How to Let Car Shoppers Search Your Inventory in Plain English

Q: Why does dropdown-and-filter inventory search lose car shoppers before they reach a VDP?

Most shoppers arrive knowing what they need, not what specs to enter. Someone looking for a reliable three-row SUV under $35k has to translate that into make, model, year, and trim before filter-based search returns anything. When they guess wrong, they get zero results and leave. Dropdown filters were built for buyers who already know the exact vehicle; they fail shoppers who are still deciding, which is most traffic arriving from Google or social ads.

Q: How do I connect Nobi to my dealership's live inventory feed?

You point Nobi at the same feed URL or file format your DMS already sends to Cars.com or Autotrader: no new data pipeline required. Nobi indexes the VDP attributes it finds (make, model, year, trim, price, mileage, body style, condition, feature data) and refreshes twice daily automatically, so a vehicle that sells in the morning is out of search results by the afternoon cycle. Adding Nobi to your existing website is a snippet your developer drops in; there is no platform migration and no changes to your existing templates.

Q: How does Nobi handle budget, body-style, and feature queries that dropdown filters miss?

Nobi parses multi-condition queries as a single intent and applies all conditions simultaneously against your VDP data. Budget phrasing like "around $30k" or "between $25k and $35k" resolves to a price-range filter without requiring the shopper to find a slider, and body-style synonyms like "crossover," "SUV," and "CUV" all resolve to the same segment. Feature queries match against the options and packages data in your inventory feed, and when a feature field is not populated in your DMS export, Nobi surfaces that query in its zero-result tracking so your team knows exactly which tagging gaps to close.

Q: How do I track what shoppers searched for but couldn't find, and what do I do with that data?

Nobi surfaces zero-result searches as a running list in your analytics view, showing every query that returned nothing from your inventory. Those queries split into two categories: a genuine inventory gap, where you don't stock the vehicle, and a tagging gap, where you have the vehicle but the attribute field doesn't match the words the shopper used. A tagging gap (for example, "blind spot monitoring" returning nothing because your DMS exports it as "BSM package") is the faster fix and doesn't require buying new inventory. After-hours queries are especially useful because they show demand from shoppers who were ready to engage and got nothing back.

Can shoppers search a dealership's inventory in plain English instead of dropdown filters?

A shopper who types "reliable family SUV under $35k" is describing exactly what they want - just not in the make, model, year, and trim language a dropdown search expects. Natural-language search reads that plain-English description and maps it to matching vehicles on your lot, with no dropdown filters at all, so a shopper who can't translate their mental picture into the right boxes still lands on the right cars instead of a zero-results page. You connect your inventory feed, drop a snippet into your site, and shoppers can search by description the same day.

Why does dropdown-and-filter inventory search lose car shoppers before they reach a VDP?

Platforms like Dealer.com build inventory search around structured dropdowns because the fields map cleanly to the DMS. But a shopper who types "third-row SUV under 30k miles" doesn't know to select body style and year range separately. Rigid filter logic returns nothing for that query, and every zero-result page is a lost VDP visit.

What is Nobi, and how does it give dealership shoppers a plain-English way to find vehicles on your lot?

Nobi is a conversational website assistant that connects to your actual inventory and the content you already publish. When a shopper types "used F-150 under $40k with leather" into your website's search bar, Nobi pulls matching vehicles from your live inventory feed: a ranked shortlist tied to the actual VDP attributes your DMS is already exporting, not a response generated by a model that's never seen your stock.

The routing is automatic. Short keyword searches - year, make, model - take a fast instant-search path and return results as the shopper types. Longer, need-based queries ("third-row SUV under 30k miles") get semantic matching against VDP attributes, so the shopper's words don't have to match your listing titles for the right vehicles to surface.

Dealers who've heard the Chevy Watsonville story (a ChatGPT-powered bot that agreed to sell a Tahoe for $1) sometimes conclude AI has no place in inventory search. The real lesson is narrower: a general-purpose model with no guardrails and no connection to your lot is what caused that failure. A system grounded in structured VDP data and your approved content has different constraints, though no AI system eliminates the possibility of errors entirely, and the quality of your inventory data directly shapes what it can get right.

