What "AI pre-sales" actually means for an operating store
Pre-sales is the question a buyer asks before they pay. It is "will this ship before the 20th," "is the navy darker in person," "do you do returns." Answer it fast and accurately and the order usually lands; leave it sitting and the tab closes.
The SERP for this keyword is mostly vendor pages promising to "turn browsers into buyers" with almost no numbers behind it. One widely-cited article asserts most customers are likelier to buy after a fast reply, but attaches no source to the figure (ecommercegermany.com). You run a real store — you need the math, not the slogan.
Here is the honest map. The AI that can handle pre-sales for you comes in three layers, and you are already using the first.
Layer 1 — automation already inside your stack
Your platforms ship AI that works pre-sales inside their own walls. Klaviyo's generally-available Customer Agent handles order tracking, returns, and subscription questions across chat, SMS, email, and WhatsApp (Klaviyo). Shopify's Sidekick can draft product answers and content and "continues working in the background," presenting changes "for your review before applying them" (Shopify).
The catch: each one is blind outside its own tool. Klaviyo's agent can't see your Printify production status; Sidekick can't touch your Meta budget. A pre-sale question rarely stays inside one tool.
Layer 2 — single-surface support agents (priced per resolution)
The most mature AI-agent category is support, and pre-sales rides on it. Gorgias charges per resolved conversation — "$0.90 on most plans," with starter plans at "$1 per resolved conversation" (Gorgias). Zendesk's AI agents are billed on "the successful outcomes they deliver," with Suite plans starting at fifty-five dollars per agent per month billed yearly (Zendesk).
Two structural facts matter here. Support AI is now priced like an outcome, not a seat — you pay when the conversation is fully resolved without a human. And both vendors build in a handoff path, which is the business model admitting these agents do not handle everything.
Layer 3 — cross-tool AI employees
The newest layer works across your tools the way a hire would: read the store, the ads, and the supplier together, then take multi-step action with your approval. Analysts call the underlying capability agentic AI — a system that can "act in the real world and execute multistep processes," distinguished from a chatbot by the acting, not the chatting (Solo.io, quoting McKinsey).
This is the layer where a single pre-sale question — "where's my order and can I still change the size" — gets answered by looking up the order in your store, checking supplier status, and drafting the reply in one loop. PodVector AI's Victor is an example: an AI employee that integrates with Shopify, Meta Ads, Google Ads, Printify, Printful, Gelato, and Klaviyo, and handles customer-support email with your approval before anything sends. The cross-tool guide at AI employees for ecommerce walks the full distinction.
Chatbot vs human VA vs AI employee
These three terms get used as synonyms. They name three different things, and picking the wrong one for pre-sales wastes money.
| Chatbot / single-surface agent | Human virtual assistant | AI employee | |
|---|---|---|---|
| What it is | Converses and resolves on one surface | A remote human contractor | Multi-step actions across tools, approval-gated |
| Scope | One channel (the inbox or widget) | Whatever you train them on | Every tool it integrates with |
| Availability | Round the clock | Their working hours | Round the clock |
| Priced | Per resolution or per seat | Per hour or per month | Subscription, usually usage-based |
| Fails how | Confidently wrong answers | Slowly, visibly, recoverably | Wrong actions at scale if ungated |
Pricing and capability cells above trace to the Layer 2 sources (Gorgias; Zendesk) and the VA rate sources in the worked example below.
Three things worth spelling out. A chatbot answers; an agent acts — a widget that tells a customer how to change their order is a chatbot, a system that changes it is an agent. A "virtual assistant" in hiring still means a human, so don't confuse an AI tool with a contractor. And "AI employee" is a scope claim, not a magic claim — Gartner warns of "agent washing," rebranding chatbots as agents, estimating "only about 130 of the thousands of agentic AI vendors are real" (Gartner). For the chatbot end specifically, compare an enterprise AI chatbot for ecommerce against a full AI chatbot service for ecommerce.
What AI pre-sales handles well — and what it botches
The honest split: AI closes the routine questions reliably, and it is dangerous on the edge cases.
Automates well: order-status, shipping-time, returns-policy, and "does it come in another size" questions. These resolve from structured data, which is why Gartner predicts agentic AI will "autonomously resolve 80% of common customer service issues" by a few years out — the word "common" is the whole qualifier (Gartner).
Botches badly: the ambiguous, high-stakes promise. Air Canada's chatbot told a customer he could claim a bereavement discount after flying — contradicting policy — and a tribunal ordered the airline to pay, rejecting its argument that the bot was "a separate legal entity responsible for its own actions" (CBC; ABA). Whatever your pre-sales AI promises a buyer, you own it. That is exactly why every serious vendor gates consequential actions on human review — Victor drafts the reply and you approve the send.
