{"id":164933,"date":"2026-10-22T20:37:57","date_gmt":"2026-10-22T13:37:57","guid":{"rendered":"https:\/\/bap-software.net\/?post_type=knowledge&#038;p=164933"},"modified":"2026-10-22T20:37:57","modified_gmt":"2026-10-22T13:37:57","slug":"rag-development-process-cost","status":"publish","type":"knowledge","link":"https:\/\/bap-software.net\/en\/knowledge\/rag-development-process-cost\/","title":{"rendered":"How to Build a RAG System: Process, Cost and Timeline \u2014 Lessons from an 8-Week PoC"},"content":{"rendered":"<p>More and more companies are turning to RAG (Retrieval-Augmented Generation) to put their internal documents and business data to work with generative AI. But as soon as a project gets going, the same questions come up:<\/p>\n<ul style=\"margin: 0;padding-left: 20px\">\n<li style=\"margin: 6px 0\">How much does it cost to build a RAG system?<\/li>\n<li style=\"margin: 6px 0\">How many months does it take to go from PoC to production?<\/li>\n<li style=\"margin: 6px 0\">What data do we need to prepare?<\/li>\n<li style=\"margin: 6px 0\">Can we use Excel files and databases, not just documents?<\/li>\n<li style=\"margin: 6px 0\">How do we know whether the PoC has succeeded?<\/li>\n<\/ul>\n<div style=\"background: #f1f6fd;border-left: 5px solid #0d47a1;border-radius: 0 8px 8px 0;padding: 16px 22px;margin: 22px 0\">\n<p style=\"margin: 0 0 6px;font-weight: bold;color: #0d47a1\">The short answer<\/p>\n<div>A technical validation with a narrow data scope can start in <strong>2\u20134 weeks<\/strong>. Evaluating answer quality and business fit with real data typically takes a <strong>1\u20133 month<\/strong> PoC. In a project for a Japanese enterprise, BAP Solution Japan combined structured data such as Excel with internal documents and completed a <strong>PoC integrating SQL search and RAG search in 8 weeks<\/strong>.<\/div>\n<\/div>\n<div style=\"background: #f1f6fd;border: 1px solid #c5d8f3;border-radius: 8px;padding: 18px 24px;margin: 24px 0\">\n<p style=\"margin: 0 0 8px;font-weight: bold;color: #0d47a1\">What you will learn<\/p>\n<ul style=\"margin: 0;padding-left: 20px\">\n<li style=\"margin: 6px 0\">How RAG works, and typical timelines and costs<\/li>\n<li style=\"margin: 6px 0\">Six factors that drive cost \u2014 and how prototypes, PoCs and production systems differ<\/li>\n<li style=\"margin: 6px 0\">A six-step process for building a RAG system<\/li>\n<li style=\"margin: 6px 0\">An 8-week SQL + RAG PoC case study, and how to judge PoC success<\/li>\n<\/ul>\n<\/div>\n<hr class=\"wp-block-separator\" \/>\n<h2 style=\"border-left: 6px solid #0d47a1;padding: 6px 0 6px 14px;margin: 44px 0 18px\">What Is RAG?<\/h2>\n<p>RAG stands for Retrieval-Augmented Generation. When a user asks a question, the system searches internal documents and databases for relevant information, passes it to a large language model (LLM), and generates an answer grounded in that information.<\/p>\n<div style=\"display: flex;flex-wrap: wrap;gap: 10px;margin: 20px 0;align-items: stretch\">\n<div style=\"flex: 1 1 170px;background: #e3f2fd;border: 1px solid #90b4e6;border-radius: 8px;padding: 12px 14px\">\n<p style=\"margin: 0 0 4px;font-size: 0.8em;font-weight: bold;color: #0d47a1\">STEP 1<\/p>\n<p style=\"margin: 0 0 4px;font-weight: bold\">Retrieve<\/p>\n<p style=\"margin: 0;font-size: 0.9em\">Find information relevant to the question in documents and databases<\/p>\n<\/div>\n<div style=\"flex: 1 1 170px;background: #e8f5e9;border: 1px solid #a5d6a7;border-radius: 8px;padding: 12px 14px\">\n<p style=\"margin: 0 0 4px;font-size: 0.8em;font-weight: bold;color: #0d47a1\">STEP 2<\/p>\n<p style=\"margin: 0 0 4px;font-weight: bold\">Augment<\/p>\n<p style=\"margin: 0;font-size: 0.9em\">Pass the search results to the LLM as context<\/p>\n<\/div>\n<div style=\"flex: 1 1 170px;background: #e3f2fd;border: 1px solid #90b4e6;border-radius: 8px;padding: 12px 14px\">\n<p style=\"margin: 0 0 4px;font-size: 0.8em;font-weight: bold;color: #0d47a1\">STEP 3<\/p>\n<p style=\"margin: 0 0 4px;font-weight: bold\">Generate<\/p>\n<p style=\"margin: 0;font-size: 0.9em\">Produce an answer based on the retrieved information<\/p>\n<\/div>\n<\/div>\n<p>A general-purpose generative AI answers only from what it learned during training. With RAG, answers can be <strong>grounded in your own information<\/strong> \u2014 internal policies, product manuals, technical documents, past support tickets \u2014 and sources can be shown alongside each answer.<\/p>\n<div style=\"background: #fff3e0;border-left: 5px solid #e65100;border-radius: 0 8px 8px 0;padding: 16px 22px;margin: 22px 0\">\n<p style=\"margin: 0 0 6px;font-weight: bold;color: #e65100\">Keep in mind<\/p>\n<div>RAG does not eliminate wrong answers. Data quality, retrieval accuracy, prompts and the handling of questions the system cannot answer all have to be designed deliberately.<\/div>\n<\/div>\n<h2 style=\"border-left: 6px solid #0d47a1;padding: 6px 0 6px 14px;margin: 44px 0 18px\">How Long Does It Take to Build a RAG System?<\/h2>\n<p>RAG projects usually move from technical validation to PoC, then to production development and company-wide rollout.<\/p>\n<div style=\"margin: 20px 0\">\n<table style=\"width: 100%;border-collapse: collapse;min-width: 560px;font-size: 0.95em\">\n<caption style=\"caption-side: top;text-align: left;font-weight: bold;padding: 0 0 8px\">Typical timeline by phase<\/caption>\n<thead>\n<tr>\n<th style=\"background: #0d47a1;color: #fff;padding: 10px 12px;border: 1px solid #0d47a1;text-align: left\">Phase<\/th>\n<th style=\"background: #0d47a1;color: #fff;padding: 10px 12px;border: 1px solid #0d47a1;text-align: left\">Typical duration<\/th>\n<th style=\"background: #0d47a1;color: #fff;padding: 10px 12px;border: 1px solid #0d47a1;text-align: left\">Main purpose<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr style=\"background: #fff\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>Technical validation \/ prototype<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">2\u20134 weeks<\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Check retrieval and answers with limited data<\/td>\n<\/tr>\n<tr style=\"background: #f7f9fc\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>Small-scale PoC<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">1\u20132 months<\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Evaluate accuracy and value with real data<\/td>\n<\/tr>\n<tr style=\"background: #fff\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>PoC with mixed data<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">2\u20133 months<\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Validate documents and databases together<\/td>\n<\/tr>\n<tr style=\"background: #f7f9fc\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>Departmental production system<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">2\u20134 months<\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Implement authentication, permissions, logging and integrations<\/td>\n<\/tr>\n<tr style=\"background: #fff\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>Company-wide or customer-facing rollout<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">4\u20138+ months<\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">High availability, auditing, multi-department rollout<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p>A 2\u20134 week technical validation narrows the documents and questions in scope to prove basic feasibility. A business PoC uses the questions real users would ask to evaluate retrieval accuracy, answer quality and business impact. If data preparation or integration with several systems is needed, expect 2\u20133 months.