Leveraging AI to Accelerate the Product Development Lifecycle

KEYWORDS: Artificial Intelligence, Product Development Lifecycle, Manufacturing, PLxAI, Retrieval Augmented Generation Framework (RAG), Design Validation Plan (DVP), Failure Modes and Effects Analysis (FMEA)

Overview

The product design and development lifecycle can become light years more efficient with the application of new technologies like AI. As it is with many use cases in the industrial world, AI has the potential to reduce the burden of manual and rote work placed on developers and can provide significant productivity improvements.

LTTS is harnessing artificial intelligence (AI) to revolutionize the product design and development lifecycle in the manufacturing industry. By automating manual, repetitive tasks and enabling smarter processes, AI significantly boosts productivity and efficiency for developers. The introduction of PLxAI, LTTS’s unified deployment platform for the product development lifecycle (PDLC), allows organizations to create tailored use cases quickly, leading to immediate productivity improvements and substantial cost savings across the entire lifecycle.

PLxAI is LTTS’s enterprise-grade, scalable, secure, and reusable GenAI platform for the Product Development Lifecycle (PDLC). It standardizes AI-enabled workflows across concept, design, development, production, validation, and aftermarket, providing modular assistants and reusable components that can be rapidly tailored to domain-specific needs. Beyond single use cases, PLxAI orchestrates data ingestion, knowledge retrieval, and agentic automation to deliver improved productivity and reduced time to market.

Challenges in the PDLC, such as time-consuming manual checks, inconsistent standards, error-prone design comparisons, and loss of tribal knowledge, are addressed through intelligent AI-powered solutions. LTTS deploys AI in various stages, from initial requirements and design to validation and prototyping, integrating advanced tools like generative design and digital twins. PLxAI leverages a Retrieval-Augmented Generation (RAG) framework, connecting large language models to authoritative knowledge bases, ensuring accurate, up-to-date responses and reducing errors in development, ultimately speeding up the design cycle and maximizing ROI.

Product Development Lifecycle Challenges

The manufacturing industry faces many challenges when it comes to the product development lifecycle. Manual checks across the design, development, and validation phases are time consuming and prone to errors. The inconsistent application of design and development standards and guidelines across the various components and subcomponents of a system can make the final product inconsistent with an organization’s standards. Similarly, comparing various versions of the same design is also  error prone and, in many cases, a manual process, which leads to incompatible subcomponents making it into the final assembly process.

Much of the design and engineering knowledge that exists in organizations today is also tribal and is often compartmentalized, so the engineering and design teams do not leverage this knowledge across the enterprise. Tribal knowledge can also leave the organization when experienced employees leave, and in most cases, there is no effective way to capture this tribal knowledge digitally to preserve it for the benefit of the enterprise. Collectively, these challenges slow the end-to-end design cycle, increase rework and late-stage changes, and force organizations to operate in a more reactive mode, undermining productivity, consistency and time to market.

How AI Can Address PDLC Challenges

All of these challenges can be addressed with intelligent application of AI and software in the product development lifecycle. LTTS, a leading design and software firm serving the manufacturing industry, is introducing a new solution called PLxAI that provides a unified deployment platform for PDLC and enables users to build their own domain specific use cases within a few weeks to a month. It promises quicker development, deployment, and faster ROI. According to LTTS, users have experienced productivity improvements of up to 15 to 25 percent almost immediately. When combining all use cases over the product development lifecycle, more significant savings of 30 to 50 percent can potentially be achieved.

How LTTS Deploys AI Across the Product Development Lifecycle

LTTS is leveraging AI across several domains of the development lifecycle, from product concept and requirement specification to acceptance, validation, and the final product. AI-based design assistance is deployed at the early stages during assembly requirements development, sub assembly design, and component design. AI-based generative design is used during the 3D modeling phase, while AI-based computer assisted engineering (CAE) is used for the CAE modeling phase. Finally, AI-based product digital twin is utilized in the component verification, integration and verification, and prototyping phase. What LTTS is essentially doing is creating an AI-powered framework that weaves together the OEM design process, commercial software, and component design that can be applied to scores of use cases.