The pre-sales math: a worked example
Say your store does 340 orders a month at a $31 AOV, with $2,800 a month in Meta spend. You field roughly 300 pre-sale questions a month — mostly sizing, shipping, and "is it in stock."
Compare the options for handling them. (VA rates: offshore mid-level six to ten dollars an hour per DDIY; US fully-loaded twenty-eight to sixty-five dollars an hour per CallForce. AI per-resolution rate from Gorgias.)
A human VA handling all 300 at eight minutes each is 40 hours. At $8/hour offshore that is 40 × $8 = $320/month; at $40/hour US it is 40 × $40 = $1,600/month — and bounded by their timezone.
An AI agent resolving the easy half (150) at $0.90 is 150 × $0.90 = $135, plus the helpdesk subscription. The remaining 150 at eight minutes is 20 human hours — about $160 offshore or $800 US. The AI's half answers instantly, around the clock.
The reading: against a US cost baseline, per-resolution AI is dramatically cheaper. Against a cheap offshore VA the dollar gap on 300 questions is small — so the real AI argument at this volume is instant, around-the-clock answers on the questions that convert, not raw price.
The profit angle every pre-sales pitch skips
Here is the number that changes the decision. A converted pre-sale is worth your per-order profit, not your $31 revenue.
Walk it on that same store. Say the blank plus fulfillment from your POD supplier runs $14, and payment and platform fees run about $1.20. Your ad cost per order is $2,800 ÷ 340 = $8.24. So per-order profit is $31 − $14 − $1.20 − $8.24 = $7.56.
That reframes everything. Paying $0.90 to resolve a pre-sale question that lands a $7.56-profit order is a clean trade; paying a US VA $5.33 (eight minutes at $40/hour) to do the same is most of the profit gone. It also means not every pre-sale is worth equal effort — a question on your thinnest-margin SKU can convert at a loss once the ad cost is counted.
This is where a cross-tool AI employee separates from a support widget. A widget knows the answer; it does not know the order's true per-order profit. Victor computes true per-order profit across your store and your POD suppliers, so the same system that drafts the pre-sale reply also knows which orders are actually worth chasing — and saves the report to your own Google Drive.
What to expect if you adopt one
Expect a ramp, not a switch. Gorgias says the automation rate "emerges from usage over time" — the AI needs your policies, shipping rules, and catalog before its resolution share climbs (Gorgias).
Expect to keep reviewing. Liability sits with you (the Air Canada precedent), so budget review time as the cost that replaces answering time. And expect vendor churn — Gartner predicts "over 40% of agentic AI projects will be canceled" within a couple of years (Gartner). Favor tools whose work product — replies, reports, order edits — lives in your accounts, so the artifacts survive the tool.
The defensible headline metric is time off your calendar, not a revenue-lift promise. What you do with the reclaimed hours decides the P&L effect. If you want a human-plus-AI support build-out, see customer-service AI chatbot solutions for ecommerce, and if you are weighing building versus buying, the piece on when to hire AI developers lays out the tradeoff.
Want an AI employee that drafts your pre-sale replies, checks the order and supplier, and tells you each order's true profit before you chase it? Meet Victor.
FAQs
What is an AI pre-sales solution for ecommerce?
It is software that detects and answers a shopper's pre-purchase question — sizing, shipping time, stock, returns — fast enough to convert the browse into an order. It ranges from a chat widget on one surface to a cross-tool AI employee that answers the question and acts on it across your store, ads, and supplier.
Is pre-sales AI the same as a customer-service chatbot?
Mostly it rides on the same engine, but the goal differs. A support chatbot resolves a problem after purchase; a pre-sales solution catches a buyer before checkout, where the payoff is a closed order. The best ones use the same structured store data for both.
How much does AI pre-sales cost?
The dominant model is per resolved conversation — about $0.90 on most Gorgias plans, with Zendesk billing on outcomes delivered (Gorgias; Zendesk). Judge it against your per-order profit, not your AOV: at a $7.56 profit per order, a sub-dollar resolution that converts is a clear win.
Can AI pre-sales replace my support VA?
No — it concentrates your human on the hard half. Outcome-priced AI is built on the handoff: you pay only for what it fully resolves, and the rest routes to a person. The team shrinks per question; it does not vanish.
Does an AI employee answer live chat?
Victor handles customer-support email with your approval — it drafts the reply, and you approve the send before anything goes out. The point is that it does not answer in a vacuum: it can look up the order in Shopify and the production status with your POD supplier first, so the answer is grounded in your live data, not a guess.
What's the risk if the AI gets a pre-sale answer wrong?
You own it. A tribunal held Air Canada liable for its chatbot's wrong answer and rejected the "separate legal entity" defense (CBC). That is why approval gates on consequential replies are a feature, not a limitation — the human stays the decision-maker of record.