S Synthevo

Quick answer

What to Ask a Vendor Before Buying AI Lead-Response Software

Before you sign a contract, ask these 12 questions to any AI lead-response vendor — and learn which answers should make you walk away.

The Synthevo Team ·

TL;DR

Before buying AI lead-response software, ask about CRM integration depth, response latency guarantees, AI disclosure compliance, escalation logic, and contract exit terms — most vendors will dodge at least three of these, and that tells you everything.

Dealers who signed AI lead-response contracts in 2024 without asking the right questions are now paying for software that averages a 6-minute first response — when the industry benchmark for contact rate starts collapsing after 5. The demo looked great. The questions nobody asked would have ended the conversation.

This page gives you all 12 of those questions, plus the specific answers that should make you close your laptop and walk out.

Two men shake hands in a car dealership, sealing a business deal.
Photo by Vitaly Gariev on Pexels

Why Most Vendor Demos Are Designed to Hide the Answers You Need

Here is the contrarian truth that every GM needs to hear before the next vendor call: the most dangerous AI lead-response vendors are the ones with the slickest demos. Polished product walkthroughs are specifically engineered to avoid the seven or eight questions that would disqualify them. The demo environment runs on clean, staged data. It does not show you what happens when a lead comes in at 11:47 PM on Thanksgiving, the CRM throws a timeout error, and your BDC is offline.

The vendors worth buying from will not flinch when you go off-script. They will pull up real conversation logs on the spot. They will hand you their standard contract before you ask. If a rep needs to “check with the product team” on any question in the list below, you have your answer.

Q1: How Fast Does the AI Actually Respond — and Can You Prove It with Real Logs?

“Under two minutes” is the only acceptable answer. MIT’s 2024 lead-response study put contact-rate drop-off at five minutes — but in automotive, where a shopper has three other tabs open on Cars.com and AutoTrader, two minutes is already generous.

Ask the vendor to screen-share a production log showing timestamp-in (lead received) versus timestamp-out (first AI message sent) for at least 50 recent leads across different times of day. If they can only show you aggregate averages, ask why they cannot show individual records. Average response time of 90 seconds means nothing if 20% of leads wait 8 minutes because of batch processing.

Q2: Which CRMs Do You Natively Integrate With, and What Does ‘Integration’ Actually Mean?

“We integrate with VinSolutions” can mean anything from a real-time bidirectional API to a nightly CSV import. Those are not the same thing. Ask the vendor to define specifically: does it read lead status from the CRM in real time? Does it write activity notes back to the lead record? Does a human BDC rep logging into eLead or Reynolds see full AI conversation history without switching tabs?

Vague integration claims are the single most common source of post-contract regret. Dealerships running Synthevo today — including multi-rooftop groups like Vanguard Auto Group in Sterling, VA — see CRM activity written back to VinSolutions and eLead in real time, so BDC managers never have to reconcile two separate systems to know where a lead stands.

Q3: Does Your AI Disclose It’s an AI — and How Do You Handle FTC and State Disclosure Rules?

This is a legal question as much as a product question. The FTC’s updated guidance on AI-generated communications and state-level rules in California, Colorado, and Texas each have different triggers — but the common thread is that when a consumer directly asks whether they are speaking with a human, the AI must disclose.

Ask the vendor to show you exactly where that disclosure fires in a real conversation, what language it uses, and whether it is configurable or locked. Then ask how they stay current as rules change. A vendor with no documented compliance process is a liability you are signing onto.

Q4: What Triggers a Handoff to a Human, and How Fast Does That Happen?

The AI should handle qualification, not the close. Any vendor claiming their AI closes deals without human involvement is either overselling or serving a very different type of dealership than yours. Ask for the specific trigger list: price negotiation requests, trade-in appraisals, financing questions, angry sentiment, direct requests to speak with a person.

