Snipp Interactive today opened a ChatGPT plugin and a Claude connector that let shoppers find and enroll in brand offers without leaving an AI conversation.
In Short
A company that runs the behind-the-scenes machinery for coupons, cashback and rewards programs has connected that machinery to two AI chatbots. If you ask ChatGPT or Claude about a product, brands using this system can now have their rebate or loyalty offer appear in the answer, and you can sign up without visiting a website or downloading an app. The change matters because it moves offer discovery away from search results and brand sites, where marketers have spent two decades learning to control it.
What shipped
The Vancouver company, listed on the TSX Venture Exchange as SPN and quoted on OTC Pink as SNIPF, put out the news at 07:00 on Monday, September 14, 2026, through ACCESS Newswire. Two products carry the same underlying service: the $nipp ChatGPT Plugin and the $nipp Connector for Claude. Both were cleared by OpenAI and Anthropic under the verification processes each company applies to third-party connections, according to Snipp.
Availability is immediate. ChatGPT users install from the plugin directory at chatgpt.com/plugins. Claude users add the connector from the directory at claude.ai/directory/snipp. Snipp states that both routes resolve to the same promotions and loyalty platform, producing consistent discovery, purchase validation and enrollment across the two assistants.
The commercial framing is unusual for a channel launch. Snipp lists no additional platform fees, no new marketplace commissions and no extra platform costs as a property of the integration. That places the offering outside the take-rate model that has defined most commerce plumbing attached to conversational interfaces so far.
The stack behind the connector
What the connector exposes is not new. Snipp has spent more than two decades building receipt processing for purchase validation, receipt intelligence, basket and SKU-level data capture, and fraud prevention across hundreds of retailers and points of sale, according to the company. That infrastructure was built to run promotions, rebates and loyalty programs. The connector reroutes it.
The published capability list runs to eight items: receipt OCR parsing, SKU-level validation, fraud risk scoring, offer rules engines, reward fulfillment automation, identity resolution, campaign analytics, and compliance audit logging. Brands keep managing campaigns inside the Snipp platform. Snipp states that it handles AI discoverability, purchase validation, enrollment and real-time synchronization behind the scenes.
Atul Sabharwal, chief executive and founder of Snipp Interactive, framed the move as an extension rather than a new product line. "AI is rapidly becoming the front door to commerce, with consumers increasingly turning to AI to discover products, compare options, and find the best value," he said. He went further on positioning: "We aim to be the 'intelligence within the intelligence' when it comes to building valued and loyal connections within an AI surface for our brands."
Christopher Cubba, chief revenue officer, put the pitch in operational terms. "Our goal is to help brands participate in this emerging engagement channel using the same trusted promotions and loyalty infrastructure they rely on today, without adding complexity or operational overhead," he said.
A protocol layer for loyalty data
Alongside the two integrations, Snipp says it is releasing the first Model Context Protocol server built for shopper and loyalty marketing. The claim is category-specific rather than general. MCP servers have been arriving across advertising technology since Anthropic published the standard in November 2024, and the pattern is well documented: Google released an open-source server for its Ads API in October 2025, Amazon Ads opened a closed beta in November 2025, and Meta opened its ad system to Claude and ChatGPT in April 2026. Semrush published a native Claude connector covering its search datasets in August 2026. What has been missing from that list is the promotions and loyalty layer.
The Claude side of the launch sits in a directory whose model dates to July 29, 2025, when HubSpot published the first customer relationship management connector for Anthropic's assistant. Contentsquare described itself as among the first experience analytics vendors in that directory when it connected behavioral data to AI agents in June 2026.
Snipp also names Alpic as a technology provider working with it on further ChatGPT plugin capability. No timeline, contract value or scope accompanies that reference.
Where the numbers come from
The announcement carries several consumer statistics, and their sourcing is uneven.
Snipp states that more than 42% of consumers have used an AI tool to shop within the past month, and that AI-powered shopping assistants could influence 385 billion dollars in United States e-commerce sales by 2030. The second figure originates with Morgan Stanley Research, which published it in late 2025 as the upper bound of a range running from 190 billion dollars, corresponding to a 10% to 20% share of United States e-commerce. The release cites the top of the range without the floor. Morgan Stanley's own survey work at the time put roughly 23% of Americans as having made a purchase using AI in the preceding month, roughly half the figure Snipp cites for AI shopping usage.
That spread is not unique to this release. Measured adoption of AI in shopping lands across a very wide band depending on how the question is asked, a dispersion PPC Land examined when NIQ reported on social commerce and AI usage in late August 2026. Equativ research put AI product discovery at 60% across seven markets in October 2025. Retail Economics fieldwork across three countries reached 73%. RTB House found AI lengthening purchase decisions for 42% of United States shoppers rather than shortening them.
Three further figures appear in the consumer section of the release without attribution to any named study: 86% of shoppers saying rewards influence where and how they shop, nearly one-third citing lengthy or complicated enrollment as the biggest barrier to participation, and 28% abandoning programs that require another app. Materials circulated with the announcement add two more unsourced numbers, putting expected AI shopping usage at 65% of consumers against 8% of retailers described as ready for them.