PLxAI: Hybrid RAG + Knowledge Graph with Agentic Workflows That Leverages RLAIF

LTTS is deploying PLxAI as a RAG framework. Retrieval-Augmented Generation (RAG) is an AI framework that enhances Large Language Models (LLMs) by connecting them to external, authoritative knowledge bases (like company docs or the web) to fetch relevant information before generating an answer, making responses more accurate, up-to-date, and grounded in facts, thereby reducing hallucinations and allowing access to specific, current data without constant retraining. Think of it like an LLM acting as a smart student who quickly looks up specific facts in a library (the knowledge base) before answering your question, providing citations to show where they got the information.

LTTS PLxAI is a proprietary, enterprise-grade generative AI (GenAI) framework that utilizes a form of RAG architecture to accelerate and optimize the entire product development lifecycle. It is not merely a standard RAG framework, but a specialized, integrated solution combining GenAI with conventional AI technologies and agentic workflows to leverage specific organizational knowledge. PLxAI incorporates core RAG principles, which involve retrieving relevant information from an external, authoritative knowledge base to provide context to a large language model (LLM), thus generating more accurate and relevant responses than an LLM alone. 

PLxAI combines RAG with a Knowledge Graph (KG) for hybrid retrieval. The workflow covers parse/organize/classify → chunking → embeddings → context retrieval → LLM orchestration → response consolidation → summary/model generation, closed by Reinforcement Learning with AI Feedback (RLAIF). Enterprise integrations ensure responses are actionable within existing engineering tools and PLM/CAE environments. Customer data remains within their network boundary and is not used to train or tune public LLMs; PLxAI enforces strict residency and access controls to support compliance and auditability.

PLxAI includes Domain-tuned Reciprocal Rank Fusion (RRF) re-ranking and Hierarchical Navigable Small World (HNSW) indexing for balanced performance and precision. A smart query router targets segmented vector stores by use case/content type. Metadata-aware pre/post chunking is applied to improve recall/precision across heterogeneous corpora.

The framework is designed to capture and utilize existing “organizational and tribal knowledge,” which serves as the specific, proprietary knowledge base for the RAG system. This is crucial because generic public LLMs typically lack this domain-specific information. Instead of relying solely on the LLM’s initial training data, PLxAI retrieves relevant data from the knowledge base in response to a user’s prompt. This information is then used to augment the prompt with rich, context-aware insights, enhancing the quality and relevance of the output in real time.

With PLxAI, LTTS integrates GenAI with conventional AI technologies, creating a hybrid approach that provides comprehensive and reusable solutions across all PDLC stages. Moving beyond simple RAG, the framework also employs “proactive agents” that automate multi-step engineering tasks and self-correct, effectively creating more autonomous systems for complex problem solving. PLxAI also integrates with enterprise tools such as third-party PLM, CAD, and CAE solutions as well as document repositories to operationalize answers, maintain audit trails, and support compliance frameworks such as ISO processes.

Use Cases for PLxAI

As mentioned previously, PLxAI has many valid use cases in the PDLC. It can be used as a design assistant to automate design processes, reduce errors, and improve efficiency. PLxAI can also provide AI-based Design Validation Plans (DVPs) and can assist in tasks such as risk assessment, compliance, and failure mode analysis. It can also provide automated image/design drawing comparison for enhanced accuracy and reduced errors.

The knowledge management functions of PLxAI can be used for knowledge management (KM) applications by providing functions such as data extraction, interpretation, and report generation. Warranty analysis functions can automate things like claims processing, fraud detection, predictive maintenance, and warranty term optimization. Smart pricing functions can provide predictive pricing models.

The customer expects to improve overall productivity by around 50 percent for the engineering and R&D teams within six months of rollout, especially in the areas of document search, calculations, and smart prompting. The scalable and reusable components can also be reused for other design components such as seating, lighting, and more.

Design Validation Plan (DVP) Assistant

The DVP Assistant is designed to streamline the creation of Design Validation Plans (DVPs), tailored specifically for research and development teams. By automating the collection and analysis of information from a wide range of specifications and regulatory sources, it supports comprehensive compliance with all necessary requirements. The platform empowers R&D professionals to achieve significant time savings and operational efficiencies.

