrate external agentic-commerce applications into their own customer interface.
Retailers want to build their own capability. They do not want to hand the customer relationship to someone else's platform.
72.6% of retailers expect a small number of large external agents to dominate access to customers.
Source: EHI Retail Institute study, commissioned by SoftServe.
This concentration of access could deepen retailers' dependence on major platforms and LLM providers. Retailers are not expecting one side to win outright, though: 63.1% expect strong brands to run their own agents and 61.9% expect owned and external systems to interoperate. The future is hybrid, and the strategic question is what you own inside it.
So the ownership question has two layers. Who owns agentic commerce internally, and what part of the customer relationship you still own once agents sit in the middle of it.
The category has no leader yet
Measured impact of agentic commerce today is low. Only about one in ten retailers see a major influence on retail right now. Nobody needs to panic, and anyone selling urgency is overselling.
There is nevertheless a strong argument for starting now, while the market is still forming. No retailer has yet been observed scaling agentic commerce end to end. So, the leadership position is open, and it goes to whoever has the data foundation in place, orchestrated agentic systems that learn from their own results, and the internal clarity to move when the investment window opens.
What to do in the next two quarters
Three moves are available to aspiring leaders now:
- Name an Owner. One accountable structure, not a standing committee. Agentic commerce spans e-commerce, IT, strategy, and marketing, and in practice legal too.
- Audit Data Readiness. Pick a single journey, trace the data an agent needs to complete it, and find where it breaks. That gives you a backlog, not a maturity score.
- Ship Search First. Product search and advice launches inside a quarter and is public enough to teach you something, but only if you instrument the feedback loop.
All three fit inside two quarters, and each one is a decision your organization can make this week. Nearly half of retailers have not decided how they will build, which is usually a sign that nobody has measured the distance between where they are and where they need to be. That is where SoftServe supports: a short assessment and gap analysis, then a roadmap built on what it finds, rather than an open-ended transformation programme.
Get the full study
The study has the rest of the picture: how fast retailers expect the shift to land online versus in store, when the money starts moving, and how far the market has travelled from AI-assisted search toward autonomous procurement.