Deepthi Andi, executive vice president of product at Snipp, described the friction those numbers are meant to illustrate. "Consumers don't think in terms of promotions, rebates, or loyalty programs," she said. "They simply want to know how to get the best value on the products they're already planning to buy."
Plugin is a current word again
The word plugin carries history that complicates a quick reading of this launch. OpenAI shut down its original plugin store in April 2024 and replaced it with GPTs, then with connectors and apps built on MCP. The vocabulary returned on July 9, 2026, when the App Directory became the Plugin Directory, organizing skills, apps and app templates. A plugin in the 2026 sense is a package rather than a single-purpose extension, and it authenticates through the same app layer that preceded the rename.
That directory is now also where OpenAI distributes its own advertising tooling. PPC Land documented setup running through the Plugins directory when the company adjusted its ChatGPT ad exclusion audiences in late August 2026. A promotions connector and an ad campaign assistant now sit in the same catalogue.
The discovery problem it targets
Conversational surfaces have been absorbing commerce functions for a year. OpenAI added instant checkout to ChatGPT with Stripe on September 29, 2025, publishing the Agentic Commerce Protocol at the same time. Skepticism followed quickly, with independent analysis arguing that intermediated buying violates retailer incentives to own the customer relationship. Walmart then supplied the first hard performance disclosure: conversion rates for products sold inside the chatbot ran three times lower than for journeys that clicked out to its own site.
Snipp's positioning sidesteps the checkout question. Discovery, eligibility and enrollment happen in the conversation; purchase validation still runs on receipts submitted afterwards, which is the mechanism the company already operates. That design does not depend on a shopper completing a transaction inside the assistant, which is precisely where the published evidence has been weakest.
Advertising has been moving into the same window on a separate track. Amazon Ads confirmed a ChatGPT advertising partnership earlier this month, with Delta Vacations among the first testers. Google extended loyalty program features to 14 countries and pushed member pricing into AI Mode and Gemini on March 25, 2026, then began surfacing loyalty membership inside bidding and reporting through the Google Ads API in August 2026. Snipp is arriving at the same destination from the promotions side rather than the media side.
The measurement gap nobody has closed
An offer surfaced inside a conversation generates no click, no session and no referral line in an analytics report. That absence is the structural problem the measurement industry has been building toward all year. NIQ and Similarweb announced a product on September 2, 2026 intended to measure shopping inside ChatGPT and Gemini from the fourth quarter, covering consumer intent and what NIQ calls agentic shelf visibility. A later reading of that partnership noted clickstream research finding that 55.9% of traffic produced by AI recommendations arrived through channels attribution models could not connect back to the original conversation.
Snipp's receipt validation offers a partial answer to that gap, since a validated receipt is a purchase record independent of any referral trail. Whether an offer discovered in a chat can be attributed to that chat is a different question, and the release does not address it.
Derrick Horner, chief executive of H1 Ventures and a strategic advisor to Snipp, supplied the outside endorsement in the release: "Snipp is building exactly the kind of intelligent, purchase-verified shopper marketing infrastructure the market has been waiting for - an autonomous platform that removes friction for shoppers, delivers measurable ROI for brands, and positions the company at the center of the agentic commerce transformation." Horner's firm is not identified as an investor or a customer in the document.
Company standing
Snipp Interactive describes itself as a value-added software company in the loyalty and promotions sector, serving Fortune 500 brands, agencies and partners across short-term and always-on programs. Headquarters are in Vancouver, with operations across the United States, Canada, Ireland, Europe and India. Malcolm Davidson is listed as interim chief financial officer and the contact for further information.
No customer names, campaign volumes, program counts or pilot results accompany the launch. No revenue expectation is attached. The release carries standard forward-looking statement language and the TSX Venture Exchange disclaimer, and the roadmap section describes further AI platforms and predictive analytics without dates.
Why this matters for the marketing community
For brands running promotions and rebates, the operational question is where offer discovery happens once a growing share of product research moves into an assistant. Loyalty enrollment has historically been a funnel with a website or an app at the end of it, and both of those endpoints are now optional in a conversational flow. Contentsquare research covered by PPC Land in late August found 23% of AI-referred shoppers going to a rival when the destination site disappointed them, and Similarweb data published this month showed brands named by AI picking up measurable traffic, with Capital One gaining 14.2% shopper pickup when an assistant mentioned it. Removing the site visit from the enrollment path changes what those findings imply.
For retail media and shopper marketing teams, the agentic layer adds a second discovery surface that sits outside the retailer's own property and outside the retailer's measurement. Amazon has been building toward controlled integrationswith third-party assistants since spring 2026 rather than ceding the connection. A promotions vendor reaching the same assistants independently of any retailer is a different route into the same conversation.
For agencies, the near-term practical consequence is a new inventory of connector decisions. Each assistant directory is a distribution channel with its own approval process, its own permission model and, so far, no comparable performance benchmark. The evidence base for what a promotion surfaced inside a chat actually produces remains thin, and this launch adds capability rather than data.
