アンドワン Product Lineup

This document is a white paper aimed at organizing the structural reasons behind "what is valued in the AI era and why many web initiatives have ceased to function." It provides a detailed explanation of the evaluation (points) of the past and the evaluation (structure) in the AI era, as well as the structural consistency of the entire company as seen by AI, and the true web redesign process. AndOne is not a company that "creates" websites. We are a partner that organizes the structures of sales, business, and information, and designs the web by working backward from results. [Contents (partial)] ■ Introduction | Why have improvements to the web stopped yielding results? ■ The real changes occurring in the AI era ■ AI evaluates "companies" rather than "content" ■ Failures of companies where the sales structure and web are disconnected ■ Why have "correct web initiatives" stopped producing results? *For more details, please download the PDF or feel free to contact us.

This document provides a detailed explanation of the new fundamentals of customer acquisition that small and medium-sized enterprises should know now. We also introduce the benefits that GEO brings to companies, such as being "chosen even without being searched," the knowledge of all employees being turned into assets, and long-term evaluations that do not decline. AI tools like ChatGPT, Gemini, and Copilot are looking for "reliable sources of information to generate answers." As a result, the need for GEO (Generative Engine Optimization) has arisen. GEO is a mechanism that allows companies to gain trust by being cited by AI. 【Contents (partial)】 ■ Introduction | The End of the SEO Era and the Emergence of "GEO" ■ Vol.1 From "Discovery" to "Trust" in the AI Era ■ Vol.2 The Shift in Search from "Human → AI" ■ Vol.3 The Core of GEO is Creating "Reasons to be Cited" ■ Vol.4 Fighting in a "Forest of Context" Rather than a Sea of Information *For more details, please download the PDF or feel free to contact us.

This book is not a book that provides answers. It is a "diagnostic report" to calmly verify whether your company is an explainable entity for AI. It clearly explains the reasons why, even with technology, AI may not favor you, and the logic behind why AI includes certain candidates in its "recommendations." If you find yourself feeling uncomfortable while reading or become anxious about your company's website, that is not a problem. The fact that you have realized this is already a significant step. 【Contents】 ■ How does ChatGPT perceive "companies"? ■ Seven common structures of "companies not evaluated" by ChatGPT ■ The moment ChatGPT determines "this company is trustworthy" ■ What manufacturing websites are fundamentally getting wrong right now ■ So, how can you become a "company not disliked by AI"? ■ Why AndOne can write on this topic *For more details, please download the PDF or feel free to contact us.

Even if a site has an FAQ, the reasons for not reducing inquiries or increasing business negotiations are due to "inability to find or read" the information. Users will leave if they cannot quickly reach the answers they need. Furthermore, in B2B, questions tend to be complex (usage × conditions × specifications × delivery time), and simply having a list of FAQs does not resolve these issues. In this service, we will organize existing FAQs (or FAQs that will be developed) into a structure that "AI can answer" and implement an AI chat on the site. This is not just an automated response; we will streamline the process from "answer → related page suggestion → document download/quotation/consultation," achieving both efficiency in handling inquiries and lead acquisition. ■ Service Offerings (3 points) 1. AI optimization of FAQs (structuring, adjusting granularity, adding missing FAQs) 2. Implementation of AI chat (site embedding, answer design, process integration) 3. Operational design (guard against incorrect answers, log improvement, update flow) *First, please tell us the "questions you want to resolve via chat" and the "final goal (quotation/consultation/download)." We will design it as quickly as possible.

