Par Robert G. Cooper

L’intelligence artificielle (IA) est sur le point de révolutionner tous les aspects des affaires, notamment le processus de développement de nouveaux produits (NPD). Robert G. Cooper, créateur du modèle Stage-Gate et auteur de nombreux ouvrages et articles, nous donne ici un aperçu des applications de l’IA à toutes les phases de ce processus, illustrées par de nombreux exemples en entreprise. Selon Cooper, l’IA a déjà entraîné des changements spectaculaires dans des entreprises pionnières, améliorant la vitesse, l’efficacité et la qualité de leur processus de développement de produits.

AI is posed to transform new-product development, and is already cutting development times by 50% for leading early-adopter firms.

The 9 milestones of new-product development displayed at the 2023 PDMA Annual Conference. Artificial Intelligence will become the 10^th milestone by the end of this decade and will alter the landscape of product innovation (AI poster #10 added to photo).

Overview: Artificial Intelligence (AI) is poised to revolutionize all aspects of business, particularly new- product development (NPD). Currently, our approach to NPD has remained largely unchanged for decades, yielding stubbornly poor results: only 30% of new-product projects succeed commercially. However, the Artificial Intelligence revolution is set to alter this landscape significantly!

Leading early-adopter firms demonstrate that AI not only finds many applications in NPD, but also offers substantial payoffs, such as a 50% reduction in development times. This article provides an outline of the diverse and powerful applications of AI in NPD, offering numerous examples from leading companies. Our exploratory journey begins at the idea stage and traverses the entire new-product process to the post-launch period. Some examples are nothing short of astounding: Unilever’s $120-million lab, staffed entirely with robots for creating and testing new personal care products, end to end; Nestle’s “concept generator” that scans multiple sources for data and insights, and then generates new product concepts based on the scanning results; or Moderna’s AI-based drug discovery tool that yielded the mRNA Covid vaccine. Additionally, digital twins, employed by leaders like GE, Siemens, and Tesla, play a pivotal role during development (field testing) and even during post-launch, enhancing customer satisfaction.

While AI might still resemble science fiction to many, that future is no longer fiction – it’s here now. AI has arrived in full force! The numerous applications of AI in NPD have already ushered in dramatic changes, enhancing the speed, efficiency, and quality of NPD for these early-adopter firms. However, the journey is not about adopting AI on a piecemeal basis; rather, this must be a holistic transformation guided by a bold, enterprise-wide strategy that is created and championed by your business’s leadership team.

Citation: Cooper, R.G. 2023. “The Artificial Intelligence Revolution in New-Product Development,” Penn State Institute for the Study of Business Markets Members Conference, Chicago IL: Sept 2023.

1. The Coming Revolution

Artificial Intelligence (AI) is poised to transform all aspects of business, and with it, new-product development. Business historians, a hundred years from today, will look back and see this decade as the turning point for AI, much like the early days of the Industrial Revolution or the Electrification Era a century later. Both changed the world, mostly for the better.

Today we still approach new-product development much as we did when Edison opened his first R&D laboratory in 1876. Unfortunately, results have remained stubbornly poor, with only 30% of new-product projects succeeding commercially.¹ This is about to change, however, as the Artificial Intelligence Revolution takes hold in new-product development (NPD)!² Leading firms have adopted AI for a various reasons such as improving efficiency, business agility, or productivity; but a full 35% of early adopters globally report that the main benefit is increased innovation, making it the number one improvement realized.³

Consider Nestlé, which has increased its pace of product development by 60% in 6 years using AI.⁴ CTO Stefan Palzer stresses that AI and Machine Learning (ML) are essential product development tools: “Concept research, formulation development, plant breeding, clinical data mining, raw material quality assurance, improved process control, and early problem detection are just a few of the ways that AI is already being applied throughout the enterprise.”

1.2 But What Counts as AI? Stories abound in the media about AI since the wide availability of ChatGPT. But just what is AI when it comes to NPD? Does a simple sales forecasting algorithm based on statistical analysis count as AI? Or does a portfolio optimization routine using linear programming count? Probably not, according to ChatGPT.

So, the question was put to Artificial Intelligence itself: Who or what are you?:

“AI refers to the simulation of human intelligence in machines that are programmed to think, reason, and perform tasks that typically require human intelligence. AI enables computers or machines to mimic cognitive functions such as learning, problem-solving and reasoning, perception, and decision-making. The goal of AI is to create systems that can perform tasks independently, adapt to new situations, and improve their performance over time without explicit programming” (source: ChatGPT).

AI can be implemented through various techniques and technologies, including machine learning (ML), natural language processing (NLP), computer vision, robotics, and expert systems. Machine learning, in particular, plays a significant role in modern AI development, enabling systems to learn from data and improve their performance through experience.

AI has been defined more simply for business as a “prediction technology that reduces the cost of predictions”.⁵ Where we make predications today, such as market forecasting, AI will makes them faster, better, and much cheaper. And for problems that today do not involve prediction, AI will take the traditional problem – such as creating the drawings for a new concept car – and turn it into a prediction model: AI will predict how a creative artist would draw the car, given specific instructions. Prediction is at the heart of making decisions under uncertainty, which is a major part of a new-product project, hence the transformative role for AI in NPD.

