Your team has:<\/p>\n\n\n\n
- \n
- 500+ page user manuals<\/li>\n<\/ul>\n\n\n\n
- \n
- Multiple product versions<\/li>\n<\/ul>\n\n\n\n
- \n
- Role-specific guides (admin, sales, manager)<\/li>\n<\/ul>\n\n\n\n
- \n
- API documentation<\/li>\n<\/ul>\n\n\n\n
- \n
- Troubleshooting FAQs<\/li>\n<\/ul>\n\n\n\n
- \n
- Video tutorials<\/li>\n<\/ul>\n\n\n\n
- \n
- Release notes<\/li>\n<\/ul>\n\n\n\n
When a user needs help, they spend 15-30 minutes<\/strong> searching through PDFs, switching between tabs, and piecing together information from multiple sources. Often, they give up and create a support ticket.<\/p>\n\n\n\n
The impact?<\/strong> Support teams are overwhelmed with routine questions that should be self-service, while customers waste valuable time searching for information that’s technically available but practically inaccessible.<\/p>\n\n\n\n
\n\n\n\nUnderstanding RAG vs Agentic RAG<\/h2>\n\n\n\n
What is RAG?<\/h3>\n\n\n\n
RAG (Retrieval-Augmented Generation)<\/strong> combines document search with AI language models to answer questions based on your documentation. It’s the foundation of modern AI assistants.<\/p>\n\n\n\n
But there’s a significant difference between simple RAG<\/strong> and Agentic RAG<\/strong>.<\/p>\n\n\n\n
\n\n\n\nTraditional RAG: Simple But Limited<\/h2>\n\n\n\n
Traditional RAG follows a straightforward, linear process:<\/p>\n\n\n
\n![\"\"](\"https:\/\/www.ucssolutions.com\/blog\/wp-content\/uploads\/2025\/12\/Screenshot_2025-12-05_at_08.41.55-354x1024.png\")
<\/figure><\/div>\n\n\nTraditional RAG Architecture<\/h3>\n\n\n\n
TRADITIONAL RAG\n------------------------------\n\nUser\n \u2193\n\"How to login to SalesWorx?\"\n \u2193\n\nEmbedding Model\n\u2022 Convert query to vector\n \u2193\n\nVector Database\n\u2022 Search similar documents\n\u2022 Return top 5-10 results\n \u2193\n\nCombine Context\n\u2022 Concatenate all documents\n \u2193\n\nLLM (Single Call)\n\u2022 Generate answer\n \u2193\n\nAnswer (no citations)<\/code><\/pre>\n\n\n\nProblems with Traditional RAG<\/h3>\n\n\n\n
\u274c Single search pass<\/strong> – Misses context if first search is poor<\/p>\n\n\n\n
\u274c No self-correction<\/strong> – Wrong search = wrong answer<\/p>\n\n\n\n
\u274c No citations<\/strong> – Users can’t verify information<\/p>\n\n\n\n
\u274c Can’t handle complex questions<\/strong> – No decomposition<\/p>\n\n\n\n
\u274c One-size-fits-all<\/strong> – Same approach for all queries<\/p>\n\n\n\n
\u274c No reasoning<\/strong> – Just retrieval + generation<\/p>\n\n\n\n
Example Problem:<\/strong><\/p>\n\n\n\n
Query: \"How to configure multi-currency reporting?\"\n\nTraditional RAG:\n\u251c\u2500 Search: \"multi-currency reporting\"\n\u251c\u2500 Finds: 5 documents about reporting (misses currency setup)\n\u2514\u2500 Answer: Incomplete (only covers reporting basics)\n\nResult: User still confused, creates support ticket\n<\/code><\/pre>\n\n\n\n
\n\n\n\nAgentic RAG: Intelligent Multi-Agent Research System<\/h2>\n\n\n\n
Agentic RAG<\/strong> transforms the process by introducing multiple specialized agents<\/strong> that work together, plan research strategies, and self-correct their approach.<\/p>\n\n\n
\n![\"\"](\"https:\/\/www.ucssolutions.com\/blog\/wp-content\/uploads\/2025\/12\/Screenshot_2025-12-05_at_09.44.37-711x1024.png\")
<\/figure><\/div>\n\n\nAgentic RAG Architecture: 7 Specialized Agents<\/h3>\n\n\n\n
