Byte Digital is building AI search optimisation (AI SEO) into the new website for Alltype Roofing (opens in a new tab), a roofing merchant. The brief connects two customer decisions: can someone asking an AI assistant where to buy a particular roofing product identify Alltype as a supplier, and can they then find the right item and order it on their phone?
A buying question the website should be able to answer
Consider the question: “Where can I buy a Marley Ludlow Plus roof tile in smooth grey?” This is an illustrative question based on a product in Alltype’s existing catalogue (opens in a new tab). It does not name the retailer. That makes it useful for testing whether an assistant such as Gemini discovers Alltype, rather than whether it can repeat facts about a company the buyer already knows.
A useful answer would identify the right tile and link to a page where the buyer can check it. A link to Alltype’s homepage is less useful to someone ready to order. If the buyer also asks about local collection, evidence that Alltype sells the tile and evidence that a particular depot has it available are two different things. An accurate answer should only go as far as the source supports.
What Byte has built for AI-assisted product discovery
As of 2 November 2026, Byte has created the visual design and built the HTML and CSS front end. The work includes schema markup, location information and FAQs. Each has a specific job in the AI SEO brief: the markup should describe the same product the buyer sees, the location information should make clear where the retailer operates, and the FAQs should answer buying questions that the product name alone cannot resolve.
The next part of the journey is commercial. The previous design felt cumbersome on a phone. Once someone follows a product link, the aim is for them to recognise the item they need and place the order themselves.
Alongside the website, Byte is using AI to build internal tools intended to help the Alltype team work more productively.
How the project can be assessed after launch
This is a work-in-progress implementation case study. A useful post-launch test would repeat the unbranded tile question in fresh Gemini chats and retain the answers and sources, including misses. Record whether Gemini names Alltype, recommends it as a supplier, cites the correct product page and accurately describes the item. A separate question naming Alltype would test factual accuracy about a known retailer.
The business objective is more completed online orders, particularly from customers shopping on mobile. Identifiable AI referrals, mobile orders and mobile purchase conversion should be measured separately: a citation does not show that someone visited, and a mobile order alone does not identify how the buyer found Alltype. No post-launch AI visibility or order results are claimed here.
References
- Byte Digital project update, 2 November 2026: website scope, delivery stage, audience and commercial objective.
- Alltype Roofing: Marley Ludlow Plus Tile Smooth Grey 28 (opens in a new tab) — existing public product example.
- Byte Digital: AI search for ecommerce retailers and AI search measurement.
- Alltype Roofing paid media — separate historical project.