Nobi also handles the questions shoppers ask after they find a vehicle they like. Warranty coverage, financing options, the trade-in process: Nobi answers from the content you've already published, with an inline citation on every answer showing the exact source.

There's no search engineer required, no synonym lists to maintain, and no ranking rules to tune by hand after setup. Nobi connects to your existing inventory feed with a small site-side install. Pricing starts at $25/month, which includes 2,500 searches and 250 messages; beyond that, $0.01 per additional search and $0.10 per additional message.

How does Nobi translate a shopper's plain-English description into matching vehicles from your inventory?

Nobi uses semantic matching to read the intent behind a query, not just the exact words. When a shopper types "family SUV with third row under 30k miles," Nobi maps "family SUV" to body-style and segment attributes in your VDP data, "third row" to a seating or feature field, and "under 30k miles" to the mileage attribute, then returns vehicles from your live inventory that satisfy all three conditions. The match runs against the structured VDP fields your DMS already tracks (make, model, trim, body style, price, mileage, feature packages, condition), so every result is grounded in data you control.

Synonym resolution is built in. "Third row," "three-row seating," and "seven-passenger SUV" all resolve to the same attribute in your feed. The shopper doesn't need to know your listing terminology to find the right vehicle.

Multi-condition queries run as an AND intersection. A shopper who types "crossover with AWD under $35k and under 20k miles" applies four filters at once. Nobi returns a shortlist of vehicles that clear every condition. Results rank by relevance, so the vehicle that best matches the stated criteria appears first.

Short spec-based searches work differently. A query like "2022 Camry" takes a fast instant-search path and returns results as the shopper types, without the extra processing step a full conversational query requires. Both paths draw from the same live inventory index built from the VDP attributes your DMS already exports, so every result ties back to a real vehicle with real specs.

How do I connect Nobi to my dealership's live inventory feed?

Connecting that feed takes the same format your DMS already sends to Cars.com or Autotrader: a feed URL or file upload. You point Nobi at it, and it indexes the VDP attributes it finds: make, model, year, trim, price, mileage, body style, condition, and available feature data. The index refreshes twice daily automatically, so lot changes reach search results within hours.

No new data pipeline required. Nobi reads the standard inventory feed formats dealers already produce for third-party listing sites. If the export goes to Autotrader today, it works with Nobi tomorrow. You don't go back to your DMS vendor for a new configuration.

The twice-daily refresh has a practical upside your BDC will notice: a vehicle that sells in the morning is gone from search results by the afternoon cycle. Shoppers stop calling about cars you already delivered. And a vehicle that comes in on trade gets indexed at the next refresh without anyone touching a setting.

Dealers on Dealer.com, DealerOn, or any other website platform can add Nobi as an overlay. Your developer drops in the Nobi snippet and points it at the slot where the search bar should appear. No platform migration, no touching existing templates. Cox Automotive Dealer.com customers use the same path.

Implementation runs in hours, not a quarter. You connect the feed, verify that the attribute mapping looks right, and search is live.

Once the feed is connected, Nobi's query override feature lets you lock exact answers to high-stakes questions: return policy, CPO terms, warranty coverage. Any question you pin gets the exact answer your team approved, word for word, every time a shopper asks it. The rest of the query flow stays AI-driven; overrides are for the handful of questions where you can't afford variation.

How does Nobi handle budget, body-style, and feature queries that dropdown filters miss?

Budget, body-style, and feature queries are exactly where dropdown filters break down. A shopper who types "under $40k" gets a price-range filter applied against your inventory; "third row" maps to a seating or feature field; "leather seats" maps to an options or packages field. Nobi parses multi-condition queries as a single intent, so "crossover with AWD and heated seats under $35k" resolves to four simultaneous filters against your VDP data in one pass.

Budget phrasing doesn't need to be exact. "Around $30k," "under $40k," and "between $25k and $35k" all resolve to price-range filters without requiring a shopper to find the price slider. Body-style synonyms are handled the same way: "crossover," "SUV," and "CUV" resolve to the same segment, and "pickup," "truck," and "work truck" resolve to the same body type.

Feature queries match against the options and packages data in your inventory feed ("heated seats," "sunroof," "backup camera," "blind spot monitoring") to the extent those fields are populated in your DMS export. If a feature field isn't filled in, Nobi can't match against it, and the query comes back empty. Those zero-result queries appear in Nobi's zero-result tracking, which gives your team a concrete list of tagging gaps to close in the DMS. A zero result for "heated seats" isn't a search failure; it's a signal that the attribute isn't in the feed.

