Price Lab 2.0Benchmark ReportNovember 21, 2026

AI Price Comparison Test: AI Price Search vs ShopSavvy — November 2026

10 products tested

A transparent ten-product comparison separating verified pre-tax delivered totals from advertised prices, stale offers, and wrong variants.

Price and shipping evidence

Verified prices use independently corroborated item and delivery evidence. “Free Shipping Confirmed” requires verified shipping evidence. Offers labeled “Shipping Unverified” or “Shipping Not Confirmed” retain their advertised item price but cannot qualify as a verified delivered Best Price.

Overall Results
1stNo universal winner
2ndResults varied by product and evidence quality

Competitors Tested

ShopSavvy

What AI Price Search Got Right

  • Strict product identity
  • Verified shipping requirements
  • Wrong-generation and bundle exclusions
  • Ring not-Plus correction verified live

What AI Price Search Got Wrong

  • Seller-specific marketplace coverage
  • Partial provider coverage
  • Some destination-shipping verification gaps
Run This Search Yourself

Verify these results on AI Price Search:

AI Price Comparison Test: AI Price Search vs ShopSavvy — November 2026

Shopping comparison tools can show very different prices for the same search. The hard part is deciding whether the offers are actually the same product, still available, new rather than refurbished, and deliverable to the shopper at the displayed price.

On November 21, 2026, we ran a ten-product AI price comparison between AI Price Search (APS) and ShopSavvy for ZIP 71941 in the United States. We checked exact model identity, generation, condition, availability, price freshness, and shipping evidence. This was a point-in-time test, not a claim that either service always finds the lowest price.

Testing independence disclosure: AI Price Search conducted this comparison and is evaluating its own service. ShopSavvy did not sponsor, review, or approve the test. We report the evidence that favored ShopSavvy as well as the evidence that favored APS, including APS failures and coverage limitations.

Affiliate disclosure: AI Price Search may earn a commission when a shopper follows certain retailer links and completes a purchase, at no additional cost to the shopper. Affiliate relationships do not determine APS ranking or Best Price eligibility. ShopSavvy source links are included for evidence and comparison.

If you want to price compare a product yourself, visit AI Price Search or start a shopping search. You can also browse The Price Lab for more transparent comparison tests.

Advertised price vs. verified delivered total

The two price types in this report are not interchangeable:

  • Advertised price is the item price shown in a search result or retailer listing. Shipping, availability, condition, seller, or destination eligibility may still be unconfirmed.
  • Verified pre-tax delivered total means the current item price plus confirmed shipping to the test destination, excluding sales tax. APS also requires the requested product identity and availability to pass its verification checks before the offer can become Best Price.
  • No Best Price means APS did not have enough evidence to name a qualifying delivered offer. It does not mean no retailer advertised the product.

A lower advertised price can be real and still remain ineligible for Best Price when its delivered total cannot be independently verified.

Ten-product comparison

| Product tested | APS result | ShopSavvy finding investigated | What the evidence showed | |---|---|---|---| | Starlink Mini Kit | Amazon $449.99 verified pre-tax delivered total | Home Depot $199 advertised; West Marine was reported at $199 | Home Depot matched the exact kit, but delivery to ZIP 71941 was not independently verified. West Marine showed $249 and out of stock during the investigation; the reported $199 price was stale or conditional. APS kept the higher verified offer, but its coverage of the cheaper Home Depot offer remains a genuine limitation. | | Nintendo Switch 2 — console only, no bundle | No verified Best Price | No current exact console-only ShopSavvy price was confirmed | APS correctly excluded Mario Kart and other bundles from a console-only query. Target and Macy's advertised $499.99 listings were discovered, but delivered totals were not verified. | | Sony WH-1000XM5 Black | No verified Best Price | Best Buy $169.99 advertised; Sam's Club $149.91 reported earlier | Best Buy appeared to be the exact new black model, but destination shipping was not verified. The Sam's Club price was older and stale. ShopSavvy also placed a $105 refurbished Groupon offer first, but its destination URL identified a different Sony WH-CH720N model, so it was not a valid XM5 comparison. APS's inability to promote the current exact Best Buy listing remains a genuine verification and coverage limitation. | | Logitech MX Master 3S Black with Logi Bolt receiver | Amazon $75.79 verified pre-tax delivered total on the precise model/receiver retest | Amazon $79 new but out of stock, checked November 20; Best Buy $89.99 new and in stock, last checked November 2 | The corrected competitor evidence is $79 at Amazon only as an out-of-stock new offer; the next observed in-stock new ShopSavvy offer was Best Buy at $89.99. APS qualified $75.79 only when the query pinned model 910-006556 and the receiver. A broader receiver query did not reproduce a Best Price after cached enrichment, so this result is query-sensitive and should not be generalized. | | DualSense Edge White | Amazon $169 verified pre-tax delivered total on a fresh equivalent query; an identical repeated MCP call returned no Best Price | Amazon marketplace $125.99 advertised | ShopSavvy's lower listing used the exact product ASIN and was sold by JRS GAMES, not Amazon itself. The investigation did not independently confirm complete accessories, destination shipping, and delivered total. This is a genuine APS discovery gap, and the difference between cached and fresh APS runs must be disclosed. | | AirPods Pro 3 | Amazon $179 verified pre-tax delivered total | VIPOUTLET/Walmart $132 advertised | The $132 result was stale and was not selectable on the current retailer page. The live Walmart offer was $179, new, with free shipping, sold by Walmart. Other merged catalog results included AirPods 4 and refurbished items, which were wrong for this query. | | Samsung 990 Pro 2TB — no heatsink | Amazon $389.99 verified pre-tax delivered total | Amazon marketplace $129.30 advertised, last checked November 15 | The lower listing used the exact no-heatsink ASIN, but it came from third-party seller offers that were not independently reconfirmed on November 21. ShopSavvy also mixed in a 1TB drive and a heatsink variant. APS currently samples Amazon search results rather than enumerating every seller offer, so this is a real seller-coverage limitation, not a reason to weaken identity checks. | | Shokz OpenRun Pro 2 Black | Amazon $124.95 verified pre-tax delivered total on a fresh equivalent query | Sam's Club $79.91 advertised | The cheaper Sam's Club listing was the original OpenRun Pro, not the OpenRun Pro 2. APS correctly rejected the wrong generation. An identical repeated MCP call returned cached output, while the fresh equivalent query returned the qualifying offer. | | Ring Battery Doorbell 2K base model | Amazon $39.99 verified pre-tax delivered total after the v60 production retest | $39.99 advertised for the base model | The original fresh retest exposed a genuine negation-parsing defect: “not Plus” was treated as a positive Plus request and could select the $89.99 sibling. APS v60 corrected the query interpretation and final Best Price defense. The live uncached retest selected the exact $39.99 Speckled White base model with verified destination and confirmed free shipping; no Plus listing qualified. | | Roborock original Qrevo | No exact qualifying offer | No exact qualifying offer | Neither service produced a confirmed exact original Qrevo offer. A ShopSavvy match was an accessory rather than the robot vacuum. |

