How AI Is Accelerating Product Development and Empowering Scientists to Innovate Faster

AI and Product Development: Practical Insights from Industry Experts

Companies ready to experiment, adapt, and upskill today are best positioned to shorten product development cycles and bring innovations to market faster. This consensus comes from three industry experts, including a PepsiCo R&D leader, gathered by the Institute of Food Technologists (IFT).

From Rules to Generative Intelligence

The distinction between artificial intelligence (AI) and machine learning (ML) can be confusing, but they are not the same. AI and ML coexist, with ML forming the foundation of many AI systems. Michael Slater, technical director of consulting at Improving, highlighted how the field has shifted from traditional ML models built for narrow tasks to modern generative AI.

Generative AI tools like ChatGPT or Copilot use deep learning architectures that mimic brain-like networks and can synthesize knowledge across domains. This shift allows product development teams to access insights from massive datasets in seconds, accelerating ideation and feasibility analysis. As Slater put it, generative AI “unlocks the ability to search terabytes of information in seconds and return a structured summary.”

However, he cautioned that large language models (LLMs) have limitations. They aren’t trained on peer-reviewed scientific literature and always produce an answer—even when accuracy is uncertain. To mitigate this, Slater suggests agentic workflows where autonomous AI agents make decisions and coordinate tasks with minimal human input, drawing on trusted, domain-specific sources like scientific journals. In this setup, AI acts as a “wingman,” augmenting creativity and automating routine tasks while scientists remain the final decision-makers.

AI as a Toolbox

Jay Gilbert, director of scientific programs & product development at IFT, described AI as a toolbox rather than a threat. His key message: “AI is not coming for your job, but someone with AI is.” Success depends on selecting the right AI tools for the right tasks.

Experimenting with off-the-shelf platforms like ChatGPT is valuable for learning. Still, long-term benefits come from developing or adopting domain-specific AI systems that align with organizational goals, protect intellectual property, and address data privacy risks. Gilbert warned to “always read the fine print” because proprietary information should not be inadvertently included in public AI training datasets.

Trust and transparency are crucial. Businesses need to evaluate who builds the AI models they rely on and understand the data those models use. Gilbert introduced retrieval-augmented generation (RAG) as a practical framework. RAG works like an Amazon warehouse: organizational knowledge is stored and “fetched” by AI when needed, providing referenced, validated responses. This keeps humans, especially scientists, in control during ideation and problem-solving.

Scaling AI in Enterprises

Mohamed Badaoui Najjar, R&D senior director of digital transformation & global specifications at PepsiCo, shared insights on AI adoption in large companies. While big organizations have deep internal expertise, organizing and mapping that knowledge remains challenging.

Najjar emphasized that successful AI integration depends on mindset, adoption, and upskilling. Off-the-shelf tools can support early exploration, but tailored systems and structured digital training are essential for lasting results. He urged teams to start their upskilling journey now, equipping scientists with tools designed for their field.

The ultimate aim is “first-time right” development—shortening product lifecycles by reducing iterations and ensuring alignment with customer value. Simplifying processes enables faster, profitable progress. Najjar highlighted the need for clear functionality design suited to company size and lifecycle stage, as well as smooth handoffs from R&D to commercialization.

The Road Ahead

Despite different perspectives, all three experts agree: AI won’t replace scientists, but those who embrace AI will outperform those who don’t. Current AI systems may produce errors or “hallucinations,” but these represent the earliest stage of the technology.

For product developers, the path forward involves:

• Experimenting with off-the-shelf AI tools
• Building domain-specific workflows based on trusted data
• Protecting intellectual property and ensuring transparency
• Investing in upskilling teams for upcoming digital innovation waves

Done right, AI will make product development not just faster, but smarter, more reliable, and better aligned with customer needs.

For those interested in practical AI training to support this transition, explore courses at Complete AI Training.