The ROI of AI Service Communication Is Bigger Than Your Dashboard Is Telling You

Numa’s own ROI framework, built from data across 1,300+ dealerships and more than 1 billion calls handled, starts from a specific premise: most GMs calculating ROI on AI service communication only count the easiest number to see, recovered appointments, and miss the two variables that usually matter more, retention and declined-service recapture. A complete calculation has five parts: cost, direct revenue recovery, avoided cost, retention value, and declined-service recapture. Most dealerships stop after the first two and walk away thinking the ROI is smaller than it actually is.

Why Most ROI Calculations on This Undercount the Real Number

The instinct when evaluating AI service communication is to ask a narrow question: how many more calls got answered, and how many turned into appointments. That’s a real number, and it matters, but it’s also the smallest piece of the actual financial picture. Cox Automotive’s 2026 Fixed Ops and Ownership Study found that customers who return to a dealership for service are 74% likely to repurchase their next vehicle from that same store, compared to just 44% for customers who don’t come back for service. That 30-point gap is worth more, over a customer’s ownership period, than almost any single appointment a communication tool recovers directly, and it’s the part of the ROI calculation most dealerships never put a number on at all.

Step One: Calculate the Cost Side

Start with what the tool actually costs, fully loaded, not just the subscription line. That includes the software cost itself, implementation time (typically two to four weeks for a single store), and what it’s replacing. Dealerships running separate point solutions for calls, texting, scheduling, and reviews typically spend $3,000 to $8,000 a month across those licenses alone, a baseline worth comparing against rather than evaluating a single connected system’s cost in isolation.

If the alternative under consideration is adding headcount instead of software, the comparison has its own math: a dedicated in-house hire breaks even, fully loaded, around 150 to 175 monthly contacts, and in-house BDC turnover running 35% to 50% annually means that cost recurs. Whatever the cost side ends up being, it’s the only part of this calculation most dealerships get right by default, since it’s the number on the invoice.

Step Two: Calculate Direct Revenue Recovery

This is the number most ROI conversations start and stop at: appointments that would have been lost, now booked. The calculation is straightforward: (missed calls recovered per month) × (appointment conversion rate) × (average repair order value).

For the last variable, NADA’s own topline figures, more than $164 billion in service and parts sales against more than 276 million repair orders, imply an industry-wide average RO value of roughly $594. Xtime’s own dealer data puts the figure slightly higher at $615 per RO among its network. Either number works as a reasonable input; what matters is using an actual figure rather than guessing.

After-hours capture is worth calculating separately, since it behaves differently than daytime volume. Cox Automotive’s Xtime research found after-hours appointments carry a $54 premium over a walk-in once captured, on top of being volume that wouldn’t have existed at all without coverage outside posted hours.

Step Three: Calculate Avoided Cost

The second most common gap in an ROI calculation is failing to count what a dealership avoided spending, not just what it gained. If AI communication is covering volume that would otherwise require an additional hire, that avoided cost is real money: a $45,000 to $55,000 fully-loaded hire, the training time to ramp them, and the turnover cost of refilling that seat when they leave. A dealership that would have added a night-shift or overflow position and didn’t should count that avoided cost as part of the return, not treat it as a separate, unrelated decision.

Step Four: Calculate the Retention Multiplier

This is the step almost every ROI calculation skips, and it’s usually the largest number in the entire model. The 74% versus 44% repurchase gap Cox Automotive’s research identified isn’t a soft, directional benefit. It’s calculable: take the number of service customers a dealership is retaining that it previously wasn’t, due to faster response, proactive updates, or catching a frustrated customer before they defected, and multiply by the value of a repurchase at that store, plus the same $12,000-plus lifetime service spend the same research found accumulates over a vehicle’s roughly 8.4-year ownership period.

A dealership that improves service retention by even a handful of customers a month through better communication is looking at a number that dwarfs the appointment-recovery calculation most GMs stop at. Real-time customer sentiment monitoring exists specifically to protect this number, catching the moment a customer’s sentiment turns before they become part of the defection statistic instead of the retention one.

Step Five: Calculate Declined-Service Recapture

The same Cox Automotive research found dealerships recapturing more than $35,000 a month in previously declined services once follow-up communication actually happened consistently, work that equity mining and declined-service follow-up exist specifically to automate rather than leave to whichever advisor has time that week. This is revenue that was already identified, already sitting in the DMS as a declined recommendation, and simply never followed up on. Counting it as part of an ROI calculation isn’t optimistic. It’s revenue the dealership already earned the right to and left uncollected.