Then ask: when a handoff is triggered, what happens mechanically? Does it create a VinSolutions task? Send a text to the desk manager? Page the BDC rep via Slack? And what is the SLA — how long before a human is expected to respond? If the vendor does not have a defined escalation path with measurable timelines, the “handoff” is just the AI going silent and hoping someone notices.

Q5: Can I See a Real Conversation Log from a Dealership Like Mine?

Not a demo. Not a sanitized case study. A real log — with customer name redacted if needed — from a franchised dealer of similar volume to yours. This request alone separates vendors who are live in production from vendors who are mostly in pilot.

Look at how the AI handles a customer who changes their question mid-conversation, or who provides conflicting information about their trade. That is where AI falls apart or earns its keep. For more on evaluating BDC software in real-world conditions, see our Dealership BDC Software in 2026: An Honest Comparison.

Q6: How Do You Handle Leads That Come In After Hours or on Holidays?

After-hours leads are the whole point. If the AI treats a 2 AM lead the same as a noon lead — same response speed, same qualification depth, same handoff logic — you are getting what you paid for. If not, you need to understand what degrades and why.

Ask specifically whether the AI’s response behavior changes outside business hours, whether escalation paths change (they should, since your BDC is not staffed), and how appointment-setting works when the lead wants to come in at 9 AM and it is currently 1 AM. This is a more complex problem than it sounds — we cover the full logic tree in How Should a BDC Handle Leads That Come In at 2 AM?.

Q7: What Happens to My Lead Data — Who Owns It, Where Does It Go, and Can It Train Your Model?

Your lead data is a competitive asset. Find out explicitly: does the vendor’s contract give them rights to use your conversation data to train or improve their model? Many contracts include this in buried language. Some vendors anonymize it; others do not.

Also ask where data is stored, whether it is encrypted at rest, and what happens to your data if you cancel. If you cannot export a full conversation history in a machine-readable format within 30 days of contract end, you do not actually own your data.

Q8: What Does Your Pricing Model Look Like at Scale — Per Seat, Per Lead, or Flat Fee?

Per-lead pricing sounds reasonable until you run a big OEM incentive campaign and your inbound volume triples. Flat-fee pricing sounds safe until you realize it caps at a lead volume well below your busiest months. Per-seat pricing makes no sense for AI that does not sit at a desk.

Get the math in writing. Ask the vendor to model your current monthly lead volume and your peak-month volume and show you what the invoice looks like for both. Then ask what happens operationally — not just financially — if you exceed the tier.

Q9: How Long Is the Contract, and What Does It Cost to Exit Early?

Annual contracts with 60-day written termination notice clauses are standard. That is not inherently a red flag. What is a red flag: early termination fees equal to the remaining contract value, with no performance-based exit clause.

Ask specifically: if the AI fails to meet the response-time SLA you agreed to for 60 consecutive days, can you exit without penalty? If the answer is no, or if the vendor hedges, you are being asked to absorb all the performance risk.

Q10: What’s Your Average Onboarding Time, and What Do I Need to Provide?

“We can go live in two weeks” often means “if you get us everything we need in 48 hours and your CRM admin cooperates.” Ask for the specific onboarding checklist upfront: CRM API credentials, phone number provisioning, inventory feed setup, conversation playbook approval. Then ask what the average dealer actually takes to complete each step.

Onboarding delays cost you leads. Every week you are not live is a week your BDC is still handling volume manually. Vendors who cannot give you a realistic timeline with owner accountability on both sides are not set up to run enterprise deployments.

Q11: Do You Support 10DLC-Registered SMS, and How Do You Handle TCPA Compliance?

If the AI is sending SMS — and it should be, since text outperforms email for first response — the sending number must be registered under 10DLC (10-digit long code) through your carrier. Unregistered numbers face carrier filtering and delivery failure rates that quietly kill your contact rate without any error messages.

Ask for the vendor’s 10DLC registration process: who files it, how long it takes, and whether the number is shared across multiple dealerships or dedicated to yours. Shared numbers create brand and deliverability risk. On TCPA: ask how the AI handles opt-out requests and whether that opt-out is written back to your CRM immediately.

