The vehicle stated once, every answer filtered to fit
Fitment confirmed in the chat, never assumed
Grounded in your Magento catalog and fitment data
Why fitment first
Why Auto Parts Need a Fitment Assistant, Not a Generic Chatbot
41 intents out of the box1 vehicle stated once
Auto parts are the one catalog where the right product is wrong for most buyers. A brake pad set fits a handful of the vehicles in the industry's Vehicle Configuration Database, the reference behind the ACES fitment standard, and a shopper who guesses sends the part back, with the return label, the restock and the lost margin that follow.
Generic chatbots make that worse. They read a product title, not a fitment table, so they recommend confidently and wrongly. The year, make and model selectors most stores bolt on help, until a shopper lands from search on a product page, asks a question the dropdowns cannot take, or gives up on five menus on a phone screen and calls the parts desk instead.
The ZestIQ Sales Agent treats the vehicle as the first fact of every conversation. Stated once, it filters every search, every recommendation and every add to cart to parts that fit, from the fitment data your Magento store already holds, and it says so in the answer. When the data does not confirm a fit, it says that too, and hands the shopper to a person with the vehicle attached.
Where fitment fails
Six Ways Auto Parts Stores Lose the Sale
Every one of these ends the same way: a part that did not fit, a return label, and a customer who buys the next one somewhere else.
Wrong Part Ordered, Then Returned
The shopper picked the listing that looked right. It fitted the other trim level. The return costs the margin on the sale and the next one.
Dropdowns Abandoned on a Phone
Year, make, model, engine, trim: five selectors on a phone screen before a single part shows. Half the shoppers never finish them.
Search That Ignores the Vehicle
The shopper lands from Google on a product page, types a part name into search, and gets every variant for every vehicle you sell. The selector they never saw is three pages back.
Phone Calls to Confirm Fitment
The parts desk answers the same question all day: will this fit my truck. Each call is a sale that waited and a person who could not be doing anything else.
Part Numbers Nobody Can Type
The number on the old part is half worn off, the shopper types what they can read, and search returns nothing. A prefix match would have found it.
Chatbots That Guess
A generic chatbot reads the product title and recommends with confidence. It has never seen your fitment table, and the shopper finds that out on the driveway.
In the conversation
What the Fitment Assistant Does in a Parts Conversation
-
Exact part number first, then a prefix match, then semantic search across the attributes you choose. A shopper can type the number off the old part, describe what they need, or just name the truck, and the agent shortlists what fits and says why.
-
Year, make, model and the attributes your fitment data uses are captured once. Every search, recommendation and add to cart after that is filtered to parts that fit, and the fit is confirmed in the answer rather than left to a dropdown.
-
The item goes in from the chat after the agent checks fitment, stock and options against Magento. The addition is attributed to the conversation, so you can see which parts the assistant sells.
-
Order status, carrier and tracking are shown only once the customer has verified the order is theirs, and never to anyone else. A freight shipment is explained as freight, with the collection rules your pages state.
-
Core charges, freight collection, warranty terms and returns on a part ordered for the wrong side come from the policy pages you publish, indexed as knowledge, so the agent cannot invent a rule.
-
When a fitment question needs a person, the agent opens a ticket, a live chat or a callback and passes the vehicle, the parts discussed and the order so far, so the parts desk does not start from the beginning.
YourStoreFinds the Part From a Number, a Description or a Vehicle
ZestIQ Sales AgentOnline, replies instantly
Connect, teach, go live, measure
How the Fitment Assistant Goes Live on Magento 2
The fitment data is the part that takes care. The rest is the same connection the Sales Agent makes on any Magento 2 store.
OAuth From the Admin
Catalog Sync
Fitment Attributes Mapped
Policy Pages Indexed
Vehicle Intents Tuned
Part-Number Patterns
Policy Answers Checked
Handover Rules
Widget on Hyvä or Luma
Staging First
Guest and Account Modes
Attribution On
Fitment Confirmations
Carts From the Chat
Returns Watched
Intents Report
Accuracy
What the Fitment Assistant Reads, and What It Never Invents
3 tiers of part searchDays to live on the storefront
Prices, stock, options and fitment are read from Magento at the moment of the answer, never from the model's memory, so a price change or a sold-out part shows in the next reply. Search runs in tiers: an exact part number, then a prefix match for the worn or partial numbers shoppers type, then semantic search over the attributes you chose, so the cheap, certain steps run first and the model is used only where it adds something.
If your fitment data does not say a part fits the stated vehicle, the agent says so and offers the parts desk. It does not infer a fit from a product title, a similar SKU or a vehicle one trim level away, and it does not recommend around a gap in the data. The gaps it meets are reported, which is often the first complete list of missing fitment a store has seen.
Customer emails, phone numbers, card numbers and identifiers are masked before anything is stored, prompt-injection attempts are blocked before they reach the model, and every security event is logged. The AI provider is your choice, OpenAI, Anthropic or a self-hosted model, switchable per workspace without a redeployment, which also settles where your data lives.