The main reason why AI chat fails to meet expectations is that product information is scattered in "text." Questions in the B2B manufacturing industry cannot be answered accurately unless they are structured data, such as applications, conditions, specifications, dimensions, materials, compatibility, and delivery times. Referring only to the text on the page leads to ambiguous answers and increases the risk of incorrect responses. This service will implement an AI chat that is linked to a product database (product master/specification database/compatibility tables/stock and delivery information, etc.). The chat will listen to the user's conditions, present relevant candidates from the database, and provide the basis (specification values and compatibility conditions), smoothly connecting to quotes, document downloads, and inquiries. ■ Provided Content (3 Points) 1. Product DB design/maintenance (items that can withstand AI reference/normalization) 2. AI chat implementation (database reference answers, candidate presentation, basis display) 3. Sales flow integration (quote condition collection, form/CRM integration, operational design) *First, please tell us the "number of products" and the "frequently asked questions from users (compatibility/selection/quotes)." We will design based on the optimal DB items.

Many of the reasons why B2B estimates do not progress are due to "inquiries coming in without the necessary information being gathered." Sales teams end up in a back-and-forth of additional questions → waiting for replies → re-questions, which delays the submission of estimates. As a result, the other party may turn to competitors, leading to a price competition scenario. This service automates the pre-estimate hearing process using AI. It tailors questions based on the user's objectives (wanting to know price expectations / wanting to discuss specifications / wanting to confirm delivery times) and collects the necessary information in stages. The collected information is then handed over to the sales team, preparing it for a state where "the estimate can be provided in the shortest time possible." While aiming for a higher collection rate than forms, we also design guardrails to avoid incorrect answers and legal risks. ■ Provided Content (3 points) 1. Pre-estimate hearing design (definition of questions, branching, and mandatory conditions) 2. AI hearing implementation (integration into the site/landing page, input assistance, summarization) 3. Sales handover design (CRM/ledger, email notifications, SLA, operations) *Please tell us the common types of estimate requests and the essential conditions for estimates (5-10 items). We will incorporate them into the flow.

The main reasons for the delay in responding to inquiries from overseas are "time difference" and "language." Users want answers immediately, but replies come the next day or later. English support becomes personalized, leading to inconsistent quality. As a result, comparisons with other companies progress. In this service, we will introduce an AI chat that can respond in multiple languages, primarily in English. It will provide immediate answers to questions by referencing FAQs, product pages, case studies, customs/delivery information, and agency details. If necessary, it will guide users to "pre-quotation hearings," "document downloads," "web meeting reservations," and "agency inquiries," reducing the loss of overseas leads and advancing negotiations. We will design guardrails to avoid incorrect answers and legal risks, including the operation and improvement cycle. ■ Service Offerings (3 points) 1. Multilingual response design (target languages, tone, terminology standardization) 2. AI chat implementation (site embedding, reference data organization, pathway integration) 3. Guardrails/operational design (preventing incorrect answers, log improvement, escalation) *Please let us know the target languages (English + ◯ languages) and the goals of the chat (quotation/consultation/DL/agency). We will design it as quickly as possible.

When internal knowledge becomes personalized, decision-making and responses slow down. Past proposals, estimation bases, specification judgments, trouble responses, transaction conditions, and FAQ response histories. When necessary information is in "people's heads" or "scattered files," the time spent searching increases, the quality of responses fluctuates, and handovers become disorganized. This service will build a knowledge AI (RAG) that can provide answers while searching and referencing internal documents. It is not just a simple chat; through the design of **"presenting the basis (source) and adhering to the scope of answers,"** it suppresses misinformation and speculative responses. We will create a state where sales, technical, and support teams can respond based on the same standards, thereby enhancing internal productivity and response speed. ■ Provided Content (3 points) 1. Knowledge inventory and design (target documents, classification, permissions, updates) 2. RAG construction (searching, summarizing, presenting basis, including UI/operations) 3. Guardrail maintenance (areas not to answer, audit logs, improvements) *First, please let us know the "location of the documents you want to use (Drive/SharePoint, etc.)" and the "usage purposes by department (sales/technical/CS)." We will design it as quickly as possible.*