<\/p>\n<h2 style=\"border-left: 6px solid #0d47a1;padding: 6px 0 6px 14px;margin: 44px 0 18px\">How Much Does a RAG System Cost?<\/h2>\n<p>There is no standard price for a RAG system. Based on publicly available information on RAG development in the Japanese market, typical budgets look like this:<\/p>\n<div style=\"margin: 20px 0\">\n<table style=\"width: 100%;border-collapse: collapse;min-width: 560px;font-size: 0.95em\">\n<caption style=\"caption-side: top;text-align: left;font-weight: bold;padding: 0 0 8px\">Typical cost by scope (Japanese market, JPY)<\/caption>\n<thead>\n<tr>\n<th style=\"background: #0d47a1;color: #fff;padding: 10px 12px;border: 1px solid #0d47a1;text-align: left\">Scope<\/th>\n<th style=\"background: #0d47a1;color: #fff;padding: 10px 12px;border: 1px solid #0d47a1;text-align: left\">Typical cost<\/th>\n<th style=\"background: #0d47a1;color: #fff;padding: 10px 12px;border: 1px solid #0d47a1;text-align: left\">What is usually included<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr style=\"background: #fff\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>Prototype<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">\u00a50.5M\u2013\u00a52M<\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Limited data, basic search and answers<\/td>\n<\/tr>\n<tr style=\"background: #f7f9fc\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>Small-scale PoC<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">\u00a51.5M\u2013\u00a55M<\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Requirements, data processing, accuracy evaluation<\/td>\n<\/tr>\n<tr style=\"background: #fff\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>PoC with mixed data<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">\u00a53M\u2013\u00a58M+<\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Document and database integration, business-fit evaluation<\/td>\n<\/tr>\n<tr style=\"background: #f7f9fc\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>Departmental production system<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">\u00a55M\u2013\u00a515M+<\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Authentication, permissions, logging, external integrations<\/td>\n<\/tr>\n<tr style=\"background: #fff\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>Company-wide or customer-facing system<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Custom quote<\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">High availability, auditing, load handling<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p style=\"font-size: 0.85em;color: #607d8b;margin: -8px 0 20px\">*Rough market estimates, not BAP\u2019s pricing. Actual costs are calculated after reviewing the target process, data, required accuracy, integrations and security requirements.<\/p>\n<h2 style=\"border-left: 6px solid #0d47a1;padding: 6px 0 6px 14px;margin: 44px 0 18px\">Six Factors That Drive the Cost of RAG<\/h2>\n<div style=\"margin: 20px 0\">\n<table style=\"width: 100%;border-collapse: collapse;min-width: 560px;font-size: 0.95em\">\n<thead>\n<tr>\n<th style=\"background: #0d47a1;color: #fff;padding: 10px 12px;border: 1px solid #0d47a1;text-align: left\">Factor<\/th>\n<th style=\"background: #0d47a1;color: #fff;padding: 10px 12px;border: 1px solid #0d47a1;text-align: left\">Impact on cost and timeline<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr style=\"background: #fff\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>Data types<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Document-only RAG needs a different scope from a system that also uses Excel and databases. Design must fit the data \u2014 e.g. SQL for structured data, vector search for documents<\/td>\n<\/tr>\n<tr style=\"background: #f7f9fc\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>Data condition<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Outdated documents, duplicates and multiple versions take time to clean up. Scanned PDFs, images and complex tables also need OCR or layout analysis<\/td>\n<\/tr>\n<tr style=\"background: #fff\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>Required accuracy<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">General internal search and work such as contracts, finance or quality control demand very different accuracy. High-stakes use cases need more evaluation data and human approval<\/td>\n<\/tr>\n<tr style=\"background: #f7f9fc\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>Access control<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Design is far more complex when visibility depends on department or role. Confirm permission requirements from the PoC stage<\/td>\n<\/tr>\n<tr style=\"background: #fff\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>Integrations<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Connecting to SharePoint, Google Drive, Box, CRM, ERP or document management systems requires designing APIs, authentication and data sync<\/td>\n<\/tr>\n<tr style=\"background: #f7f9fc\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>Operations after go-live<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Document updates, answer-quality monitoring, API cost control, incident response and model updates \u2014 billed as monthly maintenance on top of the initial build<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<h2 style=\"border-left: 6px solid #0d47a1;padding: 6px 0 6px 14px;margin: 44px 0 18px\">Prototype vs. PoC vs. Production<\/h2>\n<div style=\"margin: 20px 0\">\n<table style=\"width: 100%;border-collapse: collapse;min-width: 560px;font-size: 0.95em\">\n<thead>\n<tr>\n<th style=\"background: #0d47a1;color: #fff;padding: 10px 12px;border: 1px solid #0d47a1;text-align: left\">Phase<\/th>\n<th style=\"background: #0d47a1;color: #fff;padding: 10px 12px;border: 1px solid #0d47a1;text-align: left\">Goal<\/th>\n<th style=\"background: #0d47a1;color: #fff;padding: 10px 12px;border: 1px solid #0d47a1;text-align: left\">Typical deliverables<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr style=\"background: #fff\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>Prototype<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Confirm that the basic RAG mechanism works<\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Simple UI, limited search index, basic answering (business impact and permissions often not evaluated)<\/td>\n<\/tr>\n<tr style=\"background: #f7f9fc\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>PoC<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Verify that RAG fits the business, using real data and representative questions<\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Validation system, evaluation questions and expected answers, accuracy and quality results, issues and improvements, proposed production architecture, rough production schedule<\/td>\n<\/tr>\n<tr style=\"background: #fff\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>Production<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Make the system safe for everyday use<\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">User authentication, access control, activity and audit logs, data updates, quality and API cost monitoring, backups, incident response, security controls<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<div style=\"background: #f1f6fd;border-left: 5px solid #0d47a1;border-radius: 0 8px 8px 0;padding: 16px 22px;margin: 22px 0\">\n<p style=\"margin: 0 0 6px;font-weight: bold;color: #0d47a1\">Key point<\/p>\n<div>What a PoC can answer and what you can safely run in production are two different things. Keep production requirements in view from the PoC stage.