Verified reviews
What Our Clients Say
2.1s
faster largest contentful paint
This team consistently goes above and beyond in terms of helping to ascertain the most viable and feasible outputs and outcomes for our evergrowing business needs.
Anthony Perrotta
E-Commerce Operations Manager
4 yrs
as their sole technical partner
They designed for where the product was going, not just what we asked for. Early on we described a single-company setup; they built the data model and permission layer so multi-tenancy wouldn’t require a rewrite later.
Alex Tee
CTO
What it covers
What the Auto Parts Fitment Chatbot Covers
Vehicle Captured Once
Year, make, model and the attributes your fitment uses, asked once and kept for the whole conversation.
Fitment-Filtered Search
Every search and recommendation filtered to parts that fit, with the fit confirmed in the answer.
Part Number and SKU Lookup
Exact and prefix matching, so a worn or partial number still finds the part.
Live Stock and Price
Read from Magento at the moment of the answer, including customer-group prices and store views.
Add to Cart and Order Tracking
Validated add to cart from the chat, and order status for the verified owner only.
Handover to the Parts Desk
Ticket, live chat or callback with the vehicle, the parts discussed and the order attached.
Trusted by 100+ companies
Keep reading
Reading Before the Demo
FAQ
Questions, Answered
Still have questions?
Tell us about your store and we will come back with straight answers.
Book a Demo
Does the auto parts chatbot work with our ACES fitment data?
It works with the fitment data your Magento store already holds, whether it was loaded from an ACES file by your catalog tool or kept as product attributes and a fitment table. During setup we map the attributes that define a fit, year, make, model, engine, trim or whatever your data uses, and the agent filters on them. It does not need its own vehicle database subscription; it reads yours.
Does it replace our year, make and model selector?
It works alongside it. The selector still filters the catalog for shoppers who browse; the assistant catches the ones the selector misses: a shopper who lands from search on a product page, one who asks a question the dropdowns cannot take, and one on a phone who would rather type the truck than work five menus. The vehicle chosen in either place carries into the other.
What happens when a part has no fitment data?
The agent says the fit is not confirmed and offers the parts desk, with the vehicle and the part attached. It does not infer a fit from a product title or a similar SKU, because a wrong part costs you the margin twice. The gaps it finds are reported, which is often the first complete list of missing fitment data a store has seen.
Can a shopper give a VIN instead of the vehicle?
The agent works from the vehicle: year, make, model and the attributes your fitment data uses. VIN decoding is not part of the agent. If your store already decodes a VIN to a vehicle elsewhere on the page, that vehicle is what the agent filters on.
Does it work on Hyvä and on Luma?
Both. The widget loads on Hyvä and Luma storefronts without touching the theme, in guest and logged-in modes, and the conversation persists across pages. On a PWA storefront the widget is placed in the storefront code, which we scope at the demo.
How long does it take to go live?
Days from the OAuth connection to the widget on the storefront for a store whose fitment data is already in Magento. The time goes into mapping the fitment attributes, tuning the vehicle intents to how your customers name their trucks and engines, and checking the policy answers against your pages on staging before anything is live.
How is the fitment chatbot priced?
Per store, agreed after the demo on your own catalog, because the setup depends on how your fitment data is held and how many storefronts the agent serves. There is no price list on this site; the quote comes with the demo.
Will it replace our parts desk?
No. It answers the questions the parts desk answers all day so the desk handles the ones that need a person: the vehicle nobody has data for, the custom build, the freight claim. Every handover arrives with the vehicle, the parts discussed and the order attached, so the desk does not start from the beginning.
How does it handle returns, cores and freight?
From the policy pages you publish. Core charges, freight collection on heavy parts, warranty terms and the rule for a part ordered for the wrong side are indexed as knowledge, so the agent quotes your policy rather than a general one. If a page is missing, the agent says it does not know and hands over.
What will the assistant refuse to do?
Invent a fit, a price, a stock level or a policy. Everything it says about a part is read from Magento at the moment of the answer, and everything it says about rules comes from your pages. When the data is silent, it says so. That is the difference between a fitment assistant and a chatbot.
Which AI provider does it use, and where does our data go?
Your choice of OpenAI, Anthropic or a self-hosted model, switchable per workspace without a redeployment, which also settles where the data lives. Customer emails, phone numbers, card numbers and identifiers are masked before anything is stored, and prompt-injection attempts are blocked before they reach the model.
Can we see it on our own catalog before we commit?
Yes, that is the demo. We connect the agent to a staging copy of your catalog and fitment data and ask it the questions your parts desk hears most. You see the answers, the fitment confirmations and the handovers before anything is live.
Let’s build something great together!
Ready to Stop Shipping the Wrong Part?
Book a demo. We connect the agent to a staging copy of your catalog and fitment data and show it answering your customers' real questions, vehicle first.