1.3 AI’s Role in NPD Till Now The earliest use of AI in NPD can be traced back to the 1960s when computer-aided design (CAD) systems were introduced. These systems enabled engineers to design products on computers, reducing the time and cost of NPD. In the 1970s and 1980s, expert systems were developed, which could mimic the decision-making capabilities of human experts. These systems could be used to evaluate the feasibility of new product ideas.

In the 1990s and 2000s, AI began to play an increasingly important role in NPD. Data mining and predictive modeling tools were developed, which could analyze customer data from various sources such as social media, online reviews, and surveys to gain insights into what customers were seeking, and make highly accurate predictions.

2. NPD as an Information Process

A simple but profound description of the NPD process from idea to launch from a GE executive is:

“The new-product process is simply a set of tasks designed to gather information to reduce uncertainty and thereby manage risk”.⁶

This information model is portrayed in Figure 1:

• Project teams perform a set of tasks – a market study, a lab test, a field trial – to seek
• Teams analyze the information, integrate it, and make predictions, for example, a business case or engineering drawings for the new product.
• Then a management team makes a decision to continue to invest or to kill the

The information that teams gather in Figure 1 yields uncertainty reduction, predictions, and validation of assumptions about key facets of the project, such as the product’s design and the project’s economics. This simple model in Figure 1, repeated a number of times from idea though to launch, is the basis the Stage-Gate process, which was designed to reduce uncertainty and thus mitigate risk.⁷

Figure 1: The NPD process viewed as an “information process” – a series of task designed to gather information, which reduces uncertainties, enables predictions, and manages risk. AI is ideally positioned to revolutionize this information and risk-reducing process.

AI is all about information, uncertainty and risk reduction, and prediction – gathering, analyzing, integrating, and predicting. So AI is perfectly positioned to revolutionize this NPD “information process” in Figure 1.

3. Positioning AI Within NPD

One way to think about AI’s many NPD applications to visualize the role of AI is by where it occurs in the NPD process.⁸ Is this application largely for the front end, for example, idea generation? Or for the back end, for example, planning the launch?

A second way to think of AI’s role is as an Originator or a Facilitator. This construct is based on a recently developed conceptual model of AI in NPD by Brem et al (2023).⁹

• The Originator role of AI is the creative side of product development – leveraging AI as creative model and method of inventing.
• By contrast, the Facilitator role is more about improving existing processes and methods, making them more efficient and effective. It capitalizes on AI’s ability to integrate and manipulate data in innovative

The resulting AI-NPD model is shown in Figure 2 which positions the many AI-NPD applications. The Originator- Facilitator dimension is the vertical axis; the location in the NPD process is the horizontal axis. The rest of this article highlights these applications with examples, with a focus on physical or manufactured new products.

Figure 2: The AI-NPD Model position AI in NPD – where used in the NPD process and the role of AI as a Fcilitator or Originator.

4. Front End Applications

4.1 AI and Idea Generation Companies utilize AI-powered idea generation tools to generate new-product ideas. These tools employ ML algorithms to analyze vast amounts of data from various sources, such as customer feedback, social media, and market trends. This analysis identifies market gaps, emerging customer needs, and areas of untapped potential. For example, LEGO employs market insights tools based on ML learning to analyze emerging trends and predict product growth areas based on publicly available Internet data, such as likes or clicks on trending videos or memes.¹⁰

Social media AI scanners monitor the social web, including social media channels, forums, blogs, comment data, reviews, and news media. And they search for insights relevant to a product and industry sector, including share of voice, customer sentiment, demographic and psychographic characteristics of the audience, and competitive intelligence. These AI tools reveal what customers like or dislike, and can identify of customers’ points-of-pain, needs and wants, and even new product ideas. Some of the top-rated AI scanners include YouScan, Brand24, and Birdeye, according to Capterra.¹¹

The idea for a new range of skincare products by Unilever came from AI: AI-powered algorithms analyzed social media data, identified emerging skincare trends, along with customer needs and preferences, and pinpointed the opportunity. Nestlé has even built an AI “concept engine”, which transforms these scanning insights into concept proposals, which are then evaluated internally and then by potential users.¹²

Even more ideation! NPL is used to analyze open-ended text responses: It extracts insights and trends from unstructured text data, such as customer feedback, survey responses, product reviews, and social media comments. For example, the Upsiide platform uses AI to generate new product ideas by analyzing open-ended responses from surveys and interviews. AI applications include identifying key challenges or needs that users face while using a certain product and even providing methodological support (e.g., AI generates a first draft of a customer interview guide).¹³ Another example is Applied Marketing Science’s app, which scans thousands of user comments and complaints on blogs to identify needs and sought-after benefits in a product area, resulting in successful new products ideas or concepts.

Generative AI, such as ChatGPT or Bard, generates new data, content, or artifacts similar to what humans might create. Generative AI can directly generate ideas for new products for a specified market or product category when given the right prompts. A recent ideation study using ChatGPT revealed positive results: The impact of ChatGPT in generating innovative concepts was measured compared to the use of “classic” design methods (brainstorming, TRIZ, etc.). From a “novelty” perspective, ChatGPT performed quite well, contrary to expectations (since its knowledge base contains only information about things that have already happened).¹⁴ However, ChatGPT proved less helpful in suggesting “useful concepts”.

