AI Can Answer the Phone. A New Report Says the Real Advantage Is What Happens Next.

The AI conversation just moved from "answering" to "doing"

For most of the past two years, the public story about AI was about answers: chatbots that could draft an email, summarize a document, or hold a plausible conversation. In 2026, the story has shifted. The industry's attention has moved to agents - systems that do not just answer a question but carry a task to completion: booking the meeting, updating the record, sending the follow-up, and handing off cleanly when a human is needed. The enterprise AI launches that drew the most attention this summer were less about a smarter chat window and more about software that finishes multi-step work on its own.

That distinction matters at a car dealership more than almost anywhere, because a dealership does not get paid for answering. It gets paid for finishing - the booked appointment, the logged lead, the completed sale, the returning service customer. And this month, that distinction got a name in auto retail.

The myth: Buying AI is the milestone. The data: 76% of U.S. dealers plan to increase their AI budgets in 2026, and voice agents are the single most common target at 74%. Yet a new July 2026 industry report argues the advantage by 2027 will go to the stores that connect that AI to the CRM, DMS, and deal record - not to whoever simply buys the most tools. Source: Spyne, 2026.

A new report calls it the "execution gap"

On July 16, 2026, the auto-retail software firm Spyne released a quarterly report with a blunt title: "AI in US Auto Retail: The Execution Gap Becomes the Battleground." Its central argument is that the meaningful divide in the industry is no longer between dealers who use AI and dealers who do not. Nearly everyone is using something. The divide is between stores that let AI sit on the edge - a standalone chatbot here, a pricing tool there - and stores that connect AI to the workflow that actually moves a customer from inquiry to delivery.

In the words of Spyne co-founder Sanjay Varnwal, "The practical divide is no longer between dealers using AI and those who do not, but between those who let AI sit on the edge and those who connect it to the customer, vehicle, and deal record." As with any vendor-authored report, it is worth reading the framing with the source in mind - Spyne sells connected-platform software - but the survey data underneath it and the direction of travel line up with what independent forecasters are describing.

The report frames a three-year arc that will feel familiar to anyone who has watched dealership tech evolve. 2024 was largely about content generation: AI writing listings and descriptions, enhancing photos. 2025 was about lead automation: AI answering chats and firing off follow-ups. And 2026 is about execution - whether the AI can actually complete the job across inventory, pricing, the CRM, and the service lane. Spyne's survey of nearly 1,200 dealership executives across 34 states found 76% planning to increase AI budgets, with AI voice agents the top investment priority at 74%, followed by merchandising and inspection automation at 68% and pricing and analytics at 62%.

Why "answering" was always the easy part

Consider a single inbound sales call on a Saturday afternoon. A basic AI receptionist can pick up on the first ring, sound pleasant, and take a message. That is the answering tier, and it is genuinely useful - an unanswered call is a lost opportunity, and after-hours coverage alone can justify a tool. But answering is not the same as finishing. Finishing that call means checking whether the exact vehicle the caller wants is still in stock and unsold, offering two real appointment times and writing one to the calendar, creating or updating the customer's CRM record so the salesperson does not start cold on Monday, and scheduling a follow-up if the customer does not show. None of that is conversation. All of it is integration.

This is the same line the broader AI industry is drawing in 2026 between a chatbot and an agent. A chatbot returns words. An agent takes actions in the systems that run the business. At a dealership, those systems are the CRM, the DMS, the inventory feed, and the service scheduler - and a tool that can talk but cannot touch any of them leaves the hardest, most valuable part of the job undone.

The backdrop: thinner margins make execution non-optional

None of this would matter much if dealerships were swimming in gross. They are not. The market in 2026 is stable but no longer growing. Cox Automotive's July forecast holds the full-year U.S. new-vehicle pace at 15.8 million units, a 2.9% decline from 2025, with first-half sales tracking about 3.6% lower year over year. June looked strong on paper - roughly 1.36 million units, up 7.2% year over year - but Cox attributes much of that to fleet activity and cautions that "affordability remains a central constraint," with demand propped up more by household wealth than by consumer confidence.