<\/div>\n<\/div>\n<h2 style=\"border-left: 6px solid #0d47a1;padding: 6px 0 6px 14px;margin: 44px 0 18px\">How to Build a RAG System: Six Steps<\/h2>\n<div style=\"display: flex;gap: 14px;align-items: flex-start;border: 1px solid #dfe5ee;background: #f1f6fd;border-radius: 8px;padding: 14px 18px;margin: 0 0 12px\">\n<div style=\"flex: 0 0 36px;height: 36px;border-radius: 50%;background: #0d47a1;color: #fff;font-weight: bold;text-align: center;line-height: 36px\">1<\/div>\n<div style=\"flex: 1\">\n<h3 style=\"margin: 0 0 6px;font-size: 1.05em\">Define the target use case<\/h3>\n<p>\u201cMake internal information searchable\u201d is far too broad. Make it specific \u2014 and keep the first PoC to one process, department or product so it is easy to evaluate.<\/p>\n<p style=\"margin: 12px 0 20px;line-height: 2.4\"><span style=\"display: inline-block;background: #f1f6fd;border: 1px solid #c5d8f3;color: #0d47a1;border-radius: 16px;padding: 2px 12px;margin: 0 6px 0 0;font-size: 0.9em;line-height: 1.8\">Internal HR policy questions<\/span> <span style=\"display: inline-block;background: #f1f6fd;border: 1px solid #c5d8f3;color: #0d47a1;border-radius: 16px;padding: 2px 12px;margin: 0 6px 0 0;font-size: 0.9em;line-height: 1.8\">Support answers from product manuals<\/span> <span style=\"display: inline-block;background: #f1f6fd;border: 1px solid #c5d8f3;color: #0d47a1;border-radius: 16px;padding: 2px 12px;margin: 0 6px 0 0;font-size: 0.9em;line-height: 1.8\">Searching past equipment incidents<\/span> <span style=\"display: inline-block;background: #f1f6fd;border: 1px solid #c5d8f3;color: #0d47a1;border-radius: 16px;padding: 2px 12px;margin: 0 6px 0 0;font-size: 0.9em;line-height: 1.8\">Variance analysis of management figures<\/span> <span style=\"display: inline-block;background: #f1f6fd;border: 1px solid #c5d8f3;color: #0d47a1;border-radius: 16px;padding: 2px 12px;margin: 0 6px 0 0;font-size: 0.9em;line-height: 1.8\">Finding sales proposal materials<\/span><\/p>\n<\/div>\n<\/div>\n<div style=\"display: flex;gap: 14px;align-items: flex-start;border: 1px solid #dfe5ee;background: #f1f6fd;border-radius: 8px;padding: 14px 18px;margin: 0 0 12px\">\n<div style=\"flex: 0 0 36px;height: 36px;border-radius: 50%;background: #0d47a1;color: #fff;font-weight: bold;text-align: center;line-height: 36px\">2<\/div>\n<div style=\"flex: 1\">\n<h3 style=\"margin: 0 0 6px;font-size: 1.05em\">Inventory your data<\/h3>\n<p>Check file formats, document counts and data volume, database structure, storage locations, update frequency, old versions and duplicates, personal and confidential data, and access rights. Excel figures and PDF documents need different retrieval methods, so understand your data before designing retrieval.<\/p>\n<\/div>\n<\/div>\n<div style=\"display: flex;gap: 14px;align-items: flex-start;border: 1px solid #dfe5ee;background: #f1f6fd;border-radius: 8px;padding: 14px 18px;margin: 0 0 12px\">\n<div style=\"flex: 0 0 36px;height: 36px;border-radius: 50%;background: #0d47a1;color: #fff;font-weight: bold;text-align: center;line-height: 36px\">3<\/div>\n<div style=\"flex: 1\">\n<h3 style=\"margin: 0 0 6px;font-size: 1.05em\">Build an evaluation set<\/h3>\n<p>Collect questions real users will ask. For each, define the expected answer, the documents to reference, the figures to use, unacceptable answers, and when a human must review. Without an evaluation set, you cannot tell whether fluent-sounding output is actually fit for work.<\/p>\n<\/div>\n<\/div>\n<div style=\"display: flex;gap: 14px;align-items: flex-start;border: 1px solid #dfe5ee;background: #f1f6fd;border-radius: 8px;padding: 14px 18px;margin: 0 0 12px\">\n<div style=\"flex: 0 0 36px;height: 36px;border-radius: 50%;background: #0d47a1;color: #fff;font-weight: bold;text-align: center;line-height: 36px\">4<\/div>\n<div style=\"flex: 1\">\n<h3 style=\"margin: 0 0 6px;font-size: 1.05em\">Build the retrieval layer<\/h3>\n<p>For documents: text extraction, chunking, metadata tagging and embedding generation, stored in a vector database. For structured data: check data types, field names, table relationships and aggregation logic, then load it into a relational database.<\/p>\n<\/div>\n<\/div>\n<div style=\"display: flex;gap: 14px;align-items: flex-start;border: 1px solid #dfe5ee;background: #f1f6fd;border-radius: 8px;padding: 14px 18px;margin: 0 0 12px\">\n<div style=\"flex: 0 0 36px;height: 36px;border-radius: 50%;background: #0d47a1;color: #fff;font-weight: bold;text-align: center;line-height: 36px\">5<\/div>\n<div style=\"flex: 1\">\n<h3 style=\"margin: 0 0 6px;font-size: 1.05em\">Route each question to the right method<\/h3>\n<p>Classify each question: document questions go to RAG search, numeric questions to SQL, mixed questions to both, and out-of-scope questions are declined. Use predefined workflows where the procedure is fixed, and let the AI plan its own steps where the path depends on the question.<\/p>\n<\/div>\n<\/div>\n<div style=\"display: flex;gap: 14px;align-items: flex-start;border: 1px solid #dfe5ee;background: #e8f5e9;border-radius: 8px;padding: 14px 18px;margin: 0 0 12px\">\n<div style=\"flex: 0 0 36px;height: 36px;border-radius: 50%;background: #2e7d32;color: #fff;font-weight: bold;text-align: center;line-height: 36px\">6<\/div>\n<div style=\"flex: 1\">\n<h3 style=\"margin: 0 0 6px;font-size: 1.05em\">Evaluate retrieval and generation separately<\/h3>\n<p>When an answer is wrong, pin down the cause: relevant documents not retrieved, right documents but poor generation, outdated data, calculation errors, unsupported content added, or data retrieved outside the user\u2019s permissions. That tells you whether to fix the data, retrieval logic, prompts or the LLM. Microsoft\u2019s guide \u201c<a href=\"https:\/\/learn.microsoft.com\/en-us\/azure\/architecture\/ai-ml\/guide\/rag\/rag-solution-design-and-evaluation-guide\" target=\"_blank\" rel=\"noopener\">Design and develop a RAG solution<\/a>\u201d covers this approach in detail.<\/p>\n<\/div>\n<\/div>\n<h2 style=\"border-left: 6px solid #0d47a1;padding: 6px 0 6px 14px;margin: 44px 0 18px\">Case Study: An 8-Week RAG PoC for Management Accounting<\/h2>\n<p>BAP Solution Japan supported a Japanese enterprise in running a PoC that applied RAG and generative AI to management accounting.