4.2 AI’s Role in the Concept Stage The concept stage in NPD is part of the fuzzy front-end of the process, and not meant to be definitive (Figure 3); it involves defining the product concept, creating preliminary designs, and undertaking a “first-pass” business analysis. With so many varied tasks to do, and much of the data readily available, AI must play a major role in this early concept stage:

1)  Market Assessment: AI enables the analysis of vast amounts of market data, including customer or consumer behavior, trends, and historical data, to provide insights into market demand. Improved project scoping is made possible by automating the time-consuming collection and analysis of user stories as well as customer feedback from concept tests. AI automatically gathers and analyzes data on competitors, their products, pricing strategies, customer reviews, and market positioning to help identify opportunities and threats. ChatGPT can also perform a PESTEL analysis (political, economic, social, technological, environmental, and legal).¹⁵

Figure 3: AI plays a key role in the Concept stage, undertaking the tasks that are typical for this stage, but more efficiently and effectively.

2)  Concept Design:¹⁶ AI algorithms analyze large amounts of data and generate insights to assist with product design This leads to more tailored or personalized products that meet the unique needs of individual customers or segments. Generative AI algorithms create realistic 3D renderings of concepts. AI algorithms also analyze the design data and, when prompted, show design modifications; this makes performing “what if” scenarios simpler – “make it stronger, make it smaller” – and enhances the process of comparing different product concepts, for example in terms of cost or weight. Visualizing the concept makes it easier for stakeholders to see what is required to build the product, and also can be used to seek customer reactions in a concept test.

Mattel uses a generative AI image-creation tool, OpenAI’s DALL-E system, to create realistic images and art – concept designs for new Hot Wheels cars – based on natural language inputs.¹⁷ Once a design is created, for example, the designer can type “make it a convertible” and the AI tool updates the image of the car as a convertible… all done with words via typed text.

Mattel is working with a generative image-creation tool, OpenAI’s DALL-E system, for creating realistic images and art based on natural language (spoken) inputs. SourceL J. Roch, Microsoft News, Oct. 12, 2022. Endnote 17.

3)  Technical Assessment: AI can assist in evaluating the technical feasibility of a new product concept by analyzing existing technologies, potential challenges, and required resources.

4)  Financial and Risk Assessments: AI performs quick financial modeling and forecasting, helping to estimate costs, revenue projections, and potential profitability. AI can also identify potential risks associated with the proposed product and project. For example, with the right prompts, ChatGPT can identify the major market and technical risks.

5)  Regulatory and Legal: AI assists in researching relevant regulations and standards that may apply. And AI can help search for existing patents and intellectual property related to the product concept, mitigating infringement issues.

4.3 Building the Business Case: Some of the tasks here are similar to those in the Concept stage, but done in more depth and require more and more accurate data and much better forecasts for the business case. AI becomes a great help (Figure 4):

Figure 4: AI-powered tools can gather and analyse market, sales and other financial data, and make projections about the potential revenue and profitability of a new product, this helping to build the business case.

1)  Market Analysis: AI is used to understand the size of the market, the competition, and target customer segments. AI tools analyze market data from sources such as online reviews, social media, and sales data to gain insights into the market.¹⁸ AI tools also make predictions: market, sales, pricing, and costs. When the correct data is available, ML can detect patterns that can’t be discerned by humans, vastly exceeding human accuracy in making predictions. Many software tools exist which seek market data and undertake market analysis – examples include Monday, Funnel, and SAS ¹⁹

2)  Technology Analysis: AI software provides real-time insights into technology trends, investment activities, and technology disruptions (e.g., CB Insights). ChatGPT also answers questions on the fly and rapidly pulls together facts on a specific technical subject.

3)  Competitive Intelligence: Tools like Crayon use AI-powered algorithms to monitor competitors’ activities: product launches, pricing changes, and marketing campaigns. AI also provides real-time insights into competitors’ strategies, market positioning, and customer targeting, thus helping project teams make informed decisions on product differentiation and market opportunities.

4)  Financial Analysis: AI-powered tools can analyze market, sales, and other financial data to help the project team build a business case for their project. Additionally, by analyzing historical data from similar products, AI can predict how well a particular product will sell and how much money will be ²⁰ AI can thus be used to analyze financial data and make projections about the potential revenue and profitability of the new product. And AI is used to simulate different scenarios and predict the impact of various factors, e.g., pricing and competition.

5)  Risk Assessment and Management: This is one of the most developed areas of AI in Applications include the use of big data and ML to help project teams anticipate risks that might otherwise go unnoticed. These tools can already propose mitigating actions, and soon, they will be able to adjust project plans automatically to avoid certain types of risks.²¹

4.4 Decision to Go to Development The “Go to Development” decision is one of the most important resource commitment decisions in the NPD process (Gate 3 in Figure 4). While AI offers tools for better project selection and prioritization, the prospect of a robot making Go/Kill decisions in a NPD project is somewhat of a And AI expert and VP of an AI software firm in Shanghai told me: “Just like Tesla’s Autopilot driving system, [AI is not sufficiently] reliable to replace a human driver. It’s a good co-pilot or assistant to help people drive”. He went on: “You can now leverage the AI to help people and organizations do new product selection decisions…. [but AI] still needs people to verify or evaluate the result…. My guess is 5-10 years [before AI alone makes the Go/Kill decision]. ”

Figure 5: Neural network analysis of date from past new-product projects results in a preditive model of new-product success. And ML enables the AI model to “learn” and improve over time, and so potentially could be a gatekeeper at Go/Kill gate meetings.

