Coming soon: sparcy.online, a voice AI shopping assistant

sparcy.online is an AI shopping assistant that SPARCLEY will launch soon. You tell it what you need, by voice or in text, and it asks clarifying questions, compares products and services, explains the differences out loud and then carries out the purchase for you, but only after you confirm.

It is built for people who know what problem they want solved but do not want to spend an evening with dozens of browser tabs open. It is also built with online stores in mind: shops that want to be found and bought from when the shopper is an AI agent rather than a person scrolling a category page.

The idea is simple: shopping becomes a conversation instead of a search session. The assistant does the legwork, and you keep the decisions. This is conversational commerce taken one step further, from answering questions to completing the order.

• Describe the need. Say or type what you are looking for in your own words, for example "a quiet dishwasher that fits a 45 cm gap" or "someone to service my boiler next week".
• Answer a few clarifying questions. The assistant asks only what changes the recommendation: budget, size, delivery date, preferred brands, must-have features.
• Let it search and compare. It reads product data, specifications, prices, stock and delivery terms from stores and service providers, then narrows the field to a short list.
• Hear the differences explained. Instead of a wall of specs, it tells you, by voice or in text, what actually separates option A from option B and which trade-off matters for your situation.
• Choose and confirm. You pick an option or keep asking. Nothing is ordered until you have reviewed the final basket, price and delivery details and explicitly confirmed.
• Purchase and summary. After confirmation, the assistant completes the checkout steps on your behalf and gives you a clear summary of what was ordered, from whom and when it should arrive.

Most online shops are built for scanning: filters, product grids, comparison tables. That works when you already know the vocabulary of the category. When you do not, a voice shopping assistant that can explain the difference between two heat pumps or two broadband plans in plain words feels closer to talking to a knowledgeable shop assistant than to typing into a search box. In other words, an AI personal shopper you can simply talk to.

Voice also suits the moments when people actually think about purchases: in the car, while cooking, while looking at the appliance that has just stopped working. You can interrupt, ask "why that one?" or change your mind mid-sentence, and the assistant adjusts the shortlist. Text stays available for when you prefer to read, need to copy a model number or are somewhere you cannot talk.

sparcy.online is planned for both products and services. Comparing a service, such as an installation, a repair or a subscription, is often harder than comparing two products, because the important details sit in descriptions and terms rather than in tidy spec tables. That is exactly where a consulting assistant earns its place.

An AI shopping agent that can pay is only useful if people trust it. These are the design principles we are building sparcy.online around. They describe how the product is being designed; they are not certifications or compliance claims.

• Confirmation before payment. The assistant prepares the order, but the final step always needs your explicit yes. No silent purchases.
• Clear reasons. Every recommendation comes with the reasons behind it, so you can see why one option ranks above another and challenge it.
• Visible sources. You always see which store or provider each offer comes from.
• Minimal data. We aim to ask only for the information a task needs and to use it only for that task.
• Easy to stop. You can pause, change the brief or cancel at any point before you confirm.

Agentic commerce is the shift from people browsing shops to AI agents shopping on their behalf. The large platforms are already moving: OpenAI and Stripe published the Agentic Commerce Protocol for checkout inside ChatGPT, and Google introduced its Universal Commerce Protocol in January 2026. Whichever assistant a customer uses, the agent behaves the same way when it reaches your store: it reads data, not banners.

An agent cannot admire your hero image or read a product's dimensions off a photo. It needs prices, stock, variants, measurements, compatibility, delivery times and return terms in a form it can parse. If two stores sell the same item and only one of them states the specifications clearly, the agent can only recommend that one with confidence.

Being AI-agent ready is mostly good data hygiene: the same work that already improves Google Shopping listings, price-comparison sites and on-site filters. This is the checklist we use with e-commerce clients:

• Structured product data on every product page: schema.org Product markup with price, currency, availability, brand, SKU and GTIN or MPN.
• Clean, up-to-date product feeds. An XML feed in the formats Google Merchant Center and comparison portals expect is also the easiest thing for an agent to read.
• Specifications as attributes, not prose. Dimensions, materials, power, compatibility and other specs belong in fields and filters, not buried in a paragraph or an image.
• A logical category tree with consistent filters, so agents and people can narrow the catalogue the same way.
• Accurate stock and prices. If supplier stock changes several times a day, the store should too; an agent that orders an out-of-stock item will not come back.
• Plain delivery and returns terms, written as clear facts that can be quoted.

• Universal XML Feed Generator for OpenCart
• Rebuilding a category tree for 100k+ products with AI
• schema.org Product reference

sparcy.online is coming soon. We are not announcing a launch date yet; we would rather name it when the assistant is ready to be trusted with real orders.

If you run an online store or offer a service and want to be among the first sources the assistant can read and buy from, or if you are interested in partnering on the project, get in touch. In the meantime, we can review your product data and tell you how ready your store is for AI shopping agents, from feeds and structured data to filters on a live OpenCart 3 or WooCommerce shop.

• Contact us about sparcy.online
• E-commerce web development