<\/p>\n<div style=\"display: flex;flex-wrap: wrap;gap: 16px;margin: 20px 0\">\n<div style=\"flex: 1 1 260px;background: #ffebee;border-top: 4px solid #c62828;border-radius: 8px;padding: 18px 22px\">\n<h3 style=\"color: #c62828;margin: 0 0 10px;font-size: 1.1em\">The challenge<\/h3>\n<p>Collecting management figures from multiple sources, formatting them and aggregating them took significant effort. Staff were also doing the following by hand:<\/p>\n<ul style=\"margin: 0;padding-left: 20px\">\n<li style=\"margin: 6px 0\">Checking variances between actuals and plan<\/li>\n<li style=\"margin: 6px 0\">Investigating causes in historical data<\/li>\n<li style=\"margin: 6px 0\">Exploring improvement measures<\/li>\n<li style=\"margin: 6px 0\">Preparing materials for management meetings<\/li>\n<\/ul>\n<\/div>\n<div style=\"flex: 1 1 260px;background: #e8f5e9;border-top: 4px solid #2e7d32;border-radius: 8px;padding: 18px 22px\">\n<h3 style=\"color: #2e7d32;margin: 0 0 10px;font-size: 1.1em\">PoC objectives<\/h3>\n<ul style=\"margin: 0;padding-left: 20px\">\n<li style=\"margin: 6px 0\"><strong>Technical:<\/strong> retrieve historical figures and related documents accurately<\/li>\n<li style=\"margin: 6px 0\"><strong>Business:<\/strong> confirm that AI-generated explanations and proposals are sound<\/li>\n<li style=\"margin: 6px 0\"><strong>End goal:<\/strong> output good enough to serve as a first draft of management reports, always reviewed by a person<\/li>\n<\/ul>\n<\/div>\n<\/div>\n<h3 style=\"margin: 28px 0 12px;color: #0d47a1\">The 8-week plan and team<\/h3>\n<p>The project ran for two months (8 weeks). After the client walked the team through its data structures and how the data is used, BAP organized the requirements and built and tested the system feature by feature. Each cycle was checked against real data, and the feedback was used to refine the answering approach and processing logic. In the final stage, the team prepared the environment and documentation and delivered on schedule.<\/p>\n<p style=\"margin: 12px 0 20px;line-height: 2.4\"><span style=\"display: inline-block;background: #f1f6fd;border: 1px solid #c5d8f3;color: #0d47a1;border-radius: 16px;padding: 2px 12px;margin: 0 6px 0 0;font-size: 0.9em;line-height: 1.8\">Project manager<\/span> <span style=\"display: inline-block;background: #f1f6fd;border: 1px solid #c5d8f3;color: #0d47a1;border-radius: 16px;padding: 2px 12px;margin: 0 6px 0 0;font-size: 0.9em;line-height: 1.8\">Bridge SE (Japanese-speaking)<\/span> <span style=\"display: inline-block;background: #f1f6fd;border: 1px solid #c5d8f3;color: #0d47a1;border-radius: 16px;padding: 2px 12px;margin: 0 6px 0 0;font-size: 0.9em;line-height: 1.8\">AI engineer<\/span> <span style=\"display: inline-block;background: #f1f6fd;border: 1px solid #c5d8f3;color: #0d47a1;border-radius: 16px;padding: 2px 12px;margin: 0 6px 0 0;font-size: 0.9em;line-height: 1.8\">Full-stack engineer<\/span> <span style=\"display: inline-block;background: #f1f6fd;border: 1px solid #c5d8f3;color: #0d47a1;border-radius: 16px;padding: 2px 12px;margin: 0 6px 0 0;font-size: 0.9em;line-height: 1.8\">Infrastructure engineer<\/span><\/p>\n<p>The team covered not only the AI model but also business requirements, databases, the web application and the cloud environment end to end. Read more about our setup in \u201c<a href=\"https:\/\/bap-software.net\/knowledge\/vietnam-ai-development-company-japan\/\">Our AI development team for Japanese companies<\/a>.\u201d<\/p>\n<h3 style=\"margin: 28px 0 12px;color: #0d47a1\">Technology stack<\/h3>\n<div style=\"margin: 20px 0\">\n<table style=\"width: 100%;border-collapse: collapse;min-width: 560px;font-size: 0.95em\">\n<thead>\n<tr>\n<th style=\"background: #0d47a1;color: #fff;padding: 10px 12px;border: 1px solid #0d47a1;text-align: left\">Area<\/th>\n<th style=\"background: #0d47a1;color: #fff;padding: 10px 12px;border: 1px solid #0d47a1;text-align: left\">Technology<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr style=\"background: #fff\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>AI platform<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Amazon Bedrock<\/td>\n<\/tr>\n<tr style=\"background: #f7f9fc\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>LLM<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Claude Sonnet 4.5<\/td>\n<\/tr>\n<tr style=\"background: #fff\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>RAG \/ vector search<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">PostgreSQL, pgvector<\/td>\n<\/tr>\n<tr style=\"background: #f7f9fc\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>Backend<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Python, FastAPI, Uvicorn, Gunicorn<\/td>\n<\/tr>\n<tr style=\"background: #fff\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>Frontend<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">React, TypeScript<\/td>\n<\/tr>\n<tr style=\"background: #f7f9fc\">\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\"><strong>Data<\/strong><\/td>\n<td style=\"padding: 10px 12px;border: 1px solid #dfe5ee;vertical-align: top\">Structured data such as Excel; knowledge data such as Office documents<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p style=\"font-size: 0.85em;color: #607d8b;margin: -8px 0 20px\">*Chosen for this PoC\u2019s requirements and data structure. In real projects, the LLM and infrastructure are selected after reviewing data volume, security, the existing environment and required accuracy.<\/p>\n<h3 style=\"margin: 28px 0 12px;color: #0d47a1\">Architecture: combining SQL and RAG<\/h3>\n<p><img loading=\"lazy\" decoding=\"async\" width=\"1774\" height=\"887\" class=\"alignnone size-full wp-image-165232\" src=\"https:\/\/cdn.bap-software.net\/2026\/10\/22175023\/Anh-ChatGPT-19_49_52-22-thg-10-2026.png\" alt=\"\" \/><\/p>\n<p>Rather than sending every question through vector search, the PoC used an architecture that <strong>switches between SQL search and RAG search depending on the question<\/strong>.<\/p>\n<div style=\"display: flex;flex-wrap: wrap;gap: 10px;margin: 20px 0;align-items: stretch\">\n<div style=\"flex: 1 1 170px;background: #e3f2fd;border: 1px solid #90b4e6;border-radius: 8px;padding: 12px 14px\">\n<p style=\"margin: 0 0 4px;font-size: 0.8em;font-weight: bold;color: #0d47a1\">STEP 1<\/p>\n<p style=\"margin: 0 0 4px;font-weight: bold\">Data ingestion<\/p>\n<p style=\"margin: 0;font-size: 0.9em\">Office documents are parsed and chunked, with embeddings and metadata stored in a vector DB. Excel data is validated and loaded into PostgreSQL<\/p>\n<\/div>\n<div style=\"flex: 1 1 170px;background: #e8f5e9;border: 1px solid #a5d6a7;border-radius: 8px;padding: 12px 14px\">\n<p style=\"margin: 0 0 4px;font-size: 0.8em;font-weight: bold;color: #0d47a1\">STEP 2<\/p>\n<p style=\"margin: 0 0 4px;font-weight: bold\">Intent Router<\/p>\n<p style=\"margin: 0;font-size: 0.9em\">Classifies each question. Routine questions use predefined workflows; open-ended ones get an AI-generated execution plan<\/p>\n<\/div>\n<div style=\"flex: 1 1 170px;background: #e3f2fd;border: 1px solid #90b4e6;border-radius: 8px;padding: 12px 14px\">\n<p style=\"margin: 0 0 4px;font-size: 0.8em;font-weight: bold;color: #0d47a1\">STEP 3<\/p>\n<p style=\"margin: 0 0 4px;font-weight: bold\">SQL