AI can identify development-ready projects that have the right fundamentals in place and also predict which projects have higher chances of success or will deliver the highest benefits. Besides being able assess large amount of data and to recognize patterns better than a human can, an AI decision-model removes human biases from the decision. Using neural network analysis, AI will predict new product success or failure, and much more accurately than a human decision-maker (Figure 5). For example, the NewProd model makes new product success/failure predictions, based on characteristic of the project (market attractiveness, unique product benefits, competencies of the firm, etc.). The model was created using both a neural network analysis as well as traditional statistical analysis of several hundred past new-product projects with known commercial outcomes

While senior management is not ready to turn NPD decision-making over to AI, the same is not true in the world of finance: “A Hong Kong venture capitalist fund credits a single member of its management team with pulling it back from the brink of bankruptcy. But the executive is not a seasoned investment professional, nor even a human being. It is an algorithm known as Vital.” ²² Time will tell in NPD.

Finally, AI can help undertake portfolio management, namely looking at a number of projects and selecting or prioritizing them in order to create a better balance in the project portfolio or to maximize the value of the portfolio (subject to specific constraints and a desired objective). For example, Sopheon’s Accolade uses AI to optimize a set of projects in a firm’s development portfolio.

5. AI’s Role in Middle and Back End of Product Innovation 

5.1 Development While many AI applications are cited for the Development Stage (see Figure 1) – the various software vendors promise even more – the main AI applications for physical new products boil down to the four depicted in Figure 6:

1. 1. The use of AI to design the
   2. The development of digital prototypes and models to permit rapid testing of product iterations .
   3. Design
   4. The “discovery” of new products (e.g., the identification of a molecule that meets certain requirements).

1)  Product Design: AI streamlines product design by automating many design tasks, such as creating 3D models or generating technical ²³ “The time it takes to generate a physical incarnation – or even a 3D or visual incarnation of a product — requires some real physics behind it,” explains Deepan Mishra (Amazon Web Services Senior Advisor).²⁴ But AI can help create mockups and virtual protypes in just a few hours. For Mishra, “The fact that you can create content from scratch with such rapid speed, and hit a higher level of accuracy, is one of the most exciting components of all this.”

AI does much more than just engineering design. Autodesk uses prompts to direct AI to help its clients design products with the right features and dimensions for their target market.²⁵ AI also aids in designing products that are more user-friendly and aesthetically pleasing. For example, working with General Motors, MIT’s John Hauser and colleagues have developed two models:²⁶

• A generative model that creates new car designs based on prompts from designers about viewpoints, colors, body type, and image.
• A predictive model that forecasts how consumers will rate car designs with respect to aesthetic appeal or

With predictive modeling, carmakers can quickly eliminate those designs that are predicted to have low appeal to car-buyers; consequently, these options aren’t advanced to consumer “theme clinics” for judging (which take time and can cost $100K each). The cost and time savings in development are significant.

Figure 6: In the development Stage, AI creates and optimizes product designs, undertakes rapid test iterations, and even discovers products.

2)  Rapid, Iterative Product Testing: By developing digital prototypes or models, AI permits the rapid testing of multiple product iterations quicky during development. For example, Siemens uses AI to create virtual prototypes that can be tested and refined before they are built.²⁷ AI also optimizes their designs for performance, reducing the number of iterations required. Ultra-high-fidelity digital prototypes can be created for digital validation, whose validation results are then used to improve the design of the new product and the corresponding manufacturing processes (methods and examples in endnote 28).

The results of digital product validation are exciting: GE has cut design times in half by using AI for rapid design testing in turbine development.²⁹ Traditionally, it might take 2 days for engineers to run a computational analysis of the fluid dynamics of a single design for a turbine blade. Now, ML trains a surrogate model, enabling a million design variations of the blade to be evaluated in just 15 minutes.

To streamline its Automated Manual Transmission (AMT) development, Renault turned to AI-simulation to create models of the transmission.³⁰ The simulation technology relies on neural networks, AI “learning systems” modeled on the human brain. Design engineers can predict the behavior and performance of the AMT and make any necessary refinements early in the development cycle, avoiding late-stage problems and delays, thus cutting the AMT development time almost in half.

In a very different industry, biotech company Moderna uses AI to develop and test thousands of different mRNA-based medicines and vaccines.³¹ Drug companies typically spend a billion dollars developing a drug in the hopes of making billions in return, but see a success rate of just 15%. AI improves those chances to 50% and reduces the time-to-market as well. Moderna built a web-based drug-design application to streamline the process of inputting the information encoded in a synthetic mRNA molecule. AI algorithms help the molecular design decision-making by data-driven predictions on the best code-sequences to use. “The AI bet paid off for the Moderna, which was able to develop a leading COVID-19 vaccine in record time.”