Our system uses 7 specialized AI agents<\/strong> working together through a state machine, with each agent handling a specific aspect of intelligence:<\/p>\n\n\n\n
Agent<\/th> Role<\/th> What It Does<\/th><\/tr><\/thead> Planner<\/strong><\/td> Research Strategist<\/td> Validates domain relevance, decomposes complex questions into sub-tasks<\/td><\/tr> Query Rewriter<\/strong><\/td> Search Optimizer<\/td> Reformulates queries for better retrieval using context and keywords<\/td><\/tr> Retrieval Supervisor<\/strong><\/td> Strategy Selector<\/td> Chooses vector\/keyword\/hybrid search based on query characteristics<\/td><\/tr> Context Distiller<\/strong><\/td> Information Synthesizer<\/td> Compresses and removes redundancy from retrieved documents<\/td><\/tr> Reflection<\/strong><\/td> Progress Tracker<\/td> Summarizes findings after each research step<\/td><\/tr> Policy<\/strong><\/td> Decision Maker<\/td> Evaluates if enough information is gathered or more research is needed<\/td><\/tr> Answer Generator<\/strong><\/td> Response Composer<\/td> Creates comprehensive answers with inline citations<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n Complete Agentic RAG Workflow<\/h3>\n\n\n
\n![\"\"](\"https:\/\/www.ucssolutions.com\/blog\/wp-content\/uploads\/2025\/12\/Screenshot_2025-12-05_at_08.47.08-830x1024.png\")
<\/figure><\/div>\n\n\n
\n\n\n\nKey Advantages of Agentic RAG<\/h2>\n\n\n\n
1. Intelligent Domain Validation<\/strong><\/h3>\n\n\n
\n![\"\"](\"https:\/\/www.ucssolutions.com\/blog\/wp-content\/uploads\/2025\/12\/Screenshot_2025-12-05_at_09.06.24-1024x868.png\")
<\/figure><\/div>\n\n\nImpact:<\/strong> Saves resources on irrelevant queries (typically 20-30% of all interactions)<\/p>\n\n\n\n
2. Adaptive Research Strategies<\/strong><\/h3>\n\n\n
\n![\"\"](\"https:\/\/www.ucssolutions.com\/blog\/wp-content\/uploads\/2025\/12\/Screenshot_2025-12-05_at_08.48.10-1024x419.png\")
<\/figure><\/div>\n\n\nThree Search Strategies (Agent 3: Retrieval Supervisor):<\/strong><\/p>\n\n\n\n
Strategy<\/th> When Used<\/th> Example Query<\/th><\/tr><\/thead> Vector Search<\/strong><\/td> Conceptual, semantic queries<\/td> “How does inventory forecasting work?”<\/td><\/tr> Keyword Search<\/strong><\/td> Exact terms, codes, IDs<\/td> “Error code 500”<\/td><\/tr> Hybrid Search<\/strong><\/td> Most queries (default)<\/td> “How to login to SalesWorx?”<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n 3. Self-Correcting Research Loops<\/strong><\/h3>\n\n\n
\n![\"\"](\"https:\/\/www.ucssolutions.com\/blog\/wp-content\/uploads\/2025\/12\/Screenshot_2025-12-05_at_09.07.29-1024x835.png\")
<\/figure><\/div>\n\n\n4. Context Distillation<\/strong><\/h3>\n\n\n
\n![\"\"](\"https:\/\/www.ucssolutions.com\/blog\/wp-content\/uploads\/2025\/12\/Screenshot_2025-12-05_at_09.07.53-1024x857.png\")
<\/figure><\/div>\n\n\n5. Trustworthy Citations<\/strong><\/h3>\n\n\n
\n![\"\"](\"https:\/\/www.ucssolutions.com\/blog\/wp-content\/uploads\/2025\/12\/Screenshot_2025-12-05_at_09.08.23-1024x1007.png\")
<\/figure><\/div>\n\n\n6. Intelligent Web Crawling with Crawl4AI<\/strong><\/h3>\n\n\n\n
Beyond searching internal documents, Agentic RAG can also ingest content from your website and documentation portals using Crawl4AI<\/strong> – a privacy-first web crawling solution.<\/p>\n\n\n