search<\/p>\n<p style=\"margin: 0;font-size: 0.9em\">The SQL Tool pulls exact figures \u2014 sales, plan, actuals \u2014 from PostgreSQL<\/p>\n<\/div>\n<div style=\"flex: 1 1 170px;background: #e8f5e9;border: 1px solid #a5d6a7;border-radius: 8px;padding: 12px 14px\">\n<p style=\"margin: 0 0 4px;font-size: 0.8em;font-weight: bold;color: #0d47a1\">STEP 4<\/p>\n<p style=\"margin: 0 0 4px;font-weight: bold\">RAG search<\/p>\n<p style=\"margin: 0;font-size: 0.9em\">The RAG Tool retrieves business rules and past documents via vector search, keeping source references<\/p>\n<\/div>\n<div style=\"flex: 1 1 170px;background: #e3f2fd;border: 1px solid #90b4e6;border-radius: 8px;padding: 12px 14px\">\n<p style=\"margin: 0 0 4px;font-size: 0.8em;font-weight: bold;color: #0d47a1\">STEP 5<\/p>\n<p style=\"margin: 0 0 4px;font-weight: bold\">Combine &amp; answer<\/p>\n<p style=\"margin: 0;font-size: 0.9em\">Results are merged into one context; Claude Sonnet 4.5 generates the answer and explanation, shown with sources<\/p>\n<\/div>\n<\/div>\n<p>For a question like \u201cShow the variance between plan and actuals and explain the causes based on past reports,\u201d the system pulls the figures from PostgreSQL, finds related past documents with RAG, combines both and generates the answer. This produced responses far better suited to management accounting than document-only RAG.<\/p>\n<h3 style=\"margin: 28px 0 12px;color: #0d47a1\">What we evaluated \u2014 and the results<\/h3>\n<p>The question was not \u201ccan the AI write text?\u201d but <strong>\u201ccan it be used for management accounting work?\u201d<\/strong> We checked whether it could retrieve historical figures correctly, find relevant documents, explain variances reasonably, make proposals based on statistical data, explain persuasively, produce usable report drafts, and output results staff could easily review and edit. AI output was never used directly for management decisions; staff verified the sources and edited as needed.<\/p>\n<div style=\"display: flex;flex-wrap: wrap;gap: 16px;margin: 20px 0\">\n<div style=\"flex: 1 1 150px;background: #f1f6fd;border: 1px solid #c5d8f3;border-radius: 8px;padding: 16px;text-align: center\">\n<p style=\"margin: 0;font-size: 1.6em;font-weight: bold;color: #0d47a1;line-height: 1.2\">8 weeks<\/p>\n<p style=\"margin: 6px 0 0;font-size: 0.9em\">PoC completed on schedule<\/p>\n<\/div>\n<div style=\"flex: 1 1 150px;background: #f1f6fd;border: 1px solid #c5d8f3;border-radius: 8px;padding: 16px;text-align: center\">\n<p style=\"margin: 0;font-size: 1.6em;font-weight: bold;color: #0d47a1;line-height: 1.2\">SQL + RAG<\/p>\n<p style=\"margin: 6px 0 0;font-size: 0.9em\">Structured data and documents unified<\/p>\n<\/div>\n<div style=\"flex: 1 1 150px;background: #f1f6fd;border: 1px solid #c5d8f3;border-radius: 8px;padding: 16px;text-align: center\">\n<p style=\"margin: 0;font-size: 1.6em;font-weight: bold;color: #0d47a1;line-height: 1.2\">Traceable<\/p>\n<p style=\"margin: 6px 0 0;font-size: 0.9em\">Every answer linked to its sources<\/p>\n<\/div>\n<div style=\"flex: 1 1 150px;background: #f1f6fd;border: 1px solid #c5d8f3;border-radius: 8px;padding: 16px;text-align: center\">\n<p style=\"margin: 0;font-size: 1.6em;font-weight: bold;color: #0d47a1;line-height: 1.2\">Delivered<\/p>\n<p style=\"margin: 6px 0 0;font-size: 0.9em\">PoC environment and documentation handed over<\/p>\n<\/div>\n<\/div>\n<p>The team validated the use case of drafting management reports and received positive feedback from the client.<\/p>\n<p style=\"font-size: 0.85em;color: #607d8b;margin: -8px 0 20px\">*Unverified metrics, such as effort reduction or answer accuracy rates, are not included in these results.<\/p>\n<h2 style=\"border-left: 6px solid #0d47a1;padding: 6px 0 6px 14px;margin: 44px 0 18px\">How to Judge Whether a RAG PoC Has Succeeded<\/h2>\n<p>\u201cThe AI works\u201d is not a success criterion. Beyond accuracy, check these conditions with production in mind:<\/p>\n<div style=\"display: flex;flex-wrap: wrap;gap: 16px;margin: 20px 0\">\n<div style=\"flex: 1 1 260px;background: #f1f6fd;border-top: 4px solid #0d47a1;border-radius: 8px;padding: 18px 22px\">\n<h3 style=\"color: #0d47a1;margin: 0 0 10px;font-size: 1.1em\">Quality and risk<\/h3>\n<ul style=\"margin: 0;padding-left: 20px\">\n<li style=\"margin: 6px 0\">Required documents and data can be retrieved<\/li>\n<li style=\"margin: 6px 0\">Answers are grounded and sources are visible to users<\/li>\n<li style=\"margin: 6px 0\">The conditions that cause wrong answers are understood<\/li>\n<li style=\"margin: 6px 0\">Unanswerable questions are handled properly<\/li>\n<li style=\"margin: 6px 0\">The scope of human review is clearly defined<\/li>\n<\/ul>\n<\/div>\n<div style=\"flex: 1 1 260px;background: #e8f5e9;border-top: 4px solid #2e7d32;border-radius: 8px;padding: 18px 22px\">\n<h3 style=\"color: #2e7d32;margin: 0 0 10px;font-size: 1.1em\">Business and operations<\/h3>\n<ul style=\"margin: 0;padding-left: 20px\">\n<li style=\"margin: 6px 0\">A clear benefit over current work (e.g. time saved) is expected<\/li>\n<li style=\"margin: 6px 0\">Production costs can be estimated and are acceptable<\/li>\n<li style=\"margin: 6px 0\">Data update methods and an operations owner are in place<\/li>\n<li style=\"margin: 6px 0\">Security requirements can be met<\/li>\n<\/ul>\n<\/div>\n<\/div>\n<div style=\"background: #fff3e0;border-left: 5px solid #e65100;border-radius: 0 8px 8px 0;padding: 16px 22px;margin: 22px 0\">\n<p style=\"margin: 0 0 6px;font-weight: bold;color: #e65100\">For high-risk use cases<\/p>\n<div>Never use AI answers as-is. Build human review and approval into the workflow.<\/div>\n<\/div>\n<h2><img loading=\"lazy\" decoding=\"async\" width=\"1536\" height=\"1024\" class=\"alignnone size-full wp-image-165233\" src=\"https:\/\/cdn.bap-software.net\/2026\/10\/22175052\/Anh-ChatGPT-19_49_59-22-thg-10-2026.png\" alt=\"\" \/><\/h2>\n<h2 style=\"border-left: 6px solid #0d47a1;padding: 6px 0 6px 14px;margin: 44px 0 18px\">What to Prepare Before Requesting a Quote<\/h2>\n<p>Having the following ready before you talk to a RAG development partner makes it much easier to scope the PoC and get an accurate estimate:<\/p>\n<p style=\"margin: 12px 0 20px;line-height: 2.4\"><span style=\"display: inline-block;background: #f1f6fd;border: 1px solid #c5d8f3;color: #0d47a1;border-radius: 16px;padding: 2px 12px;margin: 0 6px 0 0;font-size: 0.9em;line-height: 1.8\">Business problem to solve<\/span> <span style=\"display: inline-block;background: #f1f6fd;border: 1px solid #c5d8f3;color: #0d47a1;border-radius: 16px;padding: 2px 12px;margin: 0 6px 0 0;font-size: 0.9em;line-height: 1.8\">Target users<\/span> <span style=\"display: inline-block;background: #f1f6fd;border: 1px solid #c5d8f3;color: #0d47a1;border-radius: 16px;padding: 2px 12px;margin: 0 6px 0 0;font-size: 0.9em;line-height: 1.8\">Documents and databases in scope<\/span> <span style=\"display: inline-block;background: #f1f6fd;border: 1px solid #c5d8f3;color: #0d47a1;border-radius: 16px;padding: 2px 12px;margin: 0 6px 0 0;font-size: 0.9em;line-height: 1.8\">Data formats and locations<\/span> <span style=\"display: inline-block;background: #f1f6fd;border: 1px solid #c5d8f3;color: #0d47a1;border-radius: 