3)  Design Optimization: Design optimization problems are usually solved iteratively by designers by improving an initial design through rounds of engineering analysis, interpretation, and refinement.³² This takes much time and money! Thus, a common use of AI-design is structural optimization: creating parts that provide sufficient strength, stiffness, and fatigue resistance with the minimum of Across industries ranging from automotive to aerospace, generative algorithms have reduced part cost by 6 to 20%, weight by 10 to 50%, and development time by 30 to 50%. A typical example from McKinsey is a tool manufacturer that reduced the weight of the forged steel component in Figure 7 by 38% and its cost by 15%.

Product design optimization is one of the principal ways that GE employs its digital twins.³³

Digital twins are digital models that digitally mimic products or components, manufacturing processes, and even entire systems. AI and ML are two key capabilities integrated into GE’s digital twin technology, that enables data to be collected, analyzed, and acted upon to optimize the product. GE has built up a vast catalog of digital twins that have created $1 billion in value.

4)  Discovery of Products: Major pharmaceutical firms are now using AI for drug discovery. A successful drug discovery project identifies biologically active lead molecules for the disease target that are chemically distinct from known drugs. AI accelerates the process by screening many compounds from a huge number of sources and libraries. For example, Pfizer, Lilly and Sanofil are all collaborating with the AI firm, Atomwise, to use AI deep learning to analyze millions of potential drug compounds to identify promising candidates for further testing. Atomwise developed their ML discovery engine that combines neural networks with massive chemical libraries to discover new small molecules, which ML predicts will bind to the target of interest.³⁴

In a similar vein, Unilever, working with Arzea, developed new stain-fighting enzymes for cleaning and laundry products with 50% fewer ingredients, yet increased stability, performance and sustainability.³⁵ The AI tool, Intelligent Protein Design Technology, combines physics-based design and AI to create new proteins and enzymes. This dramatic enzyme development was achieved In just 18 months – 5 times faster than previously possible. “The progress made in just 18 months with Arzeda’s technology shows how the convergence of AI and biology is a game-changer for an industry like home care,” says Peter ter Kulve, Unilever Home Care President. More recently, scientists at University of California, San Franciso, have created an AI system capable of generating artificial enzymes from scratch.³⁶ In laboratory tests, “the experiments demonstrate that NLP, though developed to read and write language text, can learn at least some of the underlying principles of biology”.

5.2 Project Management AI tools facilitate scheduling processes and can draft detailed plans and resource demands. Automated reporting is not only produced with less labor but can replace today’s reports, which are often weeks old, with real-time data. These tools also drill deeper than is currently possible, displaying project status, benefits achieved, potential slippage, and team sentiment in a clear, objective way. ³⁷ AI-based product management tools Albert Invent, Project Insight, Asana, and Trello help with efficient data management, project scheduling, and resource allocation.

AI enhances project management and planning in a wide variety of ways, according to the Project Management Institute (PMI), automating time-consuming project management tasks, such as:³⁸

• Project Planning, with dynamic and flexible timelines and Gantt charts, and resource allocation
• Data collection and analysis of task completion times, resource utilization, and project
• Predictive analytics, to analyze large amounts of project data, identify hard-to-see patterns of time and resource usage, and then to forecast project outcomes (on time, on budget performance).
• NLP to analyze project documentation, such as project charters, plans, and
• ML to learn from project data and improve the accuracy of predictions over time; and ML to optimize resource allocation, identifying the most efficient use of resources for each task.
• Real-time monitoring and control to monitor project progress and alert the project leader to potential
• Decision support to provide data-driven insights to make more informed decisions, for example, to recommend whether it’s better to allocate more resources to a task, or to adjust the project schedule.

5.3 Validation and Testing While much iterative testing occurs in the Development stage with models, virtual products, digital twins, and digital prototypes, many firms still conduct formal and rigorous field trials, beta tests, or in-home consumer tests of the pre-production physical prototype or MVP This Testing and Validation stage is critical, as tests here ensure that the “final product” meets requirements and specifications under “real use” customer conditions. AI really facilitates this vital stage (Figure 8):

1) Building Physical Prototypes: AI is used to create prototypes of products more quickly and easily. Examples were provided above on the use of AI to create optimized or pre-tested engineering CAD designs, ready to move to the next step. This automates and greatly accelerates the process of moving from a blank screen to a rapid or 3D-printed near-final prototype. or real components and products, ready for physical product testing.

In the field of consumer goods, Unilever’s Materials Innovation Factory (MIF) has been using robots and AI to develop and test products for the last 5 years. According to Unilever, “MIF has the highest concentration of robots doing material chemistry in the world”, and each machine is designed to crunch “colossal amounts of data and maintain consistency across samples and testing”.³⁹ In 2022, products developed by MIF yielded one- third of Unilever’s tech-derived product sales. A number of recently successful new products have been created with MIF’s AI and robotics: Dove Intensive Repair Shampoo and Conditioner, Living Proof Perfect Hair Day Dry Shampoo, and Hourglass Cosmetics.

Testing: How digital products, digital twins, and models are created and tested was illustrated above. Such digital twins or models are useful for field testing physical products as well. Algorithms analyze data from sensors in the test product to identify potential defects or product quality issues. For example, road-testing of new-product vehicles is done with digital twins.⁴⁰ And in the development of self-driving vehicles, the test cars contain numerous sensors that collect on-the-road data, which they feed the virtual or digital twin.

