The Adoption of AI in New Product Development

Abstract

Overview Artificial intelligence (AI) is transforming all aspects of business, including new product development (NPD). Early adopters of AI for NPD are reaping substantial rewards, witnessing remarkable reductions in development timelines, and experiencing a more rapid pace of innovation. But what about the more typical firm? This article presents the preliminary results of a study of US and Central European companies, shedding light on the current state of AI implementation across 13 key AI applications in NPD. We found that companies studied have not yet fully embraced AI across any of these applications; 75 percent have failed to adopt AI for a single application; and “intention to adopt” for all 13 applications is quite weak. European firms are only marginally ahead of US firms. We conclude with important messages for management and offer six recommendations firms can use in their efforts to integrate AI into NPD. A note of caution: Companies will need to act quickly in their AI adoption efforts; otherwise, they risk losing out.

Keywords::

• Artificial Intelligence
• New Product Development
• Product Development
• Product Innovation

The current approach to new product development (NPD) mirrors decades-old methods, resulting in persistently low success rates, typically around 25–30 percent (Knudson et al. 2023; Barczak, Griffin, and Kahn 2009). This is about to change, however, as the artificial intelligence (AI) revolution takes hold in NPD (Nieto-Rodriguez and Vargas 2023). Early adopters of AI are already reaping substantial benefits. AI is revolutionizing how companies conceive, develop, and launch new products. Experts expect AI will leave an indelible mark on the trajectory of product innovation.

Lamarre et al. (2023) point out that despite the claims about the positive impact of AI on NPD performance, “hard evidence that directly ties digital and AI transformation to improvements in operational KPIs and financial performance is scant.” Armed with significant resources, industry giants like GE, Siemens, Pfizer, and Nestlé are spearheading the early adoption of AI. What about more typical firms? Are they using AI to similarly revolutionize their NPD systems? Have they begun to adopt AI for NPD, and if so, for which of the many possible applications? How does adopting AI impact NPD performance?

We wanted to explore how more typical companies are faring. This article investigates how companies are integrating AI for 13 applications in NPD (see “The 13 Applications” on page 43). We studied companies’ current state with regard to AI adoption in NPD, improvements AI has already brought, future intentions to integrate AI into new product processes, and how adopting AI for different applications impacts NPD performance, including productivity, innovation acceleration, and agility.

Background: AI Applications and Results

Early-adopter firms have implemented AI in their businesses for various reasons, such as improving efficiency, business agility, or productivity. Thirty-five percent of these early adopters globally report that increased innovation is the primary benefit of adopting AI (Jyoti and Riley 2022). Early adopters of AI for NPD also highlight its immense potential in NPD and substantial payoffs overall, including a 50 percent reduction in development times or a significant acceleration in the pace of development (Cooper 2023).

Remarkable applications of AI in NPD abound (Cooper and McCausland 2024; Cooper 2023). GE designs turbine blades and tests virtual prototypes in minutes, cutting turbine development time in half (Nieto-Rodriguez and Vargas 2023). Siemens uses digital twins to test products during field trials and to monitor performance after launch. Moderna uses AI to design new mRNA molecules for vaccines and drugs (Euchner 2020; Iansiti and Lakhani 2020), while Pfizer employs AI for drug discovery. Nestlé leverages AI to create new product ideas and concepts by scanning the Internet and also mining technical data, increasing its pace of innovation by 60 percent (Palzer 2022). In the front end of the new product process, AI products like AIM Blueprinter improve the productivity of B2B voice-of-customer (VoC) interviews, cutting research times in half; Applied Marketing Science scans the Internet to identify customers’ points of pain, leading to innovative product concepts; and Ai Palette creates and tests new CPG product concepts online.

AI in NPD yields various benefits, including improving design thinking (Garbuio and Lin 2021), enhancing the effectiveness of open innovation (Ferràs, Nylund, and Brem 2023), generating better product ideas and concepts (Bouschery, Blazevic, and Piller 2023; Filippi 2023; Joosten et al. 2024), augmenting digital prototyping (Bilgram and Laarmann 2023), accelerating physical development and testing (Nieto-Rodriguez and Vargas, 2023), fostering more effective development teams (Bouschery, Blazevic, and Piller 2023), and facilitating better go/no-go decision-making (Cooper 2023).

