AI-powered synthesis planning: revolutionizing sustainability and efficiency

The world of chemical synthesis is becoming more complex every day. However, a recent webinar — AI in Action: Sustainability and Efficiency in Chemical Synthesis opens in new tab/window — offers good news. Innovative state-of-the-art systems are making synthesis more efficient, environmentally friendly and commercially viable.

The growing challenge of API complexity

How Lonza is using AI to transform route planning

Lonza’s award-winning AI-enabled route scouting service opens in new tab/window tackles these challenges head-on. Their experts have integrated their own custom building block library with Reaxys’s extensive database and predictive retrosynthesis capabilities to create a remarkable four-step solution:

1. They start by meeting with customers to establish priorities across both commercial and technical parameters.

2. Then they generate multiple potential synthesis routes using Reaxys’s predictive retrosynthesis tools alongside input from Lonza’s expert chemists.

3. Human experts are then called in to evaluate and rank these proposed routes based on the agreed priorities.

4. Finally, Lonza delivers structured reports to the customer with proposed routes for follow-up R&D work, detailed supply chain intelligence with pricing trends, regional availability and qualified suppliers.

The results speak for themselves. In one case study, they transformed a seven-step synthesis involving seven isolations into a streamlined four-step route with only four isolations. This saved four to six weeks of lab time, reduced the amount of required starting materials by 70%, and identified starting materials that were 50% cheaper.

Building sustainability into every step

Dr Alexei Lapkin opens in new tab/window, Director of Chemical Data Intelligence (CDI) opens in new tab/window and Professor of Sustainable Reaction Engineering at the University of Cambridge, is advancing the field significantly. He is redefining what chemistry should aspire to be.

“We want chemistry that doesn’t channel fossil carbon into the atmosphere, doesn’t harm biodiversity, and creates products that are non-toxic and safe throughout their entire lifecycle,” Alexei explains.

Using Reaxys data, his team is incorporating green chemistry principles into Computer-Aided Synthesis Planning (CASP) to create more sustainable synthesis routes. “Our design vector is now much bigger,” Alexei notes. “We're looking at performance, price, mass intensity, biodegradability, toxicity, availability of materials, resilience, ability to manufacture, hazards and reputational factors.”

One of their most impressive innovations tracks the origin of atoms in final molecules, enabling them to calculate renewable and circular carbon percentages. While still in development, this technology allows for automated evaluation of synthesis routes against emerging sustainability standards.

Read our interview with Prof Lapkin: “Sustainability is the only thing worth pursuing”

The human element remains essential

All the experts underline one crucial point: AI tools enhance human expertise rather than replace it. As Dr Timur Madzhidov opens in new tab/window, Senior Product Manager for Elsevier's Reaxys, aptly puts it: “The chemist with AI will replace the chemist without AI.”

Lonza’s approach exemplifies this philosophy of creating a greater whole by leveraging the strengths of both humans and machines. “It’s a semi-governed service. It’s absolutely not like pressing a button,” Simon explains. “Process chemists are still very much needed — more than ever because of the increasing complexity we see from discovery colleagues.”

Sustainability and profitability: two sides of the same coin

According to the webinar guests, “greener” synthesis offers a wide range of benefits:

• Protecting worker health and safety

• Minimizing environmental damage from emissions and waste

• Avoiding untreatable or toxic waste

• Reducing embedded carbon through renewable or circular feedstocks

• Enabling informed discussions with suppliers about embedded emissions

One key insight from the panel was that sustainability and profitability aren’t mutually exclusive. “Sustainability includes economics. It’s not either/or — it’s and/and,” Alexei says. Finding mutually beneficial solutions is both possible and necessary.

Simon reinforces this with concrete examples from Lonza: “We’re talking about dozens of thousands of tons. If we recycle 20% to 25% of all solvents ... it’s huge, the amount that you can save. And it’s also compatible with cost because if you save solvents, that’s money you don’t need to spend.”

Invisible tools making a visible impact

Looking ahead, the experts highlighted several exciting developments on the horizon:

• Automated full lifecycle assessment (LCA) calculations for proposed routes

• Integration of circularity metrics as standards evolve

• Manufacturing robots working around the clock to accumulate material

• Expanded supply chain intelligence capabilities

• More specific, targeted AI tools for chemistry tasks

This webinar demonstrates that AI in chemical synthesis has moved beyond theoretical promise to practical implementation, delivering measurable benefits in sustainability and efficiency.

And the time is very much now. “A couple of years ago, people were jumping on every new tool and trying to solve any problem with complex AI solutions,” notes Alexei. “I think the community has moved on from that. There has been a lot of development of very specific, target-oriented AI tools that are way more accurate and precise.”

According to Timur, many AI solutions are already embedded in everyday tools, often without the end user’s knowledge. “And this is also part of our task — to make it invisible” while still visibly improving outcomes.