Figure 8: AI accelerates the developement of physical protypes for testing, and automates the securing of useful customwe-test-results.

Many new products today are complex: They are smart, connected (IoT) products, and contain multiple systems and software that must be tested. Advanced system testing solutions for software projects will soon allow early detection of defects and provide self-correcting processes.⁴¹

Obtaining and interpreting feedback from users from in-home product tests or field trials is also enhanced by AI. When testing products with customers, project teams typically rely on focus groups, interviews, and conversations with customers to collect feedback; but much data from these sources, because the data are unstructured (e.g., written or typed notes or recorded conversations) are often not captured; so project teams miss many insights into product performance and acceptability.⁴² Fortunately, “generative AI can help convert customer feedback into data for your business,” Mishra explains.⁴³ NLP takes unstructured data – recorded conversations and product usage comments – and translates these into charts and trend lines, and analyzes them in the same way that structured data is. One can then determine product acceptability and what customers like or dislike about particular product features and benefits.

5.4 Commercialization and Launch

1)  Market Launch: AI helps companies market their new products more effectively (Figure 9). AI guides the project team in crafting the marketing launch plan for the new product: the strategies and tactics to reach the target market effectively. Project teams can now generate highly effective marketing plans with ease by employing AI-driven marketing plan generators, such as Taskaid, to streamline the planning process.

AI also assists with the execution of some of the elements of the marketing mix, and numerous AI tools are available to create, plan, aid decisions, and execute; some examples include: ⁴⁴

Figure 9: AI proves useful in the Launch Stage – both in marketing and manufacturing, with many applications in each.

Marketing communications:

• A significant 78% of B2B and 65% of B2C firms now use generative AI, such as ChatGPT, to create advertising text, images, videos or other content.⁴⁵
• Google uses AI to target its ads to people who are likely to be interested in its
• One of Amazon’s most effective sales strategies is using AI-powered product recommendations to engage customers and increase ⁴⁶ Offers like “recommended for you,” or “frequently bought together” are instances of this highly effective selling strategy using AI.

Salesforce and distribution:

• AI automates lead scoring and salesforce routing, predicts customer behavior and preferences, and suggests the next best actions for the sales team.
• Procter & Gamble uses ML algorithms to define the right product assortment for physical and virtual stores, and to analyze in-store information on product availability.⁴⁷

Pricing:

• AI helps companies set the right price for their products: Netflix uses AI to personalize pricing for its subscribers based on their viewing habits.
• Uber uses AI to optimize its pricing strategy by employing ML algorithms and Predictive Analytics to analyze customer demand and adjust its pricing dynamically based on real-time demand.

2) Manufacturing: The limitless applications and inroads that AI has made into manufacturing and supply- chain management go far beyond NPD, and is beyond the scope of this article (Forbes⁴⁸ provides an excellent review).

5.5 Post Launch After its launch, AI continues to play a pivotal role in the success of NP. AI and NLP algorithms, similar to those previously highlighted, meticulously analyze customer feedback and market data to discern use patterns, comments, and complaints. AI thus serves as a vital tool for companies to pinpoint potential issues or areas warranting improvement in the just-launched product, thus facilitating design refinements aimed at total customer These analytical insights also prove invaluable to developers for charting the course of future product development endeavors.

Examples: A digital twin exists for each new Tesla vehicle on the road; its embedded sensors diligently collect performance data and relay the data to its twin. Invaluable feedback is derived from the real-world operation of the vehicle.⁴⁹ Similarly, in the realm of aviation, GE has implemented digital twins for its GE90 engines on Boeing 777 aircraft that foresee engine degradation. Siemens, a pioneer in the domain of digital twins, has introduced ATOM, a virtual model for its gas turbines and compressors. Blockchain technology is anticipated to serve as a robust repository for vast volumes of digital twin data.

6. The Future is Here

The future that we read about in science fiction is here now. AI has arrived in full force! AI’s many applications in NPD have already created dramatic changes in the speed, efficiency and quality of NPD for a handful of early-adopter firms. The AI revolution has been compared to the Industrial Revolution (1760) which occurred over an 80-year period. But new technologies are adopted much faster today than back then, so, expect the rate of change and impact of AI on business to occur with lightning speed. The time is now to embrace this transformative technology for NPD and indeed across the entire organization.

Experts agree, however, that the journey is not about adopting AI on a piecemeal basis – one can become mesmerized by the clever individual applications of AI in NPD or in marketing or production. Rather, this must be a holistic transformation: According to the strategy consulting firm, Deloitte: “One of the most frequently cited leading practices for AI transformation is the need for a bold, enterprise-wide strategy that is set and championed by an organization’s highest leadership.” That time is now!

References:

¹ Knudsen, M. P., von Pedowitz, M., Griffin, A., and Barczak, G. 2023. “Best practices in new product development and Innovation: Results from PDMA’s 2021 global survey.” Journal of Product Innovation Management 40: 257–275. doi: 10.1111/ jpim.12663; and: Barczak, G., Griffin, A. and Kahn, K.B.2009. “Trends and drivers of success in NPD practices: Results of the 2003 PDMA best practices study.” Journal of Product Innovation Management 26: 3–23.