\n![\"\"](\"https:\/\/www.ucssolutions.com\/blog\/wp-content\/uploads\/2025\/12\/Screenshot_2025-12-05_at_09.09.07-1024x657.png\")
<\/figure><\/div>\n\n\nHow Crawl4AI Works:<\/strong><\/p>\n\n\n
\n![\"\"](\"https:\/\/www.ucssolutions.com\/blog\/wp-content\/uploads\/2025\/12\/Screenshot_2025-12-05_at_08.49.00-326x1024.png\")
<\/figure><\/div>\n\n\nExample: Ingesting SalesWorx Documentation Website<\/strong><\/p>\n\n\n\n
Step 1: Provide Sitemap URL\n\u2514\u2500 https:\/\/docs.salesworx.ae\/wp-sitemap-posts-1.xml\n\n\nStep 2: Crawl4AI Processes Locally\n\u251c\u2500 Parses XML structure\n\u251c\u2500 Discovers 98 documentation URLs\n\u251c\u2500 Opens each in Playwright browser\n\u251c\u2500 Executes JavaScript (renders dynamic content)\n\u251c\u2500 Extracts article content only\n\u2514\u2500 Removes navigation, ads, footers\n\n\nStep 3: Clean Markdown Created\n\u251c\u2500 Article Title: \"Customer Management Guide\"\n\u251c\u2500 Content: Clean paragraphs, headings, lists\n\u2514\u2500 Metadata: URL, author, date, section\n\n\nStep 4: Chunk & Embed\n\u251c\u2500 Split into 1000-character chunks\n\u251c\u2500 200-character overlap for context\n\u251c\u2500 Send to Google Gemini Embeddings API\n\u2514\u2500 Store 3072-dimensional vectors in ChromaDB\n\n\nResult: 98 articles \u2192 2,847 searchable chunks\nTime: 10 minutes (one-time process)\nReady for instant retrieval!\n<\/code><\/pre>\n\n\n\nBenefits of Crawl4AI:<\/strong><\/p>\n\n\n\n
Feature<\/th> Benefit<\/th><\/tr><\/thead> Privacy-First<\/strong><\/td> HTML processing happens locally, never sent to LLM APIs<\/td><\/tr> JavaScript Support<\/strong><\/td> Renders dynamic content with real browser (Playwright)<\/td><\/tr> Clean Extraction<\/strong><\/td> Removes navigation, ads, popups automatically<\/td><\/tr> Sitemap Support<\/strong><\/td> Discovers all pages from sitemap.xml<\/td><\/tr> Batch Processing<\/strong><\/td> Handles 100+ URLs efficiently<\/td><\/tr> One-Time Ingestion<\/strong><\/td> Content stored permanently, no re-crawling needed<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n Data Sources in Agentic RAG:<\/strong><\/p>\n\n\n\n
MULTI-SOURCE KNOWLEDGE BASE\n------------------------------------------------------------\n\nInternal Documents (Manual Upload)\n\u251c\u2500 PDF: User manuals, API docs\n\u251c\u2500 DOCX: Internal workflows\n\u251c\u2500 Markdown: Technical guides\n\u2514\u2500 TXT\/CSV: FAQs, troubleshooting\n\n\nWeb Content (Crawl4AI)\n\u251c\u2500 Documentation portals\n\u251c\u2500 Knowledge base articles\n\u251c\u2500 Blog posts and tutorials\n\u2514\u2500 Release notes\n\n\nLive Web Search (Tavily)\n\u251c\u2500 Domain-restricted to official sites\n\u251c\u2500 Real-time information\n\u2514\u2500 Recent updates and announcements\n\n\nAll sources unified in a single search experience!<\/code><\/pre>\n\n\n\n
\n\n\n\nReal Example: Agentic RAG in Action<\/h2>\n\n\n\n
Let’s see how our 7-agent Agentic RAG system processes a real SalesWorx query.<\/p>\n\n\n\n
User Question:<\/h3>\n\n\n\n
“What steps are required to log in to SalesWorx?”<\/strong><\/p>\n\n\n\n
Agentic RAG Processing Flow<\/h3>\n\n\n
\n![\"\"](\"https:\/\/www.ucssolutions.com\/blog\/wp-content\/uploads\/2025\/12\/Screenshot_2025-12-05_at_08.50.22-1024x643.png\")
<\/figure><\/div>\n\n\nDetailed Breakdown (10-20 seconds total)<\/h3>\n\n\n\n