Putting It Together: A Worked Example

A single-rooftop store recovering 40 previously missed calls a month, converting 60% into booked appointments at an average RO value of $600, generates $14,400 in direct revenue recovery. If 10 of those appointments happen after hours at the $54 premium Xtime identified, add another $540. If the same system avoids a $50,000 fully-loaded hire the store would have otherwise made, that’s avoided cost on top of the direct recovery. If improved retention keeps even three additional service customers a year who would have defected, each worth roughly $12,000 in lifetime service spend and a meaningfully higher repurchase probability, that’s another $36,000 in retained value the appointment-counting method never captures. Add even a fraction of the $35,000 a month Cox Automotive’s research found available in declined-service recapture, and the total return is routinely several multiples of what a GM sees by counting recovered appointments alone.

The scale of that gap shows up in real dealership numbers, not just the model. One Chrysler-Dodge-Jeep-Ram dealership tracked total revenue up 123% year over year after moving to connected AI communication, a number that includes exactly the categories most ROI calculations leave out: recovered calls, but also retention, after-hours capture, and follow-up that used to go undone. No single line item on a spreadsheet explains a jump that size. The five-part calculation above is what starts to.

What This Number Should Not Include

A rigorous ROI calculation is also honest about what not to claim. Not every retained customer stayed because of a communication tool specifically, and attributing 100% of a retention improvement to any single system overstates the case. The more defensible approach is calculating a plausible range, crediting the tool with the volume it demonstrably touched, missed calls it recovered, follow-ups it actually sent, heat cases it actually flagged, rather than the full retention or revenue number for the store overall. A GM who presents a conservative, well-documented number to ownership is in a stronger position than one who claims credit for revenue the tool didn’t actually influence.

The Bottom Line: The Real ROI Is Bigger Than the Easiest Number to Count

Most ROI conversations about AI service communication end at recovered appointments because that’s the easiest number to see on a dashboard. The actual return includes avoided headcount cost, after-hours premium capture, declined-service recapture, and a retention multiplier that Cox Automotive’s own research shows is worth a 30-point swing in repurchase likelihood. Numa’s own data across 1,300+ dealerships reflects all five of these categories, not just the appointment count, because a GM who only measures the easiest number to find is systematically underselling the case for the tool they’re already running.

Frequently Asked Questions

What’s the simplest way to calculate ROI from AI service communication?

At minimum, multiply the number of previously missed calls now recovered per month by the appointment conversion rate and the average repair order value, then subtract the fully loaded software cost. That gives a conservative baseline, but it excludes avoided headcount costs, after-hours revenue premiums, retention value, and declined-service recapture, all of which typically make the real return several times larger.

What is the average repair order value to use in an ROI calculation?

NADA’s national topline figures imply an industry-wide average of roughly $594 per repair order. Xtime’s own dealer network data puts the figure at approximately $615. Either is a reasonable input; the key is using an actual benchmark rather than an estimate.

How much is a retained service customer actually worth?

Cox Automotive’s 2026 Fixed Ops and Ownership research estimates more than $12,000 in service spend over a vehicle’s average 8.4-year ownership period, based on 2.4 service visits a year at roughly $615 per visit. Retained service customers are also 74% likely to repurchase their next vehicle from the same dealership, compared to 44% for customers who don’t return for service.

Should avoided headcount costs count toward ROI?

Yes, if the AI communication system is genuinely covering volume that would have otherwise required an additional hire. A fully loaded in-house BDC hire typically costs $45,000 to $55,000 annually and carries 35% to 50% annual turnover risk at many stores, both of which are real, calculable costs avoided when that volume is covered without adding the position.

Is it accurate to credit an AI communication tool with 100% of a retention or revenue improvement?

No, and doing so weakens the credibility of the calculation. The more defensible approach credits the tool with the specific volume it demonstrably touched, calls it recovered, follow-ups it sent, heat cases it flagged, rather than claiming full credit for a store’s overall retention or revenue performance, which has multiple contributing factors.

See how Numa’s own data breaks down ROI across all five categories, not just recovered appointments. Talk to Numa.