Technical documentation in the manufacturing industry is the same as "non-existent" when it is scattered. Data sheets, specifications, CAD files, inspection reports, manuals, FAQs, past response emails. When users cannot find the necessary materials, they disengage, sales teams waste time searching, and engineers repeatedly address the same questions. This service will build an AI portal that allows cross-searching of technical documents. Instead of keyword searches, it will be designed to search in the context of applications, conditions, products, standards, etc., providing a series of results including "relevant documents," "summaries of relevant sections," "related pages," and "next actions (quotes/consultations)." It will cater to both external public versions (for customers) and internal versions (for sales/technical teams), aiming to reduce inquiries and facilitate business negotiations. ■Provided Content (3 Points) 1. Inventory of documents and information design (classification, tags, permissions, update flow) 2. Construction of AI search portal (searching/summarizing/citing sources/linking pathways) 3. Operation and guardrail design (preventing incorrect answers, improving logs, managing permissions) *Please let us know the types of documents (PDF/CAD/Excel, etc.), storage locations (Drive/SharePoint, etc.), and the scope of publication (external/internal). We will design it as quickly as possible.

The reason for the delay in responding to inquiries is not the content of the replies, but rather the "slowness of sorting." When inquiries with diverse content arrive in the same inbox, time is consumed in assigning responsibilities, prioritizing, and confirming necessary information. As a result, there are delays in hot leads, technical questions remain unresolved, and support-related issues are more likely to escalate. In this service, we will design a system that automatically classifies received inquiries using AI, assigning responsibility, priority, identifying missing information, providing expected response templates, and determining the next actions. Once the process of **"immediate sorting → immediate response"** begins, both response speed and conversion rates will increase simultaneously. ■ Provided Content (3 Points) 1. Classification design (categories, priority, assignment rules, exceptions) 2. Automatic judgment logic design (AI judgment + rule combination, reliability) 3. Operational design (human verification, improvement of misjudgments, KPIs) *First, please tell us the "types of inquiries (quotes/technical/agency/recruitment/complaints, etc.)." We will create a classification system from there.*

The biggest loss in lead management is "delaying responses to hot leads." In B2B, the consideration period is long, and if all leads are pursued with the same intensity, sales teams become exhausted, responses are delayed, and hot leads are missed. Conversely, applying "strong sales" tactics to colder prospects can lead to disengagement and a negative impression. In this service, we will design an AI evaluation logic to estimate the temperature of leads (Hot/Warm/Cold) based on inquiry content, document downloads, browsing behavior, form submissions, and company attributes. Depending on the temperature, we will differentiate response templates, proposal content, and next actions (calls/meetings/documents/waiting), clarifying sales priorities. ■ Provided Content (3 Points) 1. Design of temperature definitions (what constitutes Hot, verbalizing winning strategies) 2. Estimation logic design (AI + scoring, reliability) 3. Operational design (next action branching, improvement cycles, KPIs) *First, please tell us about the "winning patterns (characteristics of inquiries close to closing)" and the "system for follow-up (how many days it takes to respond)." We will translate this into evaluation criteria.

The reasons for losing B2B sales are often not due to price or product, but rather due to "insufficient design of communication." Proposals that do not match the client's situation, lack of comparative criteria, weak handling of objections, and ambiguity in next actions can accumulate, causing negotiations to stall at "under consideration." On the other hand, top salespeople adjust the "order of conversation," "ways of comparison," and "assertiveness" according to the situation, even with the same product. In this service, we will design a system that generates sales talks (hearing questions, proposal outlines, comparative explanations, objection handling, closing) using AI, tailored to the product, application, target roles (purchasing/design/field/management), and consideration phases. This is not just about generating text; we will establish a "winning pattern" that can be replicated by anyone. ■ Offered Content (3 Points) 1. Design of winning talk patterns (frameworks by phase and role) 2. AI generation design (input items, output format, rationale/restrictions) 3. Operational design (improvement through field feedback, log auditing) *First, please tell us about your "main products" and "common reasons for losing sales (points where you lose in comparisons)." We will translate this into a talk pattern.