² Nieto-Rodriguez, A. and Vargas, R.V. Feb. 2, 2023. “How AI will transform project management.” Harvard Business Review https://hbr.org/2023/02/how-ai-will-transform-project-management

³ Jyoti, R. and Riley, S. July 2022. “AI strategies view 2022.” IDC Research Inc. AI StrategiesView 2022: Executive Summary (idc.com)

⁴ Palzer, S. Nov 29, 2022. “Meaningful innovation to unlock growth.” Nestlé Investors Seminar, Barcelona, Spain investor- seminar-2022-innovation-transcript.pdf (nestle.com)

⁵ Agrawal, A., Gans, J. and Goldfarb, A. 2018. Prediction Machines – The Simple Economics of Artificial Intelligence. Harvard Business Review Press: Boston, MA. ISBN:978-1-63369-567-2

⁶ Based on a quote from of Dr. David BenDaniel in an interview by the author in the mid 1970s; BenDaniel founded GE’s Technical Ventures operation (spin-offs from GE) at the GE Labs in Schenectady, NY. Cited in print in: Cooper, R.G. Nov. 1996. “Overhauling the new product process.” Industrial Marketing Management 25(6): 465–482. See also endnote 7.

⁷ Cooper, R. 1976. Winning the New Product Game. McGill University Press: Montreal, QC, Canada.

⁸ Nel, N. Feb. 6, 2023. “40 AI apps to streamline each stage of the product lifecycle.” Product Hunt 40 AI apps to streamline each stage of the product lifecycle | Product Hunt

⁹ Brem, A., Giones, F. and Werle, M. Feb. 2023. “The AI digital revolution in innovation: A conceptual framework of artificial intelligence technologies for the management of innovation.” IEEE Transactions on Engineering Management 70(2): 770–776 The AI Digital Revolution in Innovation: A Conceptual Framework of Artificial Intelligence Technologies for the Management of Innovation | IEEE Journals & Magazine | IEEE Xplore

¹⁰ Cooper, R. G. and Sommer, A.F. 2023. “Dynamic portfolio management for new product development.” Research- Technology Management 66(3): 19–31 https://doi.org/10.1080/08956308.2023.2183004

¹¹ Capterra. 2023. “Social media scanning tools software 2023.” Social Media Scanning Tools Software – Review Leading Systems (capterra.com)

¹² Palzer, endnote 4.

¹³ Bilgram, V. and Laarmann, F. June 2023. “Generating innovation with generative AI: AI augmented digital prototyping and innovation methods.” IEEE Engineering Mgmt Review 51(2): 18–25, Accelerating Innovation With Generative AI: AI-Augmented Digital Prototyping and Innovation Methods | IEEE Journals & Magazine | IEEE Xplore

¹⁴ Filippi, S. Aug. 21, 2023. “Measuring the impact of ChatGPT on fostering concept generation in innovative product design.” Electronics 12(16): 3535, https://doi.org/10.3390/ electronics1216353, Electronics | Free Full-Text | Measuring the Impact of ChatGPT on Fostering Concept Generation in Innovative Product Design (mdpi.com)

¹⁵ Bilgram and Laarmann, endnote 13.

¹⁶ No author. May 31, 2023 .“How generative AI is streamlining the concept phase in new product development.” Fast Company, How generative AI is streamlining the concept phase in new product dev (fastcompany.com)

¹⁷ Roch, J. Oct. 12, 2022. “From Hot Wheels to handling content: How brands are using Microsoft AI to be more productive and imaginative.” Microsoft News, From Hot Wheels to handling content: How brands are using Microsoft AI to be more productive and imaginative – Source

¹⁸ Columbus, L. July 9, 2020. “10 ways AI is improving new product development.” Forbes https://www.forbes.com/sites/louiscolumbus/2020/07/09/10-ways-ai-is-improving-new-product- development/?sh=27a90ce45d3c

¹⁹ Capterra. 2022. “Top ten marketing analysis tools: Marketing analysis software.” Marketing Analytics Software – Review Leading Systems (capterra.com)

²⁰ Helrish, T. Aug 12, 2022. “10 ways to use AI for a successful product launch.” Forbes 10 Ways To Use AI For A Successful Product Launch (forbes.com)

²¹ Nieto-Rodriguez and Vargas, endnote 2.

²² Burridge, N. “Artificial Intelligence Gets a Seat in the Boardroom,” NikkeiAsia, May 10, 2017, Artificial intelligence gets a seat in the boardroom – Nikkei Asia

²³ Nieto-Rodriguez & Vargas, endnote 2.

²⁴ Forsey, C. June 26, 2023. “How AI will revolutionize product development, and how to prepare.” HubSpot How AI Will Revolutionize Product Development, and How to Prepare [Insights from AWS’ Senior Advisor to Startups] (hubspot.com)

²⁵ Nieto-Rodriguez and Vargas, endnote 2.

²⁶ Eastwood, B. March 6, 2023. “Artificial intelligence can help design more appealing cars.” MIT Management Artificial intelligence can help design more appealing cars | MIT Sloan

²⁷ Nieto-Rodriguez and Vargas, endnote 2.