What APS got right

APS's strongest results came from conservative product identity checks:

  • A Nintendo Switch 2 bundle did not win a console-only search.
  • The original Shokz OpenRun Pro did not win an OpenRun Pro 2 search.
  • AirPods 4 and refurbished AirPods did not replace new AirPods Pro 3.
  • A Roborock accessory did not count as the original Qrevo vacuum.
  • Stale or unavailable prices were not promoted as verified Best Prices.

These distinctions matter whenever shoppers compare prices online. A lower number is not useful if it belongs to the wrong generation, a refurbished product, an accessory, or an unavailable listing.

Where ShopSavvy exposed APS limitations

This test also found legitimate APS weaknesses:

DualSense Edge White

ShopSavvy surfaced a $125.99 Amazon marketplace listing tied to the exact ASIN. APS did not discover it in the comparable search. Because the seller was JRS GAMES and the investigation could not independently verify the full accessory set and delivered total to ZIP 71941, the price remains an advertised offer rather than a verified delivered total. Still, missing the seller-specific listing is a real coverage gap.

APS also showed run variation: a repeated identical MCP request returned cached output with no Best Price, while a fresh semantically equivalent query returned Amazon at $169. This appears to be connector or tool-result caching outside the APS ranking engine, but it affects reproducibility and must be disclosed.

Starlink Mini Kit

Home Depot's $199 listing matched the requested kit and was materially cheaper than APS's $449.99 verified Amazon result. APS could not independently verify delivery to the test ZIP, so it correctly refused to promote the offer under its current rules. The limitation is not the ranking decision; it is the lack of verified destination-shipping evidence for a genuine cheaper listing.

Sony WH-1000XM5 Black

Best Buy advertised the exact black WH-1000XM5 at $169.99. APS could identify the listing but could not verify enough destination-specific delivery evidence to name it Best Price. The older Sam's Club $149.91 result was stale. APS was appropriately cautious, but a shopper still needs to know that a lower current advertised price was found.

Samsung marketplace seller coverage

ShopSavvy found exact-ASIN third-party Amazon offers far below APS's verified result. Those seller-specific prices were not freshly reconfirmed, but APS also did not enumerate them. APS presently examines a limited Amazon search-result set rather than every seller offer on an ASIN. This is a meaningful coverage limitation.

Why comparison results disagree

Stale prices

A comparison service may retain the last observed price after a retailer changes or removes it. In this test, the VIPOUTLET/Walmart AirPods price and older Sam's Club Sony price could not be selected at the reported amount during the November 21 investigation.

Wrong generations and nearby models

Retailer catalogs often group products with similar names. The original OpenRun Pro is not the OpenRun Pro 2; AirPods 4 are not AirPods Pro 3; Sony WH-CH720N headphones are not WH-1000XM5 headphones; and a 1TB SSD is not the requested 2TB model. APS intentionally keeps these offers out of Best Price.

Refurbished products

A refurbished price can be legitimate, but it should not silently compete with a query for a new product. Some lower AirPods results were refurbished and were therefore not valid new-product comparisons.