Companies that are slow to make proposals will lose, even if they have a winning strategy. In B2B, the speed of proposals after the initial meeting directly correlates with trust. However, the reality is that information is scattered, not organized in the minds of the responsible parties, and proposals are created from "scratch" each time. As a result, submissions are delayed, and competitors move forward with their evaluations. This service is designed to generate a draft proposal (structure + key points + rationale + estimation assumptions) using AI based on hearing notes, form inputs, inquiry content, and existing materials. It is not just about generating text; it fixes the winning strategy of the proposal (issues → actions → results → approach → conditions) and makes it reproducible by anyone. ■ Service Offerings (3 points) 1. Proposal template design (structure of the winning strategy, chapter organization, essential elements) 2. Draft generation design (input items, output format, reference rationale) 3. Operation and review design (quality assurance, prevention of incorrect generation, improvements) *First, please provide the "template of the proposal (the chapter organization you are currently using)" and "common types of projects." We will create the generation framework.

The heavier the specifications at the site, the slower the decision-making becomes. "What are the changes?" "What is confirmed and what is uncertain?" "What are the risks?" "Who is doing what?" If these are not clear at a glance, meetings drag on, handovers break down, and rework during implementation increases. In this service, we will design a system that uses AI to summarize, extract differences, and organize issues related to internal documents (requirement definitions, design documents, meeting minutes, estimation conditions, etc.). This is not just a simple shortening; it will output in a form necessary for decision-making and implementation (decisions made/undecided items/dependencies/risks/tasks), thereby speeding up reviews and progress management. ■ Provided Content (3 Points) 1. Summary format design (fixed output based on reading purpose) 2. AI summary design (evidence presentation, difference extraction, prohibitions) 3. Operational design (version management, review, improvement, audit logs) *First, please tell us the type of document you want to summarize (requirements/design/minutes) and the issues you are facing internally (review/handover/difference confirmation). We will decide on the output format.

Reading blueprints and images is an area that is prone to becoming personalized. While veterans can make quick judgments just by looking at the drawings, explanations tend to stop when the person in charge changes. In international dealings, explanations in English can become cumbersome. Additionally, the images and drawings sent by clients vary greatly in quality and format, making it time-consuming to extract the necessary information. This service will design an AI-assisted system that organizes blueprints and images into a format suitable for "explanation, confirmation, and handover." Instead of leaving it to AI to determine "what is depicted," we will present perspectives (confirmation points, missing information, explanation templates) to assist human judgment and incorporate guardrails to avoid misunderstandings and overconfidence. ■ Service Offerings (3 points) 1. Use case design (internal judgment/customer explanation/international response/handover) 2. Output format design (explanatory text, confirmation items, differences, points of caution) 3. Guardrails/operational design (preventing overconfidence, confidentiality, log improvement) *First, please tell us the "type of images/blueprints (photo/2D/3D/with annotations)" and "purpose (internal/customer/English)." We will refine it into the optimal output format.

The reason why inquiries do not decrease despite having an FAQ is that the FAQ is structured in a way that is "difficult to search for both AI and humans." Questions are long, points are mixed, there are many variations in notation, conclusions are delayed, and prerequisite conditions are not stated. Such FAQs are hard for users to read and resolve, and they are also prone to misunderstanding by AI. In this service, we will restructure the existing FAQs (or inquiry logs) into a format that is easier for AI to reference and respond to. This is not just a simple rewrite; we will unify it in the format of **"Conclusion → Conditions → Evidence → Next Action,"** and also incorporate links to related pages, document downloads, and pathways to estimates/consultations. This design will be effective both as a preliminary step for implementing AI chat and as an improvement of the FAQ on its own. ■ Provided Content (3 Points) 1. FAQ Inventory (organizing duplicates, deficiencies, and mixed points) 2. Structured Rewrite (AI-friendly granularity, conditions, and notation standardization) 3. Pathway Design (related links, CTAs, inquiry/document download branching) *First, please tell us the "top 20 frequently asked questions" and the "types of inquiries you want to reduce." We will determine the granularity of the FAQ.*