AGENTIC RAG PROCESSING TIMELINE\n------------------------------------------------------------\n\n\u23f1\ufe0f 0.0s - 0.5s: \ud83e\udde0 PLANNER AGENT\n------------------------------------------------------------\nInput: \"What steps are required to log in to SalesWorx?\"\n\nDomain Validation: \u2705 ON_TOPIC (SalesWorx product question)\nQuery Category: Product Documentation\n\nResearch Plan Created:\n\u251c\u2500 Step 1: Find login page URL and access method\n\u2502 \u2514\u2500 Tool: search_documents\n\u2502\n\u2514\u2500 Step 2: Find required credentials and fields\n \u2514\u2500 Tool: search_documents\n\n============================================================\n\n\u23f1\ufe0f 0.5s - 2.5s: \ud83d\udd04 RESEARCH CYCLE 1\n------------------------------------------------------------\n\n\u270d\ufe0f Query Rewriter:\n Original: \"Find login page URL and access method\"\n Rewritten: \"SalesWorx BackOffice login page URL web\n browser access authentication\"\n\n\ud83c\udfaf Retrieval Supervisor:\n Strategy Selected: HYBRID_SEARCH\n Reason: Needs both semantic (login process) and\n keywords (URL, BackOffice)\n\n\ud83d\udcda Retrieved Documents: 10 candidates\n \u2514\u2500 User_Guide.pdf (6 chunks)\n \u2514\u2500 Installation_Guide.pdf (2 chunks)\n \u2514\u2500 Quick_Start.pdf (2 chunks)\n\n\ud83d\udd04 CrossEncoder Reranking: Top 3\n \u251c\u2500 [1] User_Guide.pdf, Page 15 (Score: 0.94)\n \u251c\u2500 [2] Installation_Guide.pdf, Page 8 (Score: 0.89)\n \u2514\u2500 [3] Quick_Start.pdf, Page 3 (Score: 0.86)\n\n\ud83d\udd0d Context Distiller:\n Input: 3,200 tokens from 3 documents\n Output: 800 tokens (75% reduction)\n \"Login accessed via web browser at URL format\n http://<server>\/swx_cs\/html\/login.aspx.\n Supports Chrome, Firefox, Edge browsers.\"\n\n\ud83e\udd14 Reflection:\n \"Found login URL format and browser access method.\n Missing credential field details.\"\n\n============================================================\n\n\u23f1\ufe0f 2.5s - 3.0s: \u2696\ufe0f POLICY AGENT (Checkpoint 1)\n------------------------------------------------------------\nEvaluation:\n\u2705 Have: Login URL, browser requirements\n\u274c Missing: Required fields, captcha, password recovery\n\nDecision: CONTINUE_PLAN\nReason: \"Need credential field details and authentication\n requirements\"\n\n============================================================\n\n\u23f1\ufe0f 3.0s - 5.0s: \ud83d\udd04 RESEARCH CYCLE 2\n------------------------------------------------------------\n\n\u270d\ufe0f Query Rewriter:\n Original: \"Find required credentials and fields\"\n Rewritten: \"SalesWorx login page form fields username\n password captcha authentication required\"\n Context Added: \"Already found URL at\n http:\/\/<server>\/swx_cs\/html\/login.aspx\"\n\n\ud83c\udfaf Retrieval Supervisor:\n Strategy Selected: KEYWORD_SEARCH\n Reason: Looking for specific field names (username,\n password, captcha)\n\n\ud83d\udcda Retrieved Documents: 10 candidates\n \u2514\u2500 User_Guide.pdf (8 chunks about login form)\n \u2514\u2500 Security_Guide.pdf (2 chunks about authentication)\n\n\ud83d\udd04 CrossEncoder Reranking: Top 3\n \u251c\u2500 [1] User_Guide.pdf, Page 15 (Score: 0.97)\n \u251c\u2500 [2] Security_Guide.pdf, Page 22 (Score: 0.91)\n \u2514\u2500 [3] User_Guide.pdf, Page 16 (Score: 0.88)\n\n\ud83d\udd0d Context Distiller:\n Input: 2,800 tokens from 3 documents\n Output: 700 tokens (75% reduction)\n \"Login form contains three required fields: Username,\n Password, Enter Captcha. Forgot Your Password link\n available for password recovery.