Seller-specific offers

A marketplace product page can contain multiple sellers with different prices, availability, accessories, returns, and shipping. ShopSavvy surfaced seller-specific Amazon offers that APS's current provider path did not enumerate. Those prices are useful leads, but they need current seller and delivery verification before being described as delivered totals.

Prepublication retest evidence

A production-equivalent retest used identity-preserving queries at ZIP 71941. Every result below was correlated with a new SearchLog record and a production mcpSearchProducts invocation. All searches remained partial because at least one provider failed or timed out.

| Product | Fresh APS observation (UTC) | Fresh result | Evidence note | |---|---|---|---| | Starlink Mini Kit | 2026-11-21 21:29:05 | Amazon $449.99 verified pre-tax delivered total | Home Depot's exact kit remained $199 advertised; destination shipping was not independently confirmed. | | Nintendo Switch 2 console only | 2026-11-21 21:29:06 | No Best Price | Bundles remained excluded. | | Sony WH-1000XM5 Black | 2026-11-21 21:29:06 | No Best Price | Best Buy $169.99 remained advertised/current, but shipping and destination evidence did not qualify. | | Logitech MX Master 3S Black with receiver | 2026-11-21 21:31:52 | Amazon $75.79 verified pre-tax delivered total | Precise MPN/receiver query; the no-receiver $75.99 variant was explicitly rejected. | | DualSense Edge White | 2026-11-21 21:29:15 | Amazon $169 verified pre-tax delivered total | ShopSavvy still showed the exact-ASIN JRS GAMES offer at $125.99, checked 2026-11-21 18:51 UTC, without independently verified delivered total. | | AirPods Pro 3 new | 2026-11-21 21:29:05 | Amazon $179 verified pre-tax delivered total | Wrong generations remained excluded. | | Samsung 990 Pro 2TB, no heatsink | 2026-11-21 21:29:08 | Amazon $389.99 verified pre-tax delivered total | ShopSavvy's $129.30 exact-ASIN seller offers were last checked November 15, not freshly reconfirmed. | | Shokz OpenRun Pro 2 Black, standard size | 2026-11-21 21:32:16 | Amazon $124.95 verified pre-tax delivered total | Fresh uncached query; wrong variants remained excluded. | | Ring Battery Doorbell 2K base model | 2026-11-21 21:54:03 | Amazon $39.99 verified pre-tax delivered total | Live v60 retest: the destination-verified Speckled White base model qualified with confirmed free shipping, and no Plus listing qualified. Search coverage remained partial. | | Roborock original Qrevo | 2026-11-21 21:29:50 | No exact qualifying offer | Accessories and other Qrevo generations remained excluded. |

The backend SearchLog records show cache_hit:false for these final evidence runs. Production logs recorded corresponding partial-search invocations between 21:29 and 21:32 UTC. Repeating the same queries afterward returned cached results, as expected, and those repeats were not used as the fresh evidence timestamps above.

Methodology

  1. AI Price Search conducted this test and evaluated its own service.
  2. We used ten fixed product identities and ZIP 71941.
  3. The intended condition was new unless stated otherwise.
  4. We preserved manufacturer, model, generation, storage or capacity, color, receiver/package contents, and bundle requirements where relevant.
  5. APS and ShopSavvy results were traced to retailer or marketplace evidence.
  6. We checked exact identity, condition, price freshness, availability, seller, and shipping evidence.
  7. APS Best Price used the item price plus confirmed shipping, excluding sales tax. Advertised-only offers remained visible in the analysis but did not become verified winners.
  8. Prices and availability are point-in-time observations from November 21, 2026, unless an older source date is explicitly stated.
  9. All ten APS MCP searches reported partial provider coverage. SerpAPI shopping and Amazon discovery generally worked; CJ returned candidates; Awin often had no matching catalog result; some AliExpress or LLM-assisted calls were empty or timed out. Amazon PA-API was unavailable because of account eligibility and was not part of this investigation.

We did not change APS's ranking engine to force lower prices. Wrong variants, bundles, accessories, used or refurbished items, and offers without verified delivery remained protected by the existing rules.

Bottom line

This comparison did not produce a simple universal winner. APS was strongest when strict identity and delivered-price verification prevented misleading low prices. ShopSavvy was stronger at surfacing some seller-specific or advertised offers, particularly DualSense Edge and Samsung 990 Pro. The fresh Ring production retest first exposed a negative-variant bug, and the deployed v60 correction then passed an uncached live check: the exact $39.99 base model qualified and no Plus listing entered the qualifying set. Coverage remained partial, so this does not imply complete market coverage.

For shoppers, the practical lesson is to separate discovery from verification. Use an AI shopping comparison to find candidates, but confirm the exact model, seller, condition, availability, and delivered total before buying.

Compare prices with AI Price Search or browse more Price Lab research.

Sources checked

Retailer and comparison pages reviewed during the investigation included:

Retest Status

APS v60 deployed. Ring base-model regression passed live at ZIP 71941; full benchmark evidence remains point-in-time and provider coverage was partial.

Methodology

Data window: November 21, 2026. Analysis based on 10 searches run on aipricesearch.com. Prices are point-in-time snapshots and may have changed since capture. Free to cite with a link to this page.