The biggest risk of AI is not providing smart answers, but rather "plausible incorrect answers." Definitive pricing, performance guarantees, misunderstandings of legal and regulatory matters, leaks of confidential information, and misrepresentation of internal rules. Once this happens, it can lead to complaints, loss of trust, and damages. In this service, we will organize the risks that may arise from AI chat, internal AI, automated inquiries, etc., and design rules for **"answering/not answering," "not making definitive statements," and "always confirming."** We will prepare response templates, provide evidence, verify with people based on trust levels, maintain audit logs, and establish improvement processes to ensure that AI is in a state that can "withstand business operations." ■ Services Offered (3 Points) 1. Risk Inventory (defining unacceptable responses) 2. Guardrail Design (scope, expression, evidence, escalation) 3. Operational Design (audit, improvement, responsibility delineation, KPI) *First, please tell us the "areas where you want AI to provide answers (estimation/technology/delivery/internal regulations, etc.)." We will design based on the prohibited areas in reverse.*

The main reason why AI chat is not effective is that "logs are not being used for improvement." What did users ask? Where did they drop off? What did the AI misunderstand? Which questions led to conversions? Without examining these, accuracy and results will not improve. In this service, we will design the AI chat conversation logs in a "modifiable format." Rather than just viewing them, we will classify unanswered questions, incorrect answers, misunderstandings, insufficient FAQs, flaws in the user journey, guardrail activations, and escalations, and establish a dashboard and operational flow that highlights areas for improvement on a monthly basis. We will create a state where we can identify targets for improvement, including adding FAQs, terminology dictionaries, user journeys, question flows, and guardrails. ■ Provided Content (3 points) 1. Log collection and measurement design (what to record, granularity, anonymization) 2. Analysis design (classification axes, metrics, dashboards, alerts) 3. Improvement operation design (monthly cycle, responsibilities, priorities, KPIs) *First, please tell us the "purpose of the chat (estimation/consultation/download/self-resolution)" and "the symptoms you are currently facing (many unanswered questions/many drop-offs, etc.)." We will determine the analysis axes.

As companies become capable of creating content with AI, they are more likely to experience "quality incidents." Variations in phrasing, lack of terminology consistency, inclusion of unfounded numbers or assertions, and breaches of confidentiality can accumulate, leading to a decline in brand trust and inconsistencies in sales presentations. In this service, we will establish rules for **"what to say, in what format, based on what evidence, with what notation, and to what extent it is permissible to assert"** when creating articles, product pages, FAQs, proposals, emails, etc., using AI. Additionally, we will organize the review process, approval flow, logs, prohibited areas, and standard forms of prompts to ensure that AI utilization does not lead to "person-dependent production operations." ■ Provided Content (3 Points) 1. Establishment of Quality Standards (tone, structure, evidence, prohibitions, notation consistency) 2. Design of Production Flow (role distribution, review, approval, version control) 3. Development of Standard Templates (prompt templates, output formats, checklists) *First, please tell us the "type of content you want to create with AI (product page/blog/FAQ/email, etc.)" and the "brand tone you want to protect." We will design the rules accordingly.*

The patterns of failure in AI implementation are common. The objectives are vague, there is insufficient data, KPIs are not established, there is no risk inventory, and there is no operational framework. In this situation, the PoC ends up being just a "working demo" and cannot transition to production. In this service, we design the PoC (Proof of Concept) for AI implementation in the web domain (AI chat, RAG, inquiry automation, content generation, etc.) in a way that connects to "production." We clarify **"what needs to be validated to determine GO/NO-GO"**, and by detailing the necessary data, evaluation metrics, guardrails, and operational assumptions, we facilitate the shortest route to production implementation. ■ Provided Content (3 points) 1. PoC Design (Objectives, Hypotheses, Validation Scope, KPIs, Success Conditions) 2. Data/UX Design (Target Content, User Flow, Evidence Presentation, Evaluation) 3. Risk/Operational Design (Guardrails, Auditing, Improvement, Structure) *First, please decide on one theme you want to solve with AI (e.g., reducing inquiries, pre-estimate hearings). A focused PoC is the fastest approach.*

