Upright Launches New AI Model to Quantify Companies’ Sustainability Impacts, Risks and Opportunities

Impact data company Upright announced the launch of a new vertically-trained AI model designed to enable companies and investors to quantify sustainability impacts, risks, and opportunities.

According to Upright, the new large language model (LLM) bases its impact quantification on scientific evidence, rather than relying primarily on corporate disclosures or general-purpose AI models.

Annu Nieminen, Co-founder and CEO of Upright, said:

“For 20 years, the sustainability industry has asked companies to grade their own homework in sustainability and make business decisions based on those grades. Our model flips the question. Instead of starting with disclosures, we start with the strongest scientific evidence and quantify impacts, risks, and opportunities in monetary terms, in a way that allows for scenario modeling. That turns sustainability from a backward-looking reporting exercise into a future-looking decision-making tool.”

Upright said that the new LLM will power its platform, enabling companies, investors, and advisors to assess impacts, risks, and opportunities for any company, fund, or portfolio in real time.

The company said that the model can be applied to a broad range of sustainability-related applications, including double materiality, climate risk assessment, financial effects, net impact, EU Taxonomy, SFDR PAIs, and UN SDG analysis.

Upright added that the new model will also enable users to combine the platform’s sustainability intelligence with their own reports, operational data, and business context to benchmark against peers, assess risks, challenge existing assessments, and explore the evidence behind impact claims.

Juho Ojala, Co-founder and CTO of Upright, said:

“For impact-size questions, ground truth is sparse and there is often no single correct answer available for training. General-purpose models can therefore sound convincing while becoming inconsistent across their comparisons. We developed our model to maintain coherence across all comparisons and ground its conclusions in scientific evidence.”