Got AI Questions?
We've got answers.
Getting Started & Setup
How does does it actually work?
Vepsy's AI voice agent answers your business line with a natural, humanlike voice, handles the conversation end‑to‑end, and captures the details that matter. It can route requests, create the next step in your systems, and keep customers informed with automated updates—delivering consistent coverage 24/7 without breaks, burnout, or missed calls.
How long does it take to set up?
Most Vepsy voice agents go live in 1–2 weeks. That includes discovery, conversation design, integrations, and real-world testing.
If you need minimal integrations (answering, qualifying, routing, taking messages), we can often launch faster. More complex requirements—custom workflows, multiple departments/locations, or custom API work—may extend the timeline, and we'll confirm that upfront with a clear plan.
Do I need special hardware or equipment?
No. Vepsy works with your existing phone number and call flow—no new devices, headsets, or on‑prem hardware to install. If we connect to tools like your CRM, calendar, or ticketing system, it's done through standard integrations or secure API access, not new equipment.
What information do you need from me to get started?
Just the essentials that define a “great call” in your business: your top call types, what outcomes you want (book, qualify, support, route, take a message), and the rules your team follows to get there. If you can share any scripts/FAQs, examples of real calls, business hours, service areas, pricing or eligibility guidelines, and the tools you use (phone system, CRM, calendar, helpdesk), we'll move faster—but we can build the playbook with you even if nothing is documented yet.
Technical & Operations
What happens if the AI isn't confident or the caller goes off-script?
Vepsy is designed to fail gracefully:
- It asks clarifying questions when information is unclear.
- It switches to a simpler “safe” flow (capture details → route/escalate).
- It escalates to a human when confidence is low, the caller is frustrated, or the request falls outside your allowed scope.
You decide the thresholds: some businesses prefer aggressive containment; others prefer earlier escalation. We tune that during onboarding.
How do you handle noisy calls, accents, or poor connections?
We use robust speech recognition tuned for phone audio and real-world conditions. When audio quality is poor, Vepsy:
- confirms key details (names, numbers, addresses),
- repeats back critical fields,
- offers alternate capture methods (e.g., “I can text you a link to confirm details” if enabled),
- escalates when accuracy is at risk.
This is also why we build flows that verify important information instead of assuming it.
Can you integrate with our CRM/helpdesk/calendar/scheduling tools? Which ones?
Yes. Vepsy commonly connects to CRMs, calendars, helpdesks, and schedulers so calls create real outcomes (new lead, ticket, booking, notes). We support standard integration methods and can often connect to popular tools quickly. If you share your stack, we'll confirm what's native vs what we'll do via API/webhook.
What data do you store (audio, transcripts, extracted fields)? Can we control retention?
Vepsy can store:
- call metadata (time, duration, outcome),
- transcripts and summaries,
- extracted fields (name, intent, booking details),
- and optionally call audio (depending on configuration).
Retention is configurable. You can choose what's stored, for how long, and whether certain fields should be masked or not stored at all.
Can it handle multiple concurrent calls? What's the capacity model?
Yes—Vepsy can handle multiple calls simultaneously. Capacity is designed to scale with your demand (including peak hours), and we can set concurrency limits and overflow behavior so callers still get answered without overwhelming downstream teams.
AI Customer Experience
How natural does the AI voice sound?
Our various agent voice models are as realistic as it gets. Replicated from actual voice actors, they follow the cadence, syntax, and vernacular of natural speech.
Do you offer Spanish Language AI agents?
Yes. And coming soon: French!
Can customers tell they're talking to AI?
The AI communicates naturally with customers and understands industry terminology and regional accents. Many users cannot tell the difference between our agent voices and a real dispatcher. In order to be up front and honest with your customers, we recommend instructing the caller that they are speaking to a virtual AI dispatcher to avoid confusion.
What if the AI doesn't understand the customer?
Our AI is trained specifically for towing scenarios and understands industry terminology and regional accents. In the rare case of confusion, it knows how to ask clarifying questions. If needed, it can seamlessly transfer complex calls to a human backup.
What happens if a customer needs to speak to a human?
The AI can seamlessly transfer complex calls to your designated backup number.
Pillars, Pricing & Accounts
Can I buy just one pillar?
Yes, and most operators should. The five pillars — Attract, Qualify, Service, Collect and Retain — are sold separately, and nothing is gated behind anything else. Run the console on calls your staff take by hand. Run billing on jobs that came from a system we have never heard of. Put the voice agent on your line and pass calls to whatever you already use.
Buy the pillar you are missing. If the answer is a pillar we do not sell, we will tell you that on the first call.
What happens when I add another pillar?
The pillars you already run keep working, and the new one starts sharing what it needs. A qualified call becomes a dispatchable job instead of a message. A closed job bills itself against the authorization taken during the call. A completed job asks for the review by itself.
Adding a pillar is a conversation about configuration, not a re-implementation. There is one platform fee whether you run one software pillar or all four, so adding a second does not add a second fee — only its own usage.
Why isn't everything priced per minute?
Because only one of the pillars actually costs by the minute. Per-minute pricing on the voice agent is honest: a longer call really does cost more to run.
Applying it to the rest would be a fiction, and a bad one. The console's cost tracks jobs dispatched, not how long someone talked, and billing tracks transactions. There is also a perverse incentive buried in per-minute-everything: as the agent gets better it talks less, so we would be charging you less for a product that got better. We would rather meter each pillar in the unit it genuinely consumes.
Do my customers or field techs need an account?
No — and they never will. Only your staff sign into the console, and those accounts are provisioned by hand.
Your customers are reached by phone and text. Your field techs close jobs from a one-time link texted for that specific job: no app to install, no password to remember, and the link stops working once the job is closed. That is deliberate architecture, not a missing feature.
Are there tokens or credits to buy?
No. No tokens, no credits, no points, no synthetic units of any kind. You are billed in minutes, jobs and transactions — things you can count on your own records and argue with us about if the number looks wrong.
How is marketing billed?
Separately, always. Attract is a monthly retainer for the work plus your own ad spend, quoted per engagement. It never shares a total with the platform fee, and you can stop one without touching the other.
Questions?
Contact Us
Ready to turn your phone line into a real operator—not a voicemail box? Vepsy builds AI voice agents that answer instantly, handle the routine, and route the important calls with full context.
Book a quick call and we'll map your top call types, show a tailored demo, and outline a clear path to a live deployment that fits your systems.