Timeline
- November 2024: Anthropic publishes the Model Context Protocol
- July 29, 2025: HubSpot publishes the first CRM connector for Claude, establishing the Claude Connectors Directory model
- September 29, 2025: OpenAI adds instant checkout to ChatGPT with Stripe and publishes the Agentic Commerce Protocol
- October 2025: Independent analysis questions the viability of AI shopping agents
- Late 2025: Morgan Stanley Research publishes the 190 billion to 385 billion dollar range for agentic shopping in United States e-commerce by 2030
- March 18, 2026: Walmart discloses that in-chat conversion ran three times lower than click-out journeys
- March 25, 2026: Google extends loyalty program features to 14 countries and to AI Mode and Gemini
- April 29, 2026: Meta opens its ad system to Claude and ChatGPT through connectors
- July 9, 2026: OpenAI replaces the App Directory with the Plugin Directory
- August 26, 2026: Semrush publishes a native Claude connector
- September 2, 2026: NIQ and Similarweb set out a measurement product for AI shopping, live in the fourth quarter
- September 10, 2026: Amazon Ads confirms a ChatGPT advertising partnership with Delta Vacations testing first
- September 14, 2026: Snipp Interactive opens the $nipp ChatGPT Plugin, the $nipp Connector for Claude, and an MCP server for shopper and loyalty marketing
Related PPC Land coverage
- OpenAI launches instant checkout for ChatGPT with Stripe partnership - The September 2025 arrival of purchasing inside a conversational interface, and the protocol published with it.
- Walmart's ChatGPT checkout flopped. Here's what comes next. - The first hard conversion disclosure from a major retailer testing AI-mediated commerce.
- Skepticism grows over AI shopping agents as ChatGPT checkout launches - The structural argument that intermediated buying works against retailer incentives.
- Meta opens its ad system to Claude and ChatGPT with new AI connectors - The first major advertising platform to hand third-party assistants write access to campaigns.
- Semrush gains native Claude connector covering 28.8 billion keywords - A recent comparison point for how vendors enter the Claude directory.
- Contentsquare plugs behavioral data straight into Dust AI agents - Analytics data reaching agents through MCP, and the HubSpot precedent in the Claude directory.
- NIQ and Similarweb to measure ChatGPT and Gemini shopping by Q4 2026 - The measurement product aimed at the attribution gap conversational commerce creates.
- 68% of North Americans have never bought via social media, NIQ finds - Why AI shopping adoption figures land across such a wide band.
- AI lengthens purchase decisions for 42% of US shoppers, RTB House finds - Counter-evidence that AI extends rather than compresses the path to purchase.
- Bad websites lose 23% of AI-referred shoppers to rivals, Contentsquare finds - What happens after an assistant sends a shopper somewhere.
- Capital One gains 14.2% shopper pickup when AI names it, Similarweb finds - Measured lift for brands named inside AI shopping research.
- Amazon Ads gains a ChatGPT ad surface as Delta Vacations tests first - Paid placement arriving in the same conversational window.
- Amazon quietly builds an agentic commerce team to connect with ChatGPT - The retailer-side strategy for controlling which assistants connect.
- OpenAI drops 25,000-user floor for ChatGPT ad exclusion audiences - Recent context on the Plugins directory as a distribution channel for advertising tooling.
Summary
Who: Snipp Interactive Inc., a Vancouver-based loyalty and promotions technology company listed as TSXV:SPN and OTC PINK:SNIPF. Named executives in the announcement are chief executive and founder Atul Sabharwal, chief revenue officer Christopher Cubba, executive vice president of product Deepthi Andi, and interim chief financial officer Malcolm Davidson. Derrick Horner of H1 Ventures appears as a strategic advisor. Alpic is named as a technology provider for further ChatGPT plugin capability.
What: A ChatGPT plugin and a Claude connector, both cleared through the respective verification processes, that make brand promotions, rebates, rewards and loyalty programs discoverable inside AI conversations and allow enrollment without visiting a site or installing an app. A Model Context Protocol server for shopper and loyalty marketing is released alongside them. The integration exposes receipt OCR parsing, SKU-level validation, fraud risk scoring, offer rules engines, reward fulfillment automation, identity resolution, campaign analytics and compliance audit logging, with no additional platform fees or marketplace commissions.
When: Today, September 14, 2026, distributed at 07:00 through ACCESS Newswire. Both integrations are available immediately.
Where: Inside ChatGPT, through the plugin directory at chatgpt.com/plugins, and inside Claude, through the connector directory at claude.ai/directory/snipp. Snipp operates from Vancouver with a presence across the United States, Canada, Ireland, Europe and India.
Why: Product research and offer discovery are shifting into conversational interfaces, where the website and app endpoints that loyalty enrollment funnels depend on are optional. Snipp is extending purchase-verified promotions infrastructure into that surface rather than building a new one. The release carries no customer names, no pilot results and several unattributed consumer statistics, and it cites the upper bound of a Morgan Stanley range without the lower one, so the capability is documented while its effect is not.
Discussion