Every OEM has a digital-first retail strategy. Configurators, online checkout, direct-sales models, customer journeys designed at headquarters to feel like the brand: precise, premium, consistent.
Then the customer has a question about an actual car in an actual market, and the strategy meets its weakest link: whoever happens to answer at the local retailer, if anyone answers at all.
The brand promise is central. The delivery is local.
This is the structural problem of automotive retail digitization. The experience is designed centrally but delivered at the edge — by hundreds of retailers with different staffing, different training, different moods on a Tuesday evening. A brand can control its website pixel by pixel and still lose the customer in the first phone call to a dealership.
The markets far from headquarters get the worst of it. Digital investments roll out flagship-first; smaller markets wait years. Languages outside the top five barely get supported. The customer in Jeddah or Warsaw experiences a different brand than the customer in Munich — not by design, but by neglect of the edge.
Why importers are becoming the interesting layer
National sales companies and importers sit exactly at this seam: accountable for the brand experience in their market, but without headquarters' budgets or timelines. They can't wait for the global platform roadmap to reach them.
That constraint is turning some of them into the most advanced digital operators in their networks. The BMW importer's retail organization in Saudi Arabia runs AI sales advisors across its web presence today: full model range, list-price governed, Arabic and English, around the clock, with test-drive requests verified and delivered to sales teams as qualified leads. It didn't wait for a global program. It deployed at the edge, where its customers actually are.
Governance is the real unlock
The instinctive OEM objection to AI at the retail edge is control: what if it says the wrong thing, invents a discount, misrepresents the brand?
It's the right question, and it has a technical answer: governance enforced at platform level, not by hoping. Pricing rules that only allow list prices. Brand tone configured centrally, per market. Every conversation logged and auditable. An agent under platform governance is more consistent than any human network can be — it never freelances a discount, never has a bad Tuesday, and says exactly what the brand allows in every language it speaks.
That inverts the control argument. Today, the least controlled part of most brand experiences is the human edge. The governed agent is the consistent one.
The pilot path that doesn't die in IT
The second OEM objection is process: every innovation dies in an integration queue somewhere between procurement and IT security.
The deployments that work start outside that queue: a pilot on the importer's own infrastructure, against public data — the model range, prices, locations that are already on the website. No system integration required to prove value. Deep integration into DMS and CRM comes after the numbers make the case, when the project has champions with results, not slideware.
For OEMs and NSCs, the question is no longer whether AI belongs at the dealer edge. Customers already assume they can ask a question at 11pm and get a competent answer. The question is whether the brand's edge answers — or the aftermarket's does.