AI Phone Agents at Dealerships: What 2026's Agent Scaling Gap Means Before You Buy

For all the money dealerships pour into digital advertising and CRM tools, the highest-intent moment in the entire customer journey still happens on the phone. A shopper who calls is closer to buying — or to booking service — than almost anyone filling out a web form. That is exactly why AI voice agents have become one of the most aggressively pitched products in retail automotive this year.

The pitch is seductive: point an AI at your phones and never miss another call. But a run of fresh 2026 research from McKinsey, Gartner, and a survey released just this week suggests the smarter question is not whether to use AI on your phones, but how — because most organizations that deploy AI agents are still struggling to control them.

Myth: Putting an AI agent on your phones is a flip-the-switch win.

Data: In a survey of more than 400 IT leaders released June 17, 2026, 79% of enterprises running AI agents said they had already been forced to reverse an action an agent took, and 42% tied lost revenue to an agent failure (Kore.ai Agent Productivity Index, 2026). Separately, only 23% of organizations have scaled agentic AI anywhere in their business (McKinsey, 2025).

The phone problem is real — and bigger than most stores think

The demand side of this story is not in dispute. Call-tracking firm Car Wars tracked roughly 53 million inbound service calls across its dealership network in 2025 and found that about 19 million were missed opportunities, according to figures its CEO shared with CBT News. More than half of those missed calls (53%) went to voicemail, and 29% of callers simply hung up after waiting on hold too long. Despite the enormous volume, only about 4 million service appointments were ultimately booked.

The timing makes it worse: Car Wars found service-call volume peaks between 10 a.m. and noon — the same window when advisors are already slammed with customers in the lane. We have written before about what those missed calls actually cost, and the fixed-ops math is unforgiving.

Buyers, meanwhile, are warming to AI. In Cox Automotive's 16th annual Car Buyer Journey Study, shoppers who engaged AI tools reported among the highest satisfaction of any group, and 63% of dealers said investing in AI now is critical to long-term success. The appetite is there. The open question is execution.

Why 2026's biggest AI story is a yellow light, not a green one

Agentic AI — software that does not just answer questions but takes actions and completes multi-step tasks — is the defining enterprise-technology story of 2026. It is also, by the industry's own numbers, still immature.

McKinsey's latest State of AI survey found that 88% of organizations now use AI in at least one business function, but only 23% have scaled an agentic AI system anywhere in the enterprise, and in any single function no more than 10% have. Gartner is blunter: it predicts more than 40% of agentic AI projects will be canceled by the end of 2027, citing escalating costs, unclear business value, and inadequate risk controls — and it warns of agent washing, where ordinary chatbots and scripts get rebranded as agents.

This week's data sharpened the point. The Kore.ai Agent Productivity Index, a survey of more than 400 IT leaders at large U.S. companies released June 17, found that enterprises are adding agents faster than they can govern them.

Source: Kore.ai Agent Productivity Index, 2026 — survey of 400+ IT leaders at U.S. organizations with 2,000+ employees; margin of error plus or minus 3.0 points.

What the scaling gap means for your phones specifically

None of this means AI does not belong on a dealership's phones. It means the failure modes are real and worth designing around. A poorly governed phone agent does not just annoy a caller — it can quote a payment it should not, book a service appointment into a slot that does not exist, fumble a recall or safety question, or fail silently at 9 p.m. with no human to catch it.

The enterprises getting value are not the ones that deployed fastest. McKinsey found that high performers share two habits: they redesign the workflow around the AI rather than bolting it on, and they define in advance when a human needs to validate the output. Gartner's guidance echoes it — pursue agentic AI only where there is clear ROI, and rethink the process from the ground up rather than dropping an agent onto a broken one.

A measured way to put AI on the phones

For a dealership, that translates into a handful of practical principles:

• Start where volume is high and ambiguity is low. After-hours coverage, hold-time overflow, hours-and-directions questions, and routine service scheduling are ideal first jobs. Car Wars reports its conversational AI books a service appointment in under two minutes, versus about four minutes and 38 seconds for a human agent — a real efficiency gain on a well-bounded task.
• Keep a human escalation path. The agent should recognize what it does not know and warm-transfer a sales or finance question it cannot confidently handle.
• Demand traceability. Every call should be logged, transcribed, and written back to your CRM, with alerts when a promised follow-up does not happen. The Kore.ai data is clear that untraceable failures are where the damage compounds.
• Set the guardrails before go-live, not after. The report's structural finding is that controls bolted on after an agent is already running tend to govern only that running agent; governance works best when it is designed in from the start.
• Measure the right things. Answer rate and booked appointments, yes — but also error and reversal rate, and customer-satisfaction scores specifically on AI-handled calls.

This is also the difference between a generic AI assistant and a system built for the retail-automotive workflow — the kind of AI BDC or service-scheduling setup that integrates with your DMS and CRM, looks up recalls, respects advisor schedules, and routes lender questions to a human. That is where a phone agent either earns its keep or becomes a liability.

The bottom line

The opportunity is genuine. The phone is the highest-intent channel a dealership has, buyers increasingly reward AI-assisted experiences, and the missed-call problem is too large to simply staff your way out of. But 2026's data delivers a useful warning: the dealerships that win with AI on the phones will not be the ones that flip the switch fastest. They will be the ones that deploy with escalation paths, traceability, and measurement — treating an AI voice agent as a workflow to be managed, not a gadget to be installed.

If you are evaluating AI phone handling for your store, that governance-first lens is worth bringing to every demo. Carbuki builds AI voice agents for U.S. dealerships with human escalation and CRM logging designed in, and we are glad to talk through what a measured rollout looks like for your store.

Sources

• The state of AI in 2025: Agents, innovation, and transformation — McKinsey & Company, November 2025
• New Kore.ai Survey: 72% of Enterprises Say Their AI Agents Operate With Unmanaged Risk — Kore.ai via Business Wire, June 17, 2026
• Gartner Predicts Over 40% of Agentic AI Projects Will Be Canceled by End of 2027 — Gartner, June 2025
• Car Wars CEO Stephane Ferri on fixing the industry's biggest communication gap — CBT News, March 2026
• Cox Automotive Car Buyer Journey Study Finds Efficiency, Digital Tools and AI Drive Record Satisfaction — Cox Automotive, January 2026