Much of the evidence for building a strong business case for AI in NPD is based on single case examples, broad brush surveys by major consulting firms, and expert opinion. A few robust studies, however, do exist. An extensive analysis of almost 1,000 news articles revealed that AI used in product innovation yields “significantly lowered search costs and increased speed in R&D processes” (Johnson et al. 2022, p. 9). A study of 558 new product projects revealed that those projects where AI was used had more than double the success rates as the others (Zhang, Zhang, and Song 2021). In the latter study, however, AI was very broadly defined to include smart personal assistants, language translators, and chatbots.

There is a dearth of robust studies regarding AI’s benefits in NPD. At the time of writing, there were no comprehensive scholarly studies of AI’s impact on NPD results nor on its effect on economic or key performance indicators for the types of applications one normally sees recommended for AI in NPD—for example, the impacts of using AI for idea generation, market analysis, molecular discovery, product design, prototyping, product testing, and digital twins. This study aims to shed light into how non–early adopter companies are using AI in NPD and the impact they’re seeing.

The Research Study

We conducted a multi-firm study comprising 44 companies from the US and Central Europe. We had four objectives: assess the extent of AI adoption for NPD; identify types of applications in NPD where AI is applied; examine the results of employing AI in NPD, such as reducing time-to-market or improving decision-making; and evaluate businesses’ intent and readiness to adopt AI.

We undertook an online survey of companies from three groups: US members of the Institute for the Study of Business Markets at Penn State University; members of Quer-Kraft e.V., a German association of innovation and NPD managers (includes some from other EU countries); and members of the Innovation Research Interchange. Forty-four businesses completed the online questionnaire (approximately 40 percent from the EU and 60 percent from the US).

Given the time limitations of an online survey, we selected 13 of the more prominent applications from the comprehensive list of 40 from the “AI in NPD” map (see Figure 1 in Cooper and McCausland [2024] for a model that positions AI in NPD; the model dimensions are from Brem, Giones, and Werle [2023] and Nel [2023]). Questions that gauged the use of or “intent to use” each of the 13 AI applications had 1–5 scales, where 1 indicates “No, not at all,” and 5 indicates “Yes, very much so.” Similar 1–5 scales captured the various measures of firms’ readiness to adopt AI. Other questions sought percentage responses, such as percentage improvements realized on each of five performance metrics.

Figure 1. Current and expected future use of AI in the front end of NPD (percent of businesses that are now using or intend to)

Data analysis included converting the 1–5 scaled responses to percentages—that is, the percentage of respondents that answered 4 or 5 out of 5 were counted as a “yes” to determine the percentage of firms using (or intending to use) each application. We used a two-tail t-test to check for significant differences in the mean usage rate (percentage) of AI for EU versus US businesses. Correlational analysis (Pearson product-moment correlations) were undertaken between the use of each application and each of the five-performance metrics.

Businesses represented various industries, including chemicals, ingredients, materials, and mechanical products. All firms in the study were larger; no startups or small and medium-sized enterprises (SMEs) were included. The business unit is the unit of analysis because respondents cannot be expected to be familiar with or assess company-wide AI adoption practices.

Results: AI Usage in NPD

Of the 44 businesses studied, almost one quarter (22.7 percent) are using AI for at least one application in NPD (31 percent usage for EU firms, 18 percent for US firms). The AI usage rate reported in this study is almost double the 13 percent rate cited in a McKinsey & Company (2023) report. The increase is a reflection of the adoption that has taken place in one year since early 2023.

No significant differences exist in AI usage for specific applications or performance between US and EU firms. Accordingly, the data from both groups are combined for analysis. While there is a consistent tendency for a somewhat higher proportion of EU firms to be using AI compared to US firms, these differences are not significant.

AI Applications in the Front End of NPD

The front end of the NPD process presents ample opportunities for leveraging AI, especially for originator or more creative types of AI applications. Despite this potential, this study reveals a stark reality for typical businesses (Figure 1). Notably, AI usage in the front end is minimal, with all applications well below 10 percent usage, except for market analysis and competitive analysis. Go/No-go decision-making using AI is non-existent. Future intentions show a much higher proportion of businesses planning to use AI for NPD, signaling a major uptick in adoption.