\"\n\n\ud83e\udd14 Reflection:\n \"Found all required credential fields (Username, Password,\n Captcha) and password recovery option. Login process\n fully documented.\"\n\n============================================================\n\n\u23f1\ufe0f 5.0s - 5.5s: \u2696\ufe0f POLICY AGENT (Checkpoint 2)\n------------------------------------------------------------\nEvaluation:\n\u2705 Have: Login URL, browser requirements\n\u2705 Have: Required fields (Username, Password, Captcha)\n\u2705 Have: Password recovery process\n\nDecision: FINISH\nReason: \"All aspects of login process comprehensively covered.\n Sufficient information to generate complete answer.\"\n\n============================================================\n\n\u23f1\ufe0f 5.5s - 6.0s: \ud83d\udcdd ANSWER GENERATOR AGENT\n------------------------------------------------------------\nContext Assembled:\n\u251c\u2500 Research Step 1: Login URL and browser access\n\u2514\u2500 Research Step 2: Required credentials and fields\n\nCitations Prepared:\n[3] User_Guide.pdf, Page 15: \"To access SalesWorx\n BackOffice, launch a web browser such as Chrome...\"\n\nGenerating answer with inline citations...<\/code><\/pre>\n\n\n\nFinal Answer Generated by Agentic RAG:<\/h3>\n\n\n\n
\n\n\n\nTo log in to SalesWorx BackOffice, you need to access the application through a web browser [3].<\/strong><\/p>\n\n\n\n
Here are the steps:<\/strong><\/p>\n\n\n\n
- \n
- Launch a web browser [3].<\/strong><\/li>\n<\/ol>\n\n\n\n
- \n
- Enter the SalesWorx URL<\/strong> in the browser’s address bar. This will be either
http:\/\/<server name>\/swx_cs\/html\/login.aspx<\/code> or the specific URL for your SalesWorx instance, where<server name><\/code> is the location of your SalesWorx BackOffice site [3].<\/li>\n<\/ol>\n\n\n\n- \n
- On the Login page, enter your credentials:<\/strong>
- Username:<\/strong> Type your username into the “Username” field [3].<\/li><\/ul>
- Password:<\/strong> Enter your password in the “Password” field. If you have forgotten your password, you can click the “Forgot Your Password?” link to reset it [3].<\/li><\/ul>\n
- \n
- Captcha:<\/strong> Type the characters displayed in the image into the “Enter Captcha” field [3].<\/li>\n<\/ul>\n<\/li>\n<\/ol>\n\n\n\n
Citations:<\/h3>\n\n\n\n
[3]<\/strong> SalesWorx User Guide, Page 15<\/em>
“To access SalesWorx BackOffice, launch a web browser such as Chrome, Firefox, or Microsoft Edge. Navigate tohttp:\/\/<server name>\/swx_cs\/html\/login.aspx<\/code> where<server name><\/code> is your BackOffice location. The login page contains three required fields: Username for your account credentials, Password for authentication, and Enter Captcha to verify you are human. If you have forgotten your password, click the ‘Forgot Your Password?’ link to initiate the password recovery process…”<\/p>\n\n\n\n
\n\n\n\nWhat Happened Behind the Scenes?<\/h3>\n\n\n\n
Traditional RAG would have:<\/strong><\/p>\n\n\n\n
- \n
- \u274c Done 1 search pass<\/li>\n<\/ul>\n\n\n\n
- \n
- \u274c Possibly missed credential details<\/li>\n<\/ul>\n\n\n\n
- \n