Data inputs for the system cover a variety of sources, including OEM specifications, PPAP documentation, DFMEAs, internal records, test results, and tolerance stack reports. These diverse data streams are integrated seamlessly to support thorough validation planning. The system rigorously examines acquired data to determine relevant requirements and specifications, facilitating the development of robust DVPs. It generates comprehensive DVP documents along with curated lists of Key Product/Control Characteristics (KPCs/KCCs) that align with all pertinent standards.

For data collection and processing, the assistant integrates connectors to both internal and external knowledge repositories, enabling access to essential specifications and regulations. It utilizes advanced large language models (LLMs) to process and summarize information, execute entity recognition, and perform topic modeling, thereby uncovering critical relationships and actionable insights.

Expertly tuned LLMs are deployed to interpret requirements, construct DVPs, and review historical documentation. During DVP planning, these LLMs evaluate new plans by referencing relevant prior projects and emphasizing key considerations that inform successful outcomes. With the DVP Assistant, LTTS claims that users can experience up to a 60 percent reduction in DVP creation time, achieve a 15–20 percent decrease in duplicate test cases, and significantly minimize manual work by automating both data gathering and synthesis.

Image/Design Drawing Comparison Assistant

Manual inspection often results in a high error rate and inconsistency, as it is susceptible to missing minor discrepancies—such as subtle misalignments, omitted features, and incorrect dimensions—among thousands of details. Fatigue may further worsen these inconsistencies during quality checks. The process is also extremely time consuming, particularly when examining numerous design iterations and variations for large assemblies or complex components, which requires extensive manual effort. This not only prolongs design cycles but also delays time –to market.

Additionally, manual comparison relies heavily on individual judgement, leading to subjective evaluations that make it difficult to standardize quality control measures and maintain consistent adherence to specifications across different teams or projects. The challenges are compounded by the exponential increase in design data due to technological advancements like ADAS sensor integration and sophisticated infotainment displays, rendering manual comparison of vast image and drawing sets impractical at scale. Traditional comparison methods tend to have a limited scope, typically identifying only superficial differences without capturing deeper variances.

Generative AI (GenAI) solutions address these issues and can significantly reduce error rates. They can detect minute deviations, inconsistencies, and omissions that may escape human observation, thereby improving the precision of design validation processes. Automation enabled by GenAI leads to significant time savings, allowing rapid comparison of complex drawings and images while freeing engineers from repetitive manual tasks. As a result, design iteration speeds up, rapid prototyping becomes feasible, and new vehicle features reach the market more quickly.

GenAI also introduces objective and standardized quality control by delivering consistent, data-driven comparisons based on predefined criteria, eliminating human subjectivity. This supports the implementation of standardized quality gates throughout the design process, resulting in improved product quality and compliance. GenAI also efficiently manages large datasets, facilitating comprehensive comparison of entire vehicle assemblies and historical design modifications at scale, and supports continuous integration of ongoing design changes with immediate feedback.

Beyond basic pixel-level analysis, GenAI provides intelligent and contextual insights by interpreting design intent, identifying meaningful deviations, and suggesting potential impacts of modifications. It can reference best practices and recommend alternative solutions, transforming itself into an active and strategic assistant in the design process.

Failure Modes and Effects Analysis Assistant

Key challenges of FMEA include comprehensive lists of potential failures within processes or designs. Each possible failure must be evaluated according to its severity, likelihood of occurrence, and ease of detection. Following identification, the Risk Priority Number (RPN) is calculated for each failure mode. These documents serve as a company’s risk reference guide and require thorough documentation and secure storage. The process remains largely manual and susceptible to errors due to the complexity involved in identifying, categorizing, and ranking risks.

LTTS’ xFMEA assistant can greatly streamline the FMEA process, automating data synthesis and failure mode suggestion, providing contextualized risk assessment and prioritization, and can significantly help in forming an intelligent strategy for mitigation. The system utilizes extensive historical data, including previous FMEA reports, warranty claims, field failure records, test results, engineering specifications, industry standards, and patent databases, ensuring that all relevant information is systematically documented and readily accessible for future use.

Beyond simply suggesting failure modes, the GenAI assistant applies its contextual understanding of design to evaluate the severity, frequency, and detectability of each identified failure. Leveraging analysis of historical incidents, simulation outcomes, and material characteristics, it delivers data-driven estimations of Risk Priority Numbers (RPNs), enabling design teams to efficiently concentrate on high-risk areas.