16px;padding: 2px 12px;margin: 0 6px 0 0;font-size: 0.9em;line-height: 1.8\">Real example questions<\/span> <span style=\"display: inline-block;background: #f1f6fd;border: 1px solid #c5d8f3;color: #0d47a1;border-radius: 16px;padding: 2px 12px;margin: 0 6px 0 0;font-size: 0.9em;line-height: 1.8\">Expected answers<\/span> <span style=\"display: inline-block;background: #f1f6fd;border: 1px solid #c5d8f3;color: #0d47a1;border-radius: 16px;padding: 2px 12px;margin: 0 6px 0 0;font-size: 0.9em;line-height: 1.8\">Access requirements<\/span> <span style=\"display: inline-block;background: #f1f6fd;border: 1px solid #c5d8f3;color: #0d47a1;border-radius: 16px;padding: 2px 12px;margin: 0 6px 0 0;font-size: 0.9em;line-height: 1.8\">Systems to integrate<\/span> <span style=\"display: inline-block;background: #f1f6fd;border: 1px solid #c5d8f3;color: #0d47a1;border-radius: 16px;padding: 2px 12px;margin: 0 6px 0 0;font-size: 0.9em;line-height: 1.8\">Security requirements<\/span> <span style=\"display: inline-block;background: #f1f6fd;border: 1px solid #c5d8f3;color: #0d47a1;border-radius: 16px;padding: 2px 12px;margin: 0 6px 0 0;font-size: 0.9em;line-height: 1.8\">PoC success metrics<\/span> <span style=\"display: inline-block;background: #f1f6fd;border: 1px solid #c5d8f3;color: #0d47a1;border-radius: 16px;padding: 2px 12px;margin: 0 6px 0 0;font-size: 0.9em;line-height: 1.8\">Target timeline<\/span> <span style=\"display: inline-block;background: #f1f6fd;border: 1px solid #c5d8f3;color: #0d47a1;border-radius: 16px;padding: 2px 12px;margin: 0 6px 0 0;font-size: 0.9em;line-height: 1.8\">Post-launch operating model<\/span><\/p>\n<p>You do not need every answer up front. A clear business problem, where the data lives and a few real questions are enough to scope the first validation.<\/p>\n<h2 style=\"border-left: 6px solid #0d47a1;padding: 6px 0 6px 14px;margin: 44px 0 18px\">Frequently Asked Questions<\/h2>\n<div style=\"margin: 20px 0\">\n<div style=\"border: 1px solid #dfe5ee;border-radius: 8px;padding: 14px 20px;margin: 0 0 12px\">\n<h3 style=\"margin: 0 0 6px;font-size: 1em;color: #0d47a1\">Q. Can a RAG PoC be done in 2\u20134 weeks?<\/h3>\n<p style=\"margin: 0\">A. Yes, a 2\u20134 week technical validation is possible when the use case, data and questions are clear. Evaluation with real business data, multiple integrations or detailed accuracy tuning usually takes 1\u20133 months.<\/p>\n<\/div>\n<div style=\"border: 1px solid #dfe5ee;border-radius: 8px;padding: 14px 20px;margin: 0 0 12px\">\n<h3 style=\"margin: 0 0 6px;font-size: 1em;color: #0d47a1\">Q. Can RAG use Excel data?<\/h3>\n<p style=\"margin: 0\">A. Yes. When figures must be aggregated precisely, it is best to load Excel data into a database and query it with SQL. Combined with RAG for documents, the system can answer using both figures and explanatory material.<\/p>\n<\/div>\n<div style=\"border: 1px solid #dfe5ee;border-radius: 8px;padding: 14px 20px;margin: 0 0 12px\">\n<h3 style=\"margin: 0 0 6px;font-size: 1em;color: #0d47a1\">Q. Will RAG eliminate wrong answers?<\/h3>\n<p style=\"margin: 0\">A. Not entirely. It is combined with safeguards such as limiting source data, citing sources in answers, handling unanswerable questions, rule-based checks and human review.<\/p>\n<\/div>\n<div style=\"border: 1px solid #dfe5ee;border-radius: 8px;padding: 14px 20px;margin: 0 0 12px\">\n<h3 style=\"margin: 0 0 6px;font-size: 1em;color: #0d47a1\">Q. Which LLM should we choose?<\/h3>\n<p style=\"margin: 0\">A. Look beyond accuracy to data handling, pricing, latency, language support and fit with your cloud environment. BAP\u2019s management accounting PoC used Claude Sonnet 4.5 on Amazon Bedrock.<\/p>\n<\/div>\n<div style=\"border: 1px solid #dfe5ee;border-radius: 8px;padding: 14px 20px;margin: 0 0 12px\">\n<h3 style=\"margin: 0 0 6px;font-size: 1em;color: #0d47a1\">Q. Is ongoing maintenance needed after launch?<\/h3>\n<p style=\"margin: 0\">A. Yes. Documents and data need updating, answer quality needs checking, retrieval settings need tuning, and API costs and model changes need managing on an ongoing basis.<\/p>\n<\/div>\n<\/div>\n<h2 style=\"border-left: 6px solid #0d47a1;padding: 6px 0 6px 14px;margin: 44px 0 18px\">Summary<\/h2>\n<div style=\"background: #f5f7f9;border-left: 5px solid #37474f;border-radius: 0 8px 8px 0;padding: 16px 22px;margin: 22px 0\">\n<div>\n<ul style=\"margin: 0;padding-left: 20px\">\n<li style=\"margin: 6px 0\">RAG timelines and costs are driven by the use case, data types, accuracy, permissions, integrations and operations \u2014 not by document count<\/li>\n<li style=\"margin: 6px 0\">Expect 2\u20134 weeks for technical validation and 1\u20133 months for a business PoC<\/li>\n<li style=\"margin: 6px 0\">In BAP\u2019s case study, a PoC combining SQL search and RAG search was completed in 8 weeks<\/li>\n<\/ul>\n<\/div>\n<\/div>\n<p>Start by defining the use case and evaluation criteria, run a PoC on a limited data set, and use the results to decide on the features, budget and operating model needed for production. That is the most practical path to RAG in production.<\/p>\n<div style=\"background: #f5f7f9;border-radius: 8px;padding: 16px 22px;margin: 24px 0\">\n<p style=\"margin: 0 0 8px;font-weight: bold\">References<\/p>\n<ul style=\"margin: 0;padding-left: 20px\">\n<li style=\"margin: 6px 0\"><a href=\"https:\/\/learn.microsoft.com\/en-us\/azure\/architecture\/ai-ml\/guide\/rag\/rag-solution-design-and-evaluation-guide\" target=\"_blank\" rel=\"noopener\">Microsoft Learn: Design and develop a RAG solution<\/a><\/li>\n<li style=\"margin: 6px 0\"><a href=\"https:\/\/docs.aws.amazon.com\/nova\/latest\/userguide\/rag-systems.html\" target=\"_blank\" rel=\"noopener\">AWS Documentation: Building a RAG system with Amazon Nova<\/a><\/li>\n<\/ul>\n<\/div>\n<div style=\"background: linear-gradient(135deg,#0d47a1,#1976d2);color: #fff;border-radius: 10px;padding: 26px 28px;margin: 32px 0\">\n<p style=\"margin: 0 0 10px;font-size: 1.2em;font-weight: bold;color: #fff;text-align: center\">Talk to BAP About RAG and Generative AI PoCs<\/p>\n<p style=\"margin: 0 0 12px;color: #fff\">BAP supports you end to end \u2014 from clarifying the business problem and running a RAG PoC to production development, integration with existing systems and post-launch improvement. We also build AI solutions that combine internal documents with Excel, databases and existing business systems. Reach out if you want to:<\/p>\n<ul style=\"margin: 0 0 16px;padding-left: 20px;color: #fff\">\n<li style=\"margin: 6px 0;color: #fff\">Decide which process to start RAG with<\/li>\n<li style=\"margin: 6px 0;color: #fff\">Use internal documents and numerical data together<\/li>\n<li style=\"margin: 6px 0;color: #fff\">Confirm the timeline and team for a PoC<\/li>\n<li style=\"margin: 6px 0;color: #fff\">Embed generative AI into existing systems<\/li>\n<li style=\"margin: 6px 0;color: #fff\">Move a generative AI project stuck at PoC into production<\/li>\n<\/ul>\n<p style=\"margin: 0;text-align: center\"><a style=\"display: inline-block;background: #fff;color: #0d47a1;font-weight: bold;padding: 12px 28px;border-radius: 24px;text-decoration: none\" href=\"https:\/\/bap-software.net\/contact\/\">Discuss your RAG project<\/a><\/p>\n<p style=\"margin: 12px 0 0;font-size: 0.85em;color: #e3f2fd;text-align: center\">Even without a defined use case, we can help you scope the process, data and PoC criteria. See <a style=\"color: #fff;text-decoration: underline\" href=\"https:\/\/bap-software.net\/ai-developer-services\/\">BAP\u2019s generative AI and RAG development services<\/a> for details.