- \u274c No citations<\/li>\n<\/ul>\n\n\n\n
- \n
- \u274c No verification of completeness<\/li>\n<\/ul>\n\n\n\n
- \n
- \u23f1\ufe0f Time: 10 seconds (faster but less complete)<\/li>\n<\/ul>\n\n\n\n
Our Agentic RAG:<\/strong><\/p>\n\n\n\n
- \n
- \u2705 2 research cycles with different strategies<\/li>\n<\/ul>\n\n\n\n
- \n
- \u2705 Self-corrected when first cycle incomplete<\/li>\n<\/ul>\n\n\n\n
- \n
- \u2705 Full citations for every fact<\/li>\n<\/ul>\n\n\n\n
- \n
- \u2705 Policy agent verified completeness<\/li>\n<\/ul>\n\n\n\n
- \n
- \u23f1\ufe0f Time: 10-20 seconds (worth the extra 10 seconds)<\/li>\n<\/ul>\n\n\n\n
\n\n\n\nComparison: RAG vs Agentic RAG<\/h2>\n\n\n\n
Architecture Comparison<\/h3>\n\n\n
\n![\"\"](\"https:\/\/www.ucssolutions.com\/blog\/wp-content\/uploads\/2025\/12\/Screenshot_2025-12-05_at_09.23.07-1024x790.png\")
<\/figure><\/div>\n\n\n![\"\"](\"https:\/\/www.ucssolutions.com\/blog\/wp-content\/uploads\/2025\/12\/Screenshot_2025-12-05_at_09.23.23-1024x776.png\")
<\/figure><\/div>\n\n\nPerformance Comparison<\/h3>\n\n\n\n
Metric<\/th> Traditional RAG<\/th> Agentic RAG<\/th> Improvement<\/th><\/tr><\/thead> Answer Accuracy<\/strong><\/td> 75%<\/td> 95%<\/td> +20%<\/td><\/tr> Citation Coverage<\/strong><\/td> 0%<\/td> 100%<\/td> +100%<\/td><\/tr> Complex Query Handling<\/strong><\/td> Poor<\/td> Excellent<\/td> \u2b50\u2b50\u2b50<\/td><\/tr> Self-Correction<\/strong><\/td> None<\/td> 7 iterations<\/td> \u2705<\/td><\/tr> Response Time<\/strong><\/td> 3 seconds<\/td> 10-20 seconds<\/td> -10s (acceptable)<\/td><\/tr> Adaptability<\/strong><\/td> Fixed strategy<\/td> Dynamic<\/td> \u2705<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
\n\n\n\nDeployment & Scaling with Containerization<\/h2>\n\n\n\n
Modern Cloud-Native Architecture<\/h3>\n\n\n
\n
- \u23f1\ufe0f Time: 10-20 seconds (worth the extra 10 seconds)<\/li>\n<\/ul>\n\n\n\n
- \u2705 Policy agent verified completeness<\/li>\n<\/ul>\n\n\n\n
- \u2705 Full citations for every fact<\/li>\n<\/ul>\n\n\n\n
- \u2705 Self-corrected when first cycle incomplete<\/li>\n<\/ul>\n\n\n\n
- \u2705 2 research cycles with different strategies<\/li>\n<\/ul>\n\n\n\n
- \u23f1\ufe0f Time: 10 seconds (faster but less complete)<\/li>\n<\/ul>\n\n\n\n
- \u274c No verification of completeness<\/li>\n<\/ul>\n\n\n\n
- \u274c No citations<\/li>\n<\/ul>\n\n\n\n
- \u274c Possibly missed credential details<\/li>\n<\/ul>\n\n\n\n
- \u274c Done 1 search pass<\/li>\n<\/ul>\n\n\n\n
- Captcha:<\/strong> Type the characters displayed in the image into the “Enter Captcha” field [3].<\/li>\n<\/ul>\n<\/li>\n<\/ol>\n\n\n\n
- Password:<\/strong> Enter your password in the “Password” field. If you have forgotten your password, you can click the “Forgot Your Password?” link to reset it [3].<\/li><\/ul>\n
- Username:<\/strong> Type your username into the “Username” field [3].<\/li><\/ul>
- On the Login page, enter your credentials:<\/strong>
- Enter the SalesWorx URL<\/strong> in the browser’s address bar. This will be either
- Launch a web browser [3].<\/strong><\/li>\n<\/ol>\n\n\n\n
- Release notes<\/li>\n<\/ul>\n\n\n\n
- Video tutorials<\/li>\n<\/ul>\n\n\n\n
- Troubleshooting FAQs<\/li>\n<\/ul>\n\n\n\n
- API documentation<\/li>\n<\/ul>\n\n\n\n
- Role-specific guides (admin, sales, manager)<\/li>\n<\/ul>\n\n\n\n
- Multiple product versions<\/li>\n<\/ul>\n\n\n\n