The reason for the lack of increased inquiries is that users do not know "what to send." B2B inquiries often involve many uncertain factors such as quotation conditions, usage, specifications, delivery dates, and quantities, causing users to hesitate in front of the form. As a result, it leads to "users leaving after just viewing the materials" or "vague inquiries," which increases back-and-forth communication for the sales team. In this service, AI will conduct a hearing before the inquiry to organize the situation, confirm conditions, collect missing information, and design the next action branches (quotation/material download/consultation/agency). While reducing the burden of form input, we aim to increase inquiries with "necessary information gathered," thereby improving initial response speed and conversion rates into business negotiations. ■ Service Offerings (3 points) 1. Flow design (entry points, branching, CTAs, form integration) 2. AI hearing design (question order, required items, exceptions) 3. Operational design (log improvement, guardrails, assignment of responsibilities) *First, please decide on one type of inquiry you want to increase (quotation/technical consultation/agency/recruitment, etc.). From there, we will create the shortest flow.

The more AI utilization progresses, the more the lack of "rules" itself becomes a risk for companies. Handling of confidential information, personal data, copyright, incorrect answers, external transmission, audits, and delineation of responsibility. If these are left ambiguous and left to the field, it will result in a situation where "the only option is to stop" when an accident occurs. In this service, we will formulate AI operation policies targeting AI chat, internal AI, RAG, content generation, sales automation, and more. This is not just an ethical guideline, but rather a concrete set of practical rules that clearly defines **"what is permitted, what is prohibited, who approves, how audits are conducted, and how improvements are made."** The policies will be designed at a level that can be adhered to by the field and will be applicable to operations. ■ Provided Content (3 points) 1. Target and Risk Organization (Inventory of AI usage scope and accident patterns) 2. Formulation of Operation Policies (Permissions/Prohibitions, Responsibility Delineation, Procedures) 3. Education, Audit, and Improvement Design (Establishment and Continuous Operation) *First, please tell us the uses of AI being utilized internally (text, search, chat, development). We will design rules categorized by usage.*

We have published a white paper summarizing the structural conditions necessary for manufacturing industry websites to be evaluated and recommended by AI, as revealed in an AI demonstration experiment. ■ What does AI evaluate? In a demonstration case from February 2026, specific materials were automatically displayed as related materials solely based on algorithms, without relying on human selection or advertising budgets. AI assesses not just simple word matching (keyword volume), but rather the "similarity of meaning structures" of the entire text. ■ Five structural conditions to be chosen by AI This document explains the conditions that AI prioritizes when summarizing, comparing, and recommending information. ■ New evaluation metric for manufacturing websites: "GEO" Simply listing specifications will not lead to evaluation by AI. We will introduce a method to redefine the web not as a sales tool, but as a "blueprint that shows the structure of the company." *For more details, please download the PDF document.

Retirement of skilled workers, diversification of products in small quantities, and limitations of paper management. In the manufacturing industry, the situation where "production stops if there is no one who knows" is becoming more serious. This service provides a comprehensive solution that includes: - AI structuring of paper/Excel/oral knowledge - Tag search/trouble top display - Visual process navigation - Tablet-compatible business application development It is not just a simple AI-OCR. We go as far as translating the digitized knowledge into a "user interface that can be used on-site." ■ Service offerings (3 points) 1. Knowledge structuring design (Organization, tag design, and accident pattern inventory using AI-OCR and LLM) 2. Search and recommendation engine development (Tag search/natural language QA/related knowledge display) 3. Implementation of on-site business web applications (Checklists/visual navigation/management screens/log acquisition)