²⁸ Huang, S., Wang, G., Lei, D., and Yan, Y. 2022. “Toward digital validation for rapid product development based on digital twins: A framework.” International J. Advanced Manufacturing Technology 119: 2509–2523 https://doi.org/10.1007/s00170-021-08475-4

²⁹ Bogaisky, J. March 6, 2019. “GE says it’s leveraging Artificial Intelligence to cut product design times In half.” Forbes GE Says It’s Leveraging Artificial Intelligence To Cut Product Design Times In Half (forbes.com)

³⁰ No author. May 10, 2021. “Product design gets an AI makeover.” MIT Technology Review

https://www.technologyreview.com/2021/05/10/1024531/product-design-gets-an-ai-

makeover/#:~:text=In%20an%20effort%20to%20streamline%20its%20AMT%20development,and%20connect%20icons% 20to%20graphically%20create%20a%20model

³¹ Overby, S. 2023. “8 examples of Artificial Intelligence in action.” SAP

https://www.sap.com/norway/insights/viewpoints/eight-examples-of-artificial-intelligence-in-action.html

³² Brossard. M., Gatto, G., Gentile, A., Merle, T., and Wlezien, C. , Feb. 5, 2020. “How generative design could reshare the future of product development.” McKinsey and Company Report How generative design could reshape the future of product development | McKinsey

³³ No author. 2023. “Accelerating the breadth & depth of AI in a physics+ world.” GE Research  Industrial AI | GE Research

³⁴ No author. Aug. 17, 2022. “Atomwise signs strategic multi-target research collaboration with Sanofi for AI-powered drug discovery.” Atomwise press release  Press Release – Page 5 – Atomwise

³⁵ No author. June 12, 2023. “Unilever and Arzeda use AI to develop performance-boosting enzymes.” Unilever News Milestone reached in Unilever and Arzeda partnership | Unilever

³⁶ Kurtzman, L. Jan. 26, 2023. “AI technology generates original proteins from scratch.” University of California San Francisco press release AI Technology Generates Original Proteins from Scratch | UC San Francisco (ucsf.edu)

³⁷ Nieto-Rodriguez and Vargas, endnote 2.

³⁸ Reddi, L.R. March 26, 2023. “How AI will help project planning.” Project Management Institute report ProjectManagement.com – How AI will help Project Planning

³⁹ Dominguez, L. April 21, 2023. “How Unilever expedites product innovation with AI, automation, and robots.” Consumer Goods Technology, How Unilever Expedites Product Innovation With AI, Automation, and Robots | Consumer Goods Technology

⁴⁰Dilmegani, C. May 9, 2023.“Digital Twin applications: Use cases by industry in 2023.” AIMultiple 15 Digital Twin Applications/ Use Cases by Industry in 2023 (aimultiple.com)

⁴¹ Nieto-Rodriguez and Vargas, endnote 2.

⁴² Forsey, endnote 24.

⁴³ Forsey, endnote 24.

⁴⁴ Goyal, C. July 3, 2023. “How to create an AI marketing strategy?” Analytics Vidhya How To Create an AI Marketing Strategy? – Analytics Vidhya

⁴⁵Deeb, G. Sept. 6, 2023. “Artificial Intelligence is taking over marketing.” Forbes Artificial Intelligence Is Taking Over Marketing (forbes.com)

⁴⁶ Cohen, B. Oct 12, 2020. “Amazon’s secret to AI-powered product recommendations.” MDM Distribution Intelligence Amazon’s Secret to AI-Powered Product Recommendations (mdm.com)

⁴⁷ No author. Mar. 1, 2022. “Leveraging technology to improve the lives of P&G consumers.” P&G Leveraging Technology to Improve The Lives of P&G Consumers (pg.com)

⁴⁸ Marr, B. July 25, 2023.“The future of manufacturing: Generative AI and beyond.” Forbes The Future Of Manufacturing: Generative AI And Beyond (forbes.com)

⁴⁹No author. 2023. “What is a ‘Digital Twin’? An introduction to the NASA technology.“ Preface What is a ‘Digital Twin’? An

introduction to the NASA technology | Preface

 

____________

Dr. Robert G. Cooper est professeur émérite à la DeGroote School of Business, à l’Université McMaster de Hamilton en Ontario, “Distinguished Research Fellow” au Institute for the Study of Business Market au Penn State University’s Smeal College of Business Administration, aux États-Unis.

C’est un expert international dans le domaine de la gestion du développement de nouveaux produits. Il est l’un des compagnons Crawford du Product Development & Management Association (PDMA). Bob est le créateur du Processus Stage-Gate, largement utilisé par les entreprises leaders mondiales qui mettent en marché de nouveaux produits. Il a publié plus de 130 articles dans des revues spécialisées ce qui lui a mérité trois fois le prestigieux prix Maurice Holland du Industrial Research Institute à Washington.

Robert Cooper est un ami reconnu de l’IDP qui l’a accueilli plus d’une fois pour donner des formations et des conférences marquantes au Québec. Ces ateliers ont eu une influence déterminante sur la façon de développer des produits à succès pour plusieurs membres de l’IDP. Site web : www.bobcooper.ca ; Contact : robertcooper@cogeco.ca