Takeaways: The limited use of AI in the front end of innovation is surprising given the numerous AI tools and services available for use, as well as the number of AI tools or services available to undertake customer research or online scanning to identify new product needs, ideas, and concepts (Cooper and McCausland 2024). For example, the total non-use of AI to make investment decisions in NPD projects found in this study is directly opposite to the finance industry. According to Burridge (2017), “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.” Even better, many of the available tools are commercially available and therefore require no software development by the user firm, do not require many AI skills in IT, and are relatively low risk to try.

AI Applications in the Development and Testing Stages

As the NPD process progresses into the development and testing stages, adoption rates decline further. Here the “adoption rates” across five key applications are consistently low, averaging 4.2 percent (Figure 2). Notably, AI for product design (advanced forms of computer-aided design [CAD]) emerges as the most highly used application. The “intention to use” adoption rates are somewhat lower than for the front-end activities, averaging 26 percent.

Figure 2. Current and expected use of AI in the development and testing stages of NPD (percent of businesses that are now using AI or intend to)

Takeaways: The relatively low usage of AI for NPD in the middle NPD stages—product design, prototyping, and testing—is unexpected. AI’s absence is noteworthy, especially given the many AI solutions commercially available, plus the abundance of online user case studies showing the applications and the dramatic results:

• Siemens Digital Industries software, available as a service, includes Xcelerator (digital twins) for product design-and-test iterations and Tessent for IC product testing.

• Numerous suppliers of similar AI-powered engineering design software exist, such as PTC, Autodesk, Nemetschek Group (parent of Graphisoft and Vectorworks), and Dassault Systems (Kevin 2023).

• AI tools from firms such as Atomwise or Arzeda “discover” new molecules for the pharmaceutical, chemical, or process industries.

The development and testing stages are ripe with opportunities for the deployment of AI.

Adoption of AI for Other NPD Tasks

The study also explored three additional applications of AI: project management, portfolio management, and natural language processing (NLP):

• AI for project management entails creating project plans, optimizing Gantt charts, monitoring the progress of projects, and identifying problems or late projects.

• AI for portfolio management includes tracking projects in the portfolio, identifying potential problems, prioritizing projects, and automated preparation and distribution of project reports.

• NLP involves extracting information from unstructured data sources, such as customer reviews, social media posts, and market research interview.

The use of AI for both project and portfolio management shows even lower usage patterns; in fact, no current usage at all (Figure 3). Additionally, intentions to adopt are also fairly low, both close to 20 percent.

Figure 3. Current and expected use of AI in portfolio management, project management, and NLP (percent of businesses that are now using AI or intend to)

NLP is the one application with some traction, with about 11 percent of businesses now using it. NLP also rates near the highest “adoption intention” of all 13 applications. NLP has considerable potential in NPD for analyzing customer feedback, VoC interviews, and online blogs. NLP analyzes unstructured data, draws conclusions, and creates charts and graphs. NLP and AI applications can even generate new product concepts from this type of unstructured data analysis. The “adoption intent” for NLP is also higher than for the other applications studied.

Takeaways: Given the number of readily available and effective AI products, the low level of adoption of these AI tools was surprising. Portfolio management models analyze massive amounts of project data quickly and can extract meaningful insights that may be invisible to the human eye. With its predictive analytics capability, AI can forecast future outcomes (Janusz 2023). Companies have used portfolio optimization models for 70 years in the financial community (Markowitz 1952), yet they find little acceptance in NPD, even though the nature of the problem is much the same: select the right projects (investments) to maximize expected profits, given a resource constraint. Similarly, numerous AI-based project management tools are available, as noted by the Project Management Institute (Reddi 2023). Information management during a project is always a challenge; however, solutions exist, such as Albert Invent.

NLP is an obvious candidate for an NPD project team to adopt. Solutions available range from comprehensive products such as Microsoft’s Azure to more specific AI products like MonkeyLearn (Wolff 2023).

Nature of the Applications

Given that the businesses studied do not use AI extensively, the nature of specific applications that AI is used for yields a fairly limited but varied list (Table 1).