On warranty and vehicle-specific questions that come up after a shopper finds a match, Nobi can run a second AI review of every draft answer against the source content before it sends. A question about CPO coverage or financing terms returns what your published content actually says, not a paraphrase that could misstate a term.

How do I track what shoppers searched for but couldn't find, and what do I do with that data?

Nobi surfaces zero-result searches (the queries shoppers typed that returned nothing from your inventory) as a running list in your analytics view. Each query is a data point: either you don't have that vehicle on the lot, or you have it but it isn't tagged with the attribute the shopper used. Reviewing that list weekly gives your team a demand signal for purchasing decisions and a tagging signal for your DMS team to clean up.

Zero-result searches split into two categories, and the fix is different for each. The first is a genuine inventory gap: you don't stock that vehicle. Repeated searches for "hybrid SUV under $35k" across multiple sessions over several weeks is a conversation worth having with your GSM about days-to-turn on that segment. The data doesn't tell you what to buy, but it tells you what shoppers wanted before they left for a competitor's lot.

The second category is a tagging gap, and it's the faster fix. If "blind spot monitoring" returns zero results because your DMS exports that feature as "BSM package," a synonym mapping or a field rename on the feed side resolves it without buying new inventory. The vehicle is sitting on your lot; the shopper just couldn't find it because the attribute field doesn't match the words they used.

After-hours search behavior is especially useful here. Queries that come in at 10pm, when your BDC is closed, show you what shoppers wanted before they moved on.

Nobi's analytics go beyond raw zero-result counts. You can see which searches produced VDP clicks and which searches converted. You're not just looking at what failed; you're looking at the full funnel from query to visit to action.

The practical rhythm is a weekly review of zero-result queries, sorted by frequency. High-frequency misses with inventory on the lot point to the DMS. High-frequency misses with no matching inventory go to the GSM. Low-frequency misses, especially from odd hours, are the after-hours demand signals worth noting but not acting on immediately. That's a 20-minute weekly task that pays off in tighter tagging and smarter purchasing, not a reporting project.

When is natural-language inventory search the right call for your dealership, and when should you look at other options?

Natural-language inventory search is the right call when your current search bar is losing shoppers who think in needs rather than specs; measure this by pulling your SRP zero-result rate and your search-to-VDP conversion. Nobi fits when you want search and shopper Q&A in one product, want to go live in hours rather than quarters, and don't need a developer team to own and maintain ranking logic over time.

If you're already all-in on the Cox Automotive stack (Dealer.com website, VinSolutions CRM, Dealertrack DMS), Dealer.com's built-in search keeps your vendor relationship simple. Changing search behavior outside that stack typically requires Cox-managed services or a premium tier.

If your IT team wants full API control over ranking (custom scoring for days-on-lot, margin target, or conquest model weighting), Algolia gives engineers that control. It requires dedicated engineering resources, and NeuralSearch, the semantic layer, is available only on Algolia's top-tier Elevate plan.

Two Nobi limits worth knowing: the assistant won't execute transactions in chat (no cancellations, no real-time order lookups), and there's no mode where a BDC rep joins an active AI conversation while the AI continues alongside them.

For most single-rooftop and small franchise-group dealers whose core problem is shoppers bouncing before reaching a VDP, Nobi's $25/month base (2,500 searches and 250 messages included; $0.01 per additional search, $0.10 per additional message) and hours-to-live setup is the fastest path from filter bounce to VDP visit.

---

If your SRP zero-result rate is the problem and you want search live in hours rather than a quarter, connect your inventory feed at dashboard.nobi.ai.

<div className="my-8 flex justify-center"> <a href="https://dashboard.nobi.ai" className="inline-flex items-center justify-center gap-2 rounded-2xl font-medium transition active:scale-[.98] focus:outline-none focus-visible:ring-2 focus-visible:ring-black/10 dark:focus-visible:ring-white/20 bg-black text-white dark:bg-white dark:text-black hover:opacity-90 shadow-sm h-12 px-6 text-base no-underline" > <span>Start your free Nobi trial</span> </a> </div>