Upon identification of a high-risk failure mode, the GenAI assistant not only signals its presence but also recommends potential mitigation strategies. These are drawn from a comprehensive repository of proven solutions, industry best practices, and innovative methodologies sourced from various domains, accelerating the DFMEA problem-solving process.

Overall, this approach can enhance productivity by 25 percent through reduced research requirements, minimized human error, expedited iteration cycles, and improved allocation of expert focus. Additionally, it increases the accuracy and consistency of xFMEA reporting and formatting.

Design Standards Assistant

The Design Standards Assistant helps engineers and designers ensure compliance with established design standards during the product development process. It includes functions such as smart/similar part search; CAD automation, ECU validation & qualification, manufacturing feasibility, harness creation & intelligent routing, design/drawing checksheets, image/drawing comparison, DVP generation, and DFMEA/xFMEA acceleration.

Knowledge Management Assistants

PLxAI offers knowledge management assistants that can address low to high complexity tasks. Knowledge management assistant capabilities can include things like Product Synthesis Form (PSF) search/compare, and regulations & compliance validation (multi-destination), with automated reports and templates. Knowledge management assistants can reduce typical find-and-read cycles from 3 to 10 minutes to approximately 30 seconds to 1 minute by surfacing executive summaries, key bullets, tables, links, and next-step prompts.

Supplier RFQ Support

Supplier RFQ support functions include benchmarking (public & customer-specific), cost analytics, warranty analysis (claims, fraud, predictive maintenance), smart pricing, logistics and dispatch planning, and program cost prediction.

Innovation Assistant and Patent Assistant

The Innovation and Patent Assistant functions include automated analysis of public domain and customer-specific data to evaluate market and technical feasibility. The Patent Assistant is designed to integrate with the innovation process to help manage and protect intellectual property efficiently. This tool supports the legal and documentation aspects of patenting.

A Closer Look: PLxAI for Radiator Tank Design Assistance

To get a better idea of how PLxAI is used, let’s take a look at the use case of radiator tank design, which LTTS has deployed at a major OEM in the EU. The process of radiator tank design may seem simple on the surface, but it can actually be quite challenging. The current process or radiator tank design is quite complex, and in this particular case involves reviewing a repository of over 50 documents, including PDF files, XLS files, drawings, and other documents. Most of the documents contain hand drawn images, tables, diagrams, and process flows.

Engineers would typically spend many days going through each document, spending over a month to get a good understanding and summary of what these documents mean. Engineers often need to prepare the summary, and they need to remember several key processes and design elements.

Using PLxAI as a design assistant is helping the end user engineering teams improve their understanding of the design process, especially from design documents, diagrams, and calculation sheets (DCS). PLxAI increased the efficiency of the engineering teams in design and has enabled products to reach the market more quickly than the old design cycles.

The design team was also able to use PLxAI to display images, drawings, flow charts, tables, formula calculations, and relevant web search results as part of the prompt responses. They also used the source locator and document name finding features along with the page number of the input dataset as part of the response. If users are not able to prompt correctly, the assistant can recommend any potential missing information.

Benefits and Business Impact

In ARC’s view, PLxAI is a great example of how AI can be practically applied to the PDLC process to create real business value, reduce development costs, and reduce time to market by 20 to 30 percent. Innovation is further accelerated through the use of “Innovation Assistants,” which aid in benchmarking and technology scouting. Cost optimization is achieved as automation and reusable AI contribute to lower engineering and support expenses. By leveraging organizational knowledge, enhanced quality is realized through improved decision making. Additionally, efficiency is boosted by automating DVP and FMEA template creation, streamlining the overall documentation process.

PLxAI can provide immediate product development cost savings. According to LTTS, customers have reported immediate productivity improvements of 15 to 25 percent, with aggregated PDLC savings of 30 to 50 percent when multiple assistants are deployed end-to-end. In a Tier-1 thermal business unit trial, PLxAI saved approximately 16,000+ person hours of development over 12 months across eight teams and 100+ documents, with module-level readiness and measurable monthly gains. 

ARC Advisory Group clients can view the complete report at the ARC Client Portal.

Contact Us if you would like to speak with the author.

Obtain more ARC In-depth Research Market Analysis.