Table 1. Types of AI applications uncovered

Type of AI Application	Specific Applications Used by the Business Units Studied
Seeking market and other information online	Gaining insights and seeing innovations online Monitoring traditional media and social media Digital marketing analytics Technical information extraction
Marketing communications, sales force, and customer service	Creating advertising or promotion copy text Providing stock images, including generative imaging (AI creating images) Analyzing the need and preparing for customer meetings Customer service After sales support
Discovery and development	Predictive formulations (e.g., for chemical or materials formulation)
Operations	Operations and supply chain process optimization (several businesses) Production and logistics
Data management	Extraction of trends from complex data streams Knowledge management Mona AI (monitoring AI) RPA (robotic automation platform)—business process automation
Other	Brainstorming Fraud detection (several firms) Input detection and routing, text generation Enhancing optical character recognition (OCR) results for next process step

Takeaways: The list of apps differs from the mainstream types of products, solutions, and applications for NPD most often seen online in articles, blogs, and vendors and early-­adopter websites. The list also resembles an assortment of applications of one or two per business, apps that appear to have been adopted on a piecemeal basis rather than a comprehensive and integrated set of AI solutions based on a strategic approach.

Performance Results and Impact of AI

We present performance improvements achieved by using AI in NPD when gauged across five metrics (Figure 4). Performance results are modest, averaging just 25 percent improvement across the five metrics, with none standing out. This result is not surprising, given the low level of adoption of AI applications within the firms studied.

Figure 4. Improvements realized by using AI for NPD (percent improvement)

We also explored the specific impact of each AI application on NPD performance, operating on the hypothesis that certain applications in particular will exert a positive influence. A challenge encountered in this impact analysis is the narrow distribution of responses at the “low end” of the scale for AI usage. With fewer than 10 percent of responses at the midpoint or above (3 out of 5 or higher on the 1–5 usage scales) for most applications, conducting correlation analysis became challenging, and significantly weakened the correlations.

Despite this limitation, some AI applications do demonstrate remarkably strong positive impacts on performance results. The standout impacts are on reduced time-to-market and better decision-making, with few and limited impacts of AI applications on the other three performance metrics. We highlight those applications with stronger impacts on the two performance metrics—speed and decision-making—along with their significant correlations (one tail t-test, significant at the 0.01 or 0.05 level, as noted in Figure 5; we use a one tail t-test, as the underlying hypothesis is single directional—that is, that AI will positively affect performance). We present four performance-enhancing applications, beginning with the more impactful:

Figure 5. Performance impact of AI on improved decision-making and speed for different NPD applications (correlations)

• AI for product testing and validation—encompassing automated product testing, virtual market testing, and in-use testing—has strong impacts on both speed and improving decision-making (both strongly significant, the effect on speed being the strongest driver of all applications with a correlation of 0.61). Interestingly, product testing is not among the heavier used (or intent to use) applications.

• AI for product design—employing AI with CAD to design and optimize products—yields positive impacts across all metrics (except greater agility), with particularly strong and significant effects on speed and better decision-making (both significant at the 0.01 level, one-tail t-test).

• AI for prototyping, involving rapid prototyping and automatic translation of drawings into prototypes, has a substantial impact on speed.

• Extensive AI usage overall in NPD correlates significantly with all performance metrics, except enhanced agility. High-user firms achieve shorter times to market, a strong impact with a correlation of 0.48, and also better decision-making.

While other AI applications contribute positively to the two performance metrics, none exhibits as robust an impact as those listed above. The use of AI for simulation modeling—namely, the use of AI for simulating a product, process, or system, and the use of digital twins—impacts positively on both speed and improved decision-making, but only at the 0.05 significance level. Note that the use of AI in the front end of the NP process, although frequently mentioned in articles and blogs, has subdued impacts on both performance metrics. Only idea generation and competitive analysis using AI significantly affect both speed and improved decision-making, but only at the 0.05 level. The same is true of using NLP, often mentioned as a good application in the literature. Using AI for building the business case speeds up projects—its only positive effect—but again only significant at the 0.10 level (and thus is not shown in Figure 5). Not surprisingly, given their limited use, neither of the two major NPD management tools—AI for portfolio management and AI for project management—have any positive effects on performance.

Caveats: The values of the correlations found in this study are attenuated due to the nature of the distribution of responses in terms of AI usage—most businesses are at the very low end of the usage scale. Thus, the performance impacts of AI in the various applications in a broader sample of firms containing more adopters are likely higher than those in Figure 5. Further, note that correlations do not prove causality; higher performance could lead to AI usage, the reverse of the causality we inferred. However, since action usually precedes a result, the logical conclusion is that AI usage leads to the positive effects.