Affordability is the throughline. The average new-vehicle payment now sits around $722 a month on loan terms stretching close to 70 months, and shoppers increasingly choose by monthly payment rather than by model. When buyers are that payment-sensitive and volume is flat, every lead a store has already paid to generate becomes more valuable, and the cost of a mishandled call or an un-worked internet lead goes up. Spyne's report cites an outside projection from advisory firm Blue & Co. of a roughly 25% drop in profit per new vehicle retailed, and about 10% on used, for 2026 - a figure worth treating as one firm's estimate rather than gospel, but directionally consistent with the margin pressure dealers describe.

Here is the backdrop in one view:

When the pool of deals stops growing, you win by converting more of the demand you already have. That is an execution problem, not an answering problem.

Four questions that separate connected AI from a chatbot

If the 2026 advantage really does come from connection rather than conversation, a manager evaluating any AI tool - phone, chat, or otherwise - can pressure-test it with four questions.

1. Does it see live inventory and pricing? A tool that quotes a vehicle it cannot confirm is in stock, or a payment it cannot tie to real pricing, creates work and erodes trust. Connection starts with reading your actual data.
2. Does it write back to the CRM and DMS? A transcript emailed to the desk is not a CRM record. Ask whether the tool creates and updates the customer record, logs the interaction, and attaches next steps - so the human picks up with context, not from zero.
3. Can it complete the action end to end? Booking a real appointment on the real calendar, setting the service reminder, queuing the follow-up. If every action still requires a person to re-key it, the tool answered but did not finish.
4. Does it hand off cleanly and keep an audit trail? When the AI reaches the edge of what it should handle, does it pass the customer to the right person with the full history - and does it log what it did in a way you can review? This matters most in finance and service, where compliance and record-keeping are not optional.

A tool that clears all four is connected. A tool that clears one or two is a chatbot with good manners - useful, but not where the report says the advantage lives.

Start narrow, and measure what gets finished

The measured read of this moment is not "buy an all-in-one platform tomorrow." Even the report making the connected-AI case cautions that governed automation will scale first in bounded, lower-risk workflows - "scheduling, follow-up, inventory matching, service reminders, and deal preparation" - rather than a wholesale, AI-runs-the-store overhaul. That is the right instinct. The stores seeing results, per Spyne's survey, tend to have started with one or two well-defined jobs and connected them properly, rather than bolting on five disconnected tools.

Dealers who adopted AI in those bounded areas in 2025 self-reported gains worth checking against your own numbers rather than taking at face value:

Source: Spyne 2026 dealer operations report; figures are self-reported by adopters and will vary by store.

The practical takeaway for a dealer principal or GM this quarter is less about which vendor and more about which metric. If you are evaluating an AI phone or BDC tool, do not measure "conversations handled" or "questions answered." Measure appointments actually booked and kept, CRM records actually completed, and follow-ups actually sent. Those are the outputs the report says will separate stores by 2027 - and they are the ones that show up in the only scorecard that matters, your month-end statement.

The tools that merely answer will keep getting cheaper and more common. The edge is moving to the ones that finish the job.

If you are weighing where an AI phone agent fits in your store - and how to tell whether one is genuinely connected to your CRM, inventory, and service scheduling rather than just answering calls - the team at Carbuki works with dealerships on exactly that question.

Sources

• Spyne, "AI in US Auto Retail: The Execution Gap Becomes the Battleground" (Auto Retail Intelligence Quarterly, July 16, 2026), via PR Newswire: prnewswire.com
• Auto Remarketing, "Spyne report says 2026 will be 'first true AI Operations Year'" (survey of nearly 1,200 dealership executives; 76% increasing AI budgets; 74% voice agents; self-reported adopter gains): autoremarketing.com
• Techedge AI / PR Newswire, "Spyne Shows Connected AI Platforms Will Outpace Stand-Alone Tools in U.S. Dealerships by 2027": techedgeai.com
• Cox Automotive, "New-Vehicle Sales Pace Holds Strong Through First Half of 2026" (U.S. auto sales forecast, updated July 2026): coxautoinc.com
• U.S. News & World Report, "Average Auto Loan Interest Rates in July 2026" (average payment and term data, citing Experian): cars.usnews.com
• Blue & Co., "Automotive Tariffs" (2026 profit-per-vehicle projection, as cited by Spyne): blueandco.com