<\/p>\n<\/div>\n<p><\/p>","protected":false},"author":43,"featured_media":165231,"template":"","meta":{"_acf_changed":false},"tags":[],"blog-cat":[2058],"class_list":["post-164933","knowledge","type-knowledge","status-publish","has-post-thumbnail","hentry","blog-cat-technology"],"acf":[],"yoast_head":"<!-- This site is optimized with the Yoast SEO Premium plugin v26.1 (Yoast SEO v28.4) - https:\/\/yoast.com\/product\/yoast-seo-premium-wordpress\/ -->\n<title>(\u65e5\u672c\u8a9e) RAG\u69cb\u7bc9\u306e\u9032\u3081\u65b9\u30fb\u8cbb\u7528\u30fb\u671f\u9593\uff5c8\u9031\u9593\u306ePoC\u4e8b\u4f8b\u304b\u3089\u89e3\u8aac | BAP<\/title>\n<meta name=\"description\" content=\"(\u65e5\u672c\u8a9e) RAG\u69cb\u7bc9\u306b\u5fc5\u8981\u306a\u8cbb\u7528\u30fb\u671f\u9593\u30fb\u9032\u3081\u65b9\u3092\u89e3\u8aac\u3002\u65e5\u672c\u4f01\u696d\u5411\u3051\u306b8\u9031\u9593\u3067\u5b9f\u65bd\u3057\u305fBAP\u306ePoC\u4e8b\u4f8b\u3092\u57fa\u306b\u3001SQL\u3068RAG\u3092\u7d44\u307f\u5408\u308f\u305b\u305f\u69cb\u6210\u3001\u8a55\u4fa1\u9805\u76ee\u3001\u672c\u756a\u5c0e\u5165\u306e\u30dd\u30a4\u30f3\u30c8\u3092\u7d39\u4ecb\u3057\u307e\u3059\u3002\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/bap-software.net\/en\/knowledge\/rag-development-process-cost\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"How to Build a RAG System: Process, Cost and Timeline \u2014 Lessons from an 8-Week PoC\" \/>\n<meta property=\"og:description\" content=\"(\u65e5\u672c\u8a9e) RAG\u69cb\u7bc9\u306b\u5fc5\u8981\u306a\u8cbb\u7528\u30fb\u671f\u9593\u30fb\u9032\u3081\u65b9\u3092\u89e3\u8aac\u3002\u65e5\u672c\u4f01\u696d\u5411\u3051\u306b8\u9031\u9593\u3067\u5b9f\u65bd\u3057\u305fBAP\u306ePoC\u4e8b\u4f8b\u3092\u57fa\u306b\u3001SQL\u3068RAG\u3092\u7d44\u307f\u5408\u308f\u305b\u305f\u69cb\u6210\u3001\u8a55\u4fa1\u9805\u76ee\u3001\u672c\u756a\u5c0e\u5165\u306e\u30dd\u30a4\u30f3\u30c8\u3092\u7d39\u4ecb\u3057\u307e\u3059\u3002\" \/>\n<meta property=\"og:url\" content=\"https:\/\/bap-software.net\/en\/knowledge\/rag-development-process-cost\/\" \/>\n<meta property=\"og:site_name\" content=\"Software development - offshore service | BAP Software\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/bap32\" \/>\n<meta property=\"og:image\" content=\"https:\/\/cdn.bap-software.net\/2026\/10\/22174433\/Anh-ChatGPT-19_35_22-22-thg-10-2026.png\" \/>\n\t<meta property=\"og:image:width\" content=\"1672\" \/>\n\t<meta property=\"og:image:height\" content=\"941\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/png\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:site\" content=\"@bapsoftware\" \/>\n<meta name=\"twitter:label1\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data1\" content=\"41 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/bap-software.net\\\/knowledge\\\/rag-development-process-cost\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/bap-software.net\\\/knowledge\\\/rag-development-process-cost\\\/\"},\"author\":{\"name\":\"Mennt\",\"@id\":\"https:\\\/\\\/bap-software.net\\\/#\\\/schema\\\/person\\\/cca627b465d7ed089edcc53e6aaf2f6a\"},\"headline\":\"How to Build a RAG System: Process, Cost and Timeline \u2014 Lessons from an 8-Week PoC\",\"datePublished\":\"2026-10-22T13:37:57+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/bap-software.net\\\/knowledge\\\/rag-development-process-cost\\\/\"},\"wordCount\":8302,\"image\":{\"@id\":\"https:\\\/\\\/bap-software.net\\\/knowledge\\\/rag-development-process-cost\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/cdn.bap-software.net\\\/2026\\\/10\\\/22174433\\\/Anh-ChatGPT-19_35_22-22-thg-10-2026.png\",\"inLanguage\":\"en-US\"},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/bap-software.net\\\/en\\\/knowledge\\\/rag-development-process-cost\\\/\",\"url\":\"https:\\\/\\\/bap-software.net\\\/en\\\/knowledge\\\/rag-development-process-cost\\\/\",\"name\":\"(\u65e5\u672c\u8a9e) RAG\u69cb\u7bc9\u306e\u9032\u3081\u65b9\u30fb\u8cbb\u7528\u30fb\u671f\u9593\uff5c8\u9031\u9593\u306ePoC\u4e8b\u4f8b\u304b\u3089\u89e3\u8aac | BAP\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/bap-software.net\\\/en\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/bap-software.net\\\/en\\\/knowledge\\\/rag-development-process-cost\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/bap-software.net\\\/knowledge\\\/rag-development-process-cost\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/cdn.bap-software.net\\\/2026\\\/10\\\/22174433\\\/Anh-ChatGPT-19_35_22-22-thg-10-2026.png\",\"datePublished\":\"2026-10-22T13:37:57+00:00\",\"description\":\"(\u65e5\u672c\u8a9e) RAG\u69cb\u7bc9\u306b\u5fc5\u8981\u306a\u8cbb\u7528\u30fb\u671f\u9593\u30fb\u9032\u3081\u65b9\u3092\u89e3\u8aac\u3002\u65e5\u672c\u4f01\u696d\u5411\u3051\u306b8\u9031\u9593\u3067\u5b9f\u65bd\u3057\u305fBAP\u306ePoC\u4e8b\u4f8b\u3092\u57fa\u306b\u3001SQL\u3068RAG\u3092\u7d44\u307f\u5408\u308f\u305b\u305f\u69cb\u6210\u3001\u8a55\u4fa1\u9805\u76ee\u3001\u672c\u756a\u5c0e\u5165\u306e\u30dd\u30a4\u30f3\u30c8\u3092\u7d39\u4ecb\u3057\u307e\u3059\u3002\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/bap-software.net\\\/en\\\/knowledge\\\/rag-development-process-cost\\\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[[\"https:\\\/\\\/bap-software.net\\\/en\\\/knowledge\\\/rag-development-process-cost\\\/\"]]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/bap-software.net\\\/en\\\/knowledge\\\/rag-development-process-cost\\\/#primaryimage\",\"url\":\"https:\\\/\\\/cdn.bap-software.net\\\/2026\\\/10\\\/22174433\\\/Anh-ChatGPT-19_35_22-22-thg-10-2026.png\",\"contentUrl\":\"https:\\\/\\\/cdn.bap-software.net\\\/2026\\\/10\\\/22174433\\\/Anh-ChatGPT-19_35_22-22-thg-10-2026.png\",\"width\":1672,\"height\":941},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/bap-software.net\\\/en\\\/knowledge\\\/rag-development-process-cost\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/bap-software.net\\\/en\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Knowledge\",\"item\":\"https:\\\/\\\/bap-software.net\\\/en\\\/knowledge\\\/\"},{\"@type\":\"ListItem\",\"position\":3,\"name\":\"RAG\u69cb\u7bc9\u306e\u9032\u3081\u65b9\u30fb\u8cbb\u7528\u30fb\u671f\u9593\uff5c8\u9031\u9593\u306ePoC\u4e8b\u4f8b\u304b\u3089\u89e3\u8aac\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/bap-software.net\\\/en\\\/#website\",\"url\":\"https:\\\/\\\/bap-software.net\\\/en\\\/\",\"name\":\"Software development - offshore service | BAP