Readiness to Adopt AI

Given that the businesses studied have not begun to implement AI extensively, a major concern now is whether they are ready to do so. We asked “adoption readiness” questions and present the results (Figure 6). The first two items reconfirm the aforementioned findings—that these businesses have not embraced AI extensively in NPD, and consequently, they have not witnessed substantial performance enhancements in NPD.

Figure 6. Readiness to adopt AI for NPD (percent of businesses answering 4 or 5 on 1–5 scale)

How businesses fare on the next four readiness items is somewhat troubling: a modest to weak commitment and low willingness to adopt AI. Two-thirds of businesses have not agreed to accept the concept of the AI technology for NPD; and the executive sponsor, if there is one, lacks the capability and credibility to lead this transformation in more than two-thirds of businesses. Only the fact that senior leaders are prepared to wait for up to a year to see the benefits of AI is a positive sign, although this result is not exceptionally robust, with just over half of the businesses expressing this patience.

The very strong negative is that management is not willing to relinquish decision-making to AI. This one indicator alone could mean a tough adoption journey for many of the AI applications in NPD. Unique to the implementation of AI is the need to build trust—especially in line managers and executives—that data-driven decisions are better than gut feel, intuition, or traditional methods (Nunez 2021). Thus, managers and executives must understand what AI is and how it works, and learn to trust in the AI-produced data and also to “yield decision-making authority to an analytical process.”

Managerial Implications

Based on the study findings, we identified six recommendations for practitioners regarding the adoption of AI in NPD:

1. Educate and foster AI awareness.—Educate your management team and others about AI and its potential applications in your business operations. Develop a foundational understanding of AI concepts, terminologies, and technologies to facilitate informed decision-making and strategy development (Cooper and McCausland [2023] provide many resources about AI applications in NPD).

2. Engage with AI experts or consultants.—Consider engaging with AI experts, consultants, or specialized AI companies to assess your specific needs and develop a tailored AI strategy. Experts can help you understand the potential payoffs, recommend suitable AI technologies and vendors, and formulate a roadmap for integrating AI into your NPD process.

3. Focus.—Take a focused and stage-wise approach, gaining experience and learning along the way. Select one functional area or one major process within the business unit as a starting point. Operations or manufacturing may have already progressed with AI (Marr 2023), but NPD or research, development, and engineering (RD&E) could be next. RD&E professionals already deal with technology, making adopting new AI technologies less foreign; the NPD process has already delivered very positive AI results from early adopters; and AI-related job losses are not a major issue in RD&E (Johnson et al. 2022; McKinsey 2023).

4. Balance risk and size.—Do not adopt a piecemeal approach, implementing an application here and there with no sense of the ultimate destination. Also avoid the other extreme of undertaking a massive, enterprise-wide AI transformation, which many larger consulting firms recommend; such an approach is too ambitious for most firms. AI often fails in practice: 85 percent of AI machine learning projects fail to deliver, and only 53 percent of AI projects proceed from prototyping to final implementation (Nunez 2021).

5. Develop an AI strategy for AI for NPD.—Developing a comprehensive and longer-term strategy is vital for integrating AI into NPD. Your AI initiative must be driven from the top, beginning with executive vision, commitment, and sponsorship. Establish a task force, comprising a handpicked, dedicated team, to move the initiative forward. Use a “technology acquisition” process model, familiar to many RD&E professionals, to guide the initiative (Brem and Cooper 2024). Build in best practices, such as voice of business (VoB) and voice of process (VoP); create a fact-based business case; generate a timeline roadmap for multiple AI applications in NPD; use pilots for applications to test and learn; and ensure that senior management is fully engaged at the key decision points.

6. Start now; it’s time to take action.—The inflection point of the AI adoption curve likely occurred in early 2022 (Jafari 2022), with the peak expected before the end of this decade. That means 50 percent of adopting businesses will be on board with AI by the end of the decade (Cooper and McCausland 2024; Cooper 2024). The much-anticipated artificial general intelligence (AGI), capable of cognitive tasks on par with humans, is also expected around 2028 (Losey 2023; Jafari 2022). Anticipate a sharp rise in the use and realized benefits of AI in NPD in the next decade.