Q12: What Metrics Do You Report, and Can I Pull Them Into My Own Dashboard?

Vendors who control the reporting control the narrative. Ask whether metrics are available via API or data export — not just inside their proprietary portal. Specifically: first-response time, conversation completion rate, handoff rate, appointment-set rate, and lead-to-show correlation.

If the vendor’s reporting does not connect to your existing vAuto, Cox Automotive, or DMS reporting stack, you are going to spend the next 12 months arguing about whose numbers are right. That is a distraction you do not have time for. The ability to triangulate AI performance against actual deals closed in your DMS is what separates a line item from an accountable investment.

The One Question That Separates Real AI from Glorified Drip Email

After the twelve questions above, ask this: “If a customer responds to your AI’s first message with a completely unexpected question — something outside the standard qualification flow — what happens?”

The answer reveals whether you are buying actual conversational AI or a sophisticated drip sequence dressed up with a chat interface. Real AI reads the customer’s actual input and generates a contextually appropriate reply. Drip logic follows a script regardless of what the customer says. You can usually tell the difference in one follow-up question. Ask the vendor to demonstrate it live with a prompt you write on the spot.

Red Flags: Answers That Should End the Conversation

Vendor AnswerWhy It’s a Problem
”We integrate with all major CRMs” (no specifics)Almost certainly a connector, not a native API
”Response time averages under 2 minutes” (no log access)Averages hide outliers; outliers are where leads die
”Our AI never needs to disclose”Either wrong on the law or hiding the disclosure
”We don’t recommend sharing client conversation logs”Means they either have none or the quality is poor
”Early termination is pro-rated at the remaining contract value”You bear 100% of performance risk
”Compliance is the dealer’s responsibility”True legally, but a vendor with no compliance support is not a partner
”Our pricing is usage-based” (no tier ceiling defined)Your incentive month could cost you double

The vendors who hesitate on three or more of the questions in this list are not ready to run enterprise deployments. The market for dealership AI is moving fast — Amazon Vehicles is expanding its footprint in ways that will push digital lead volume significantly higher, and your BDC infrastructure needs to be ready. Read more on what that means operationally in Amazon Vehicles Is Expanding — Is Your BDC Ready?.

A common objection we hear: “We already have Conversica — why do we need to evaluate anything else?” Fair question. Conversica handles top-of-funnel nurture well, but dealers running Synthevo today, including the teams at Vanguard Auto Group across their 50-plus rooftops, consistently report that Synthevo’s escalation logic and CRM write-back depth handle the handoff-to-human moment in ways generic nurture tools are not built to do. The test is simple: run the 12 questions above on your current vendor. See how many they answer without hedging.

If you want to see how Synthevo answers every question on this list — with live logs, a real contract, and a working demo on your own lead data — request access to our live demo.

Frequently asked questions

How fast should AI lead-response software reply to a new lead?
Under 2 minutes for the first response is table stakes. Any vendor who cannot show you timestamped logs proving that performance at scale — not just in a demo — is selling you a promise, not a product.
Does AI lead-response software need to disclose it's an AI?
Yes. FTC guidelines and a growing number of state laws require disclosure when a consumer asks if they're speaking with a human. Any vendor that can't clearly explain their disclosure logic and show you where it fires in a real conversation is a liability.
What CRM integrations should AI lead-response software have?
At minimum, look for native two-way integration with VinSolutions, eLead, and CDK — meaning it reads lead status and writes back activity notes in real time. 'We support your CRM' via a Zapier bridge or CSV export is not a real integration.
What should I watch out for in AI lead-response vendor contracts?
Annual contracts with 60-day written termination notice and data-portability restrictions are the most common traps. Ask specifically what it costs to exit before the term ends, and whether your lead conversation data can be exported in a usable format.

Ready to see it?

Your next lead is texting right now.

Stop letting after-hours leads die in the inbox. Tell us about your store and we'll text you a live demo of Synthevo's AI Closer — usually within 30 seconds.