Software\",\"description\":\"\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/bap-software.net\\\/en\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/bap-software.net\\\/#\\\/schema\\\/person\\\/cca627b465d7ed089edcc53e6aaf2f6a\",\"name\":\"Mennt\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/3fe8294ef1ef284d6d33e6251a7a8f54353509b00f291de8495800f21a6b3949?s=96&d=identicon&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/3fe8294ef1ef284d6d33e6251a7a8f54353509b00f291de8495800f21a6b3949?s=96&d=identicon&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/3fe8294ef1ef284d6d33e6251a7a8f54353509b00f291de8495800f21a6b3949?s=96&d=identicon&r=g\",\"caption\":\"Mennt\"},\"url\":\"https:\\\/\\\/bap-software.net\\\/en\\\/author\\\/mennt\\\/\"}]}<\/script>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"(\u65e5\u672c\u8a9e) RAG\u69cb\u7bc9\u306e\u9032\u3081\u65b9\u30fb\u8cbb\u7528\u30fb\u671f\u9593\uff5c8\u9031\u9593\u306ePoC\u4e8b\u4f8b\u304b\u3089\u89e3\u8aac | BAP","description":"(\u65e5\u672c\u8a9e) RAG\u69cb\u7bc9\u306b\u5fc5\u8981\u306a\u8cbb\u7528\u30fb\u671f\u9593\u30fb\u9032\u3081\u65b9\u3092\u89e3\u8aac\u3002\u65e5\u672c\u4f01\u696d\u5411\u3051\u306b8\u9031\u9593\u3067\u5b9f\u65bd\u3057\u305fBAP\u306ePoC\u4e8b\u4f8b\u3092\u57fa\u306b\u3001SQL\u3068RAG\u3092\u7d44\u307f\u5408\u308f\u305b\u305f\u69cb\u6210\u3001\u8a55\u4fa1\u9805\u76ee\u3001\u672c\u756a\u5c0e\u5165\u306e\u30dd\u30a4\u30f3\u30c8\u3092\u7d39\u4ecb\u3057\u307e\u3059\u3002","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/bap-software.net\/en\/knowledge\/rag-development-process-cost\/","og_locale":"en_US","og_type":"article","og_title":"How to Build a RAG System: Process, Cost and Timeline \u2014 Lessons from an 8-Week PoC","og_description":"(\u65e5\u672c\u8a9e) RAG\u69cb\u7bc9\u306b\u5fc5\u8981\u306a\u8cbb\u7528\u30fb\u671f\u9593\u30fb\u9032\u3081\u65b9\u3092\u89e3\u8aac\u3002\u65e5\u672c\u4f01\u696d\u5411\u3051\u306b8\u9031\u9593\u3067\u5b9f\u65bd\u3057\u305fBAP\u306ePoC\u4e8b\u4f8b\u3092\u57fa\u306b\u3001SQL\u3068RAG\u3092\u7d44\u307f\u5408\u308f\u305b\u305f\u69cb\u6210\u3001\u8a55\u4fa1\u9805\u76ee\u3001\u672c\u756a\u5c0e\u5165\u306e\u30dd\u30a4\u30f3\u30c8\u3092\u7d39\u4ecb\u3057\u307e\u3059\u3002","og_url":"https:\/\/bap-software.net\/en\/knowledge\/rag-development-process-cost\/","og_site_name":"Software development - offshore service | BAP Software","article_publisher":"https:\/\/www.facebook.com\/bap32","og_image":[{"width":1672,"height":941,"url":"https:\/\/cdn.bap-software.net\/2026\/10\/22174433\/Anh-ChatGPT-19_35_22-22-thg-10-2026.png","type":"image\/png"}],"twitter_card":"summary_large_image","twitter_site":"@bapsoftware","twitter_misc":{"Est. reading time":"41 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/bap-software.net\/knowledge\/rag-development-process-cost\/#article","isPartOf":{"@id":"https:\/\/bap-software.net\/knowledge\/rag-development-process-cost\/"},"author":{"name":"Mennt","@id":"https:\/\/bap-software.net\/#\/schema\/person\/cca627b465d7ed089edcc53e6aaf2f6a"},"headline":"How to Build a RAG System: Process, Cost and Timeline \u2014 Lessons from an 8-Week PoC","datePublished":"2026-10-22T13:37:57+00:00","mainEntityOfPage":{"@id":"https:\/\/bap-software.net\/knowledge\/rag-development-process-cost\/"},"wordCount":8302,"image":{"@id":"https:\/\/bap-software.net\/knowledge\/rag-development-process-cost\/#primaryimage"},"thumbnailUrl":"https:\/\/cdn.bap-software.net\/2026\/10\/22174433\/Anh-ChatGPT-19_35_22-22-thg-10-2026.png","inLanguage":"en-US"},{"@type":"WebPage","@id":"https:\/\/bap-software.net\/en\/knowledge\/rag-development-process-cost\/","url":"https:\/\/bap-software.net\/en\/knowledge\/rag-development-process-cost\/","name":"(\u65e5\u672c\u8a9e) RAG\u69cb\u7bc9\u306e\u9032\u3081\u65b9\u30fb\u8cbb\u7528\u30fb\u671f\u9593\uff5c8\u9031\u9593\u306ePoC\u4e8b\u4f8b\u304b\u3089\u89e3\u8aac | BAP","isPartOf":{"@id":"https:\/\/bap-software.net\/en\/#website"},"primaryImageOfPage":{"@id":"https:\/\/bap-software.net\/en\/knowledge\/rag-development-process-cost\/#primaryimage"},"image":{"@id":"https:\/\/bap-software.net\/knowledge\/rag-development-process-cost\/#primaryimage"},"thumbnailUrl":"https:\/\/cdn.bap-software.net\/2026\/10\/22174433\/Anh-ChatGPT-19_35_22-22-thg-10-2026.png","datePublished":"2026-10-22T13:37:57+00:00","description":"(\u65e5\u672c\u8a9e) RAG\u69cb\u7bc9\u306b\u5fc5\u8981\u306a\u8cbb\u7528\u30fb\u671f\u9593\u30fb\u9032\u3081\u65b9\u3092\u89e3\u8aac\u3002\u65e5\u672c\u4f01\u696d\u5411\u3051\u306b8\u9031\u9593\u3067\u5b9f\u65bd\u3057\u305fBAP\u306ePoC\u4e8b\u4f8b\u3092\u57fa\u306b\u3001SQL\u3068RAG\u3092\u7d44\u307f\u5408\u308f\u305b\u305f\u69cb\u6210\u3001\u8a55\u4fa1\u9805\u76ee\u3001\u672c\u756a\u5c0e\u5165\u306e\u30dd\u30a4\u30f3\u30c8\u3092\u7d39\u4ecb\u3057\u307e\u3059\u3002","breadcrumb":{"@id":"https:\/\/bap-software.net\/en\/knowledge\/rag-development-process-cost\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":[["https:\/\/bap-software.net\/en\/knowledge\/rag-development-process-cost\/"]]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/bap-software.net\/en\/knowledge\/rag-development-process-cost\/#primaryimage","url":"https:\/\/cdn.bap-software.net\/2026\/10\/22174433\/Anh-ChatGPT-19_35_22-22-thg-10-2026.png","contentUrl":"https:\/\/cdn.bap-software.net\/2026\/10\/22174433\/Anh-ChatGPT-19_35_22-22-thg-10-2026.png","width":1672,"height":941},{"@type":"BreadcrumbList","@id":"https:\/\/bap-software.net\/en\/knowledge\/rag-development-process-cost\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/bap-software.net\/en\/"},{"@type":"ListItem","position":2,"name":"Knowledge","item":"https:\/\/bap-software.net\/en\/knowledge\/"},{"@type":"ListItem","position":3,"name":"RAG\u69cb\u7bc9\u306e\u9032\u3081\u65b9\u30fb\u8cbb\u7528\u30fb\u671f\u9593\uff5c8\u9031\u9593\u306ePoC\u4e8b\u4f8b\u304b\u3089\u89e3\u8aac"}]},{"@type":"WebSite","@id":"https:\/\/bap-software.net\/en\/#website","url":"https:\/\/bap-software.net\/en\/","name":"Software development - offshore service | BAP Software","description":"","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/bap-software.net\/en\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Person","@id":"https:\/\/bap-software.net\/#\/schema\/person\/cca627b465d7ed089edcc53e6aaf2f6a","name":"Mennt","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/secure.gravatar.com\/avatar\/3fe8294ef1ef284d6d33e6251a7a8f54353509b00f291de8495800f21a6b3949?s=96&d=identicon&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/3fe8294ef1ef284d6d33e6251a7a8f54353509b00f291de8495800f21a6b3949?s=96&d=identicon&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/3fe8294ef1ef284d6d33e6251a7a8f54353509b00f291de8495800f21a6b3949?s=96&d=identicon&r=g","caption":"Mennt"},"url":"https:\/\/bap-software.net\/en\/author\/mennt\/"}]}},"_links":{"self":[{"href":"https:\/\/bap-software.net\/en\/wp-json\/wp\/v2\/knowledge\/164933","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/bap-software.net\/en\/wp-json\/wp\/v2\/knowledge"}],"about":[{"href":"https:\/\/bap-software.net\/en\/wp-json\/wp\/v2\/types\/knowledge"}],"author":[{"embeddable":true,"href":"https:\/\/bap-software.net\/en\/wp-json\/wp\/v2\/users\/43"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/bap-software.net\/en\/wp-json\/wp\/v2\/media\/165231"}],"wp:attachment":[{"href":"https:\/\/bap-software.net\/en\/wp-json\/wp\/v2\/media?parent=164933"}],"wp:term":[{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/bap-software.net\/en\/wp-json\/wp\/v2\/tags?post=164933"},{"taxonomy":"blog-cat","embeddable":true,"href":"https:\/\/bap-software.net\/en\/wp-json\/wp\/v2\/blog-cat?post=164933"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}