Now is not the time to be indecisive and waiting on the sidelines. Leading early adopters embraced AI several years ago, and have seen impressive results in NPD applications, surpassing the early days of Lean and Agile NPD methods. It takes time to get up and running with AI in NPD. Given the low adoption rate of AI in NPD revealed in this study, and the lack of commitment to move forward, many firms are putting themselves at serious risk by waiting.

Future Research

The study’s sample size is only 44 businesses. We encourage researchers to do additional large-sample studies, looking at AI applications and their impacts, and also at different types of firms (startups, SMEs, and multinational corporations). A larger sample will permit more in-depth analysis of impacts, success factors, motivational issues, and barriers.

Conclusion

The use of AI in various NPD applications within the businesses studied is surprisingly low. Considering the recent attention AI has garnered, especially with some early adopters prominently showcasing their AI journeys, we might expect more widespread adoption. Lack of readiness to adopt AI for NPD is a serious concern. Many companies have disclosed that AI has a strong positive impact on both accelerated development and improved decision-making for many NPD applications. The study results point to the need for senior management to create a vision, confirm their commitment, and embark on this vital AI journey. The recommendations presented offer practitioners guidance on how to begin and where to focus.

The 13 Applications

These are the 13 applications referred to in this study and in the figures.

Building the Business Case: Analyze financial data, make sales and profit projections, do multi-scenario analysis

Competitive Analysis: Scan online for data on competitors’ performance, products, pricing; monitor activities such as product launches, pricing changes, and marketing campaigns

Go/No-Go Decisions: Make project investment decisions at gates, analyze data from previous projects to identify patterns of success

Idea Generation: Scan the internet and social media for themes and “hits,” use AI to suggest ideas

Laboratory Automation: Lab robots doing chemistry/biology, predict new drugs’ bioactivity (drug discovery), optimize conditions for chemical reactions, synthesis of new compounds

Market Analysis: Scan and analyze market data (online data, reviews, social media, sales data), ML algorithms monitor market trends, competitor activities, and customer preferences

Natural Language Processing (NLP): Extract information from unstructured data sources (customer reviews, social media, and market research reports)

Product Design: AI with CAD to design products, optimize product design, regenerative design, create multiple design options

Product Testing: Automated product testing, virtual market testing (simulation), in-use product testing and trials

Project Management: Create project plans, optimize Gantt charts, monitor progress of projects, identify problems (late projects)

Portfolio Management (support for the PMO): Track projects in the portfolio, identify potential problems, automate preparation and distribution of project reports

Prototyping: Rapid prototypes using AI (faster, more efficient), translate sketches into prototypes automatically

Simulation AI Models: Simulate the behavior of a product, system, or process; use digital twins

Teasers:

AI usage in the front end is minimal, with all applications well below 10 percent usage, except for market analysis and competitive analysis.

The use of AI for both project and portfolio management shows even lower usage patterns—in fact, no current usage at all.

The standout impacts of using AI applications in NPD are on reduced time-to-market and better decision-making.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Robert G. Cooper

Robert G. Cooper is ISBM Distinguished Research Fellow at Penn State University’s Smeal College of Business Administration, Professor Emeritus at McMaster University’s DeGroote School of Business (Canada), and a Crawford Fellow of the Product Development and Management Association. He has published 11 books and more than 150 articles on the management of new products. He has won the Innovation Research Interchange’s prestigious Maurice Holland Award three times. Cooper holds and master’s degrees in chemical engineering from McGill University in Montreal, Canada, and a PhD in business and an MBA from Western University in London, Ontario, Canada. [email protected]

Alexander M. Brem

Alexander M. Brem is director of the Institute of Entrepreneurship and Innovation Science (ENI), and Chaired Professor of Entrepreneurship in Technology and Digitization, endowed by the Mercedes-Benz Foundation in the Stifterverband, University of Stuttgart, Germany. He is also Honorary Professor of innovation and entrepreneurship at the University of Southern Denmark. He is editor of the IEEE Engineering Management Review, and a senior member of IEEE. He received his MBA and PhD in business administration from the Friedrich-Alexander-University Erlangen-Nuremberg, Erlangen, Germany, in 2004 and 2007, respectively. His research interests include technological innovation and entrepreneurship, and he is a noted scholar in the field of AI and NPD. [email protected]