AI has the power to support us in taking steps towards positive environmental and social change.
In bringing generative AI to the enterprise, Salesforce have done so with thoughtful consideration. Take, for example, the Einstein Trust Layer that safeguards customer data, and the tenets of Trusted, Ethical & Humane Al that guide all of their innovations. More consideration has been placed into sustainability – in other words, the impact on the environment as a result of the mass processing required for AI.
Sustainability is a core value that Salesforce lives by, and to help other organizations to meet sustainability goals, Salesforce brought Net Zero Cloud to the market. Salesforce follow an ESG (Environmental Social Governance) strategy, and have developed Net Zero Cloud to be an ESG platform. Now, Einstein AI will be incorporated into Net Zero Cloud to combat some of the most challenging aspects of ESG.
Einstein for Net Zero Cloud
There will be generative AI capabilities baked into Net Zero Cloud, Salesforce’s offering for carbon accounting, emissions tracking, calculating building energy intensity, and more.
In a rapidly evolving regulatory landscape, efficiency in getting to compliance is healthy for both the bottom-line, and the planet. However, ESG reporting is typically complex and time-consuming to collate and disclose reports. With this solution, Einstein suggests report content, which alleviates the burden from ESG professionals and is suitable for companies of all sizes and industries.
- ESG Report Authoring, ESG Disclosures, and Auto-Populated Responses: With more ESG regulation on the horizon, companies will need to disclose more information than they’ve had to before. Examples of ESG data include supply chain emissions and workforce representation data. To streamline ESG reporting, Einstein can access historic data, such as 10-K reports, GHD emissions, or company policies, when uploaded to Net Zero Cloud. Available Feb 2024
- CSRD Report Builder: Supports companies to comply with the CSRD regulation, a prominent regulation mandated by the EU. Available Oct 2023
- Materiality Assessment: Identify and prioritize the ESG topics that are most relevant to your business model. Available Oct 2023
AI Accelerator for Nonprofits (AI for Impact)
According to Salesforce Research, nearly three quarters of nonprofit organizations understand the importance of AI as a need to have/must have to achieve their core missions, they lack the resources and knowledge to fully understand, integrate, and invest in it. As a result, these worthwhile organizations are burdened with clunky technology.
In order to close this gap, Salesforce have launched the AI for Impact Accelerator – a $2 million pool of unrestricted grants, one-on-one consulting, and donated Salesforce products. The first cohort comprises six nonprofits and NGOs in the education and workforce development fields, serving US, UK, and global communities.
Environmental Impact of AI – Call to Action
Salesforce are hot on the topic of sustainability (being one of their core values). You may remember the segment Benioff (Founder, Chair and CEO of Salesforce) dedicated in the Dreamforce ‘22 Main Keynote. As we’ll come to shortly, sustainability once again received attention in the Dreamforce ‘23 Main Keynote.
While intangible to the user, technology usage does involve a carbon footprint. We’ve covered the considerations that anyone building on, and architecting, Salesforce could take before on SF Ben – plus Salesforce launched the ‘green code’ initiative to encourage cleaner, more efficient code, and their data centers use 100% renewable energy.
More specifically in the context of today, AI models require additional processing power. The picture we’re about to paint may sound contradictory, but bear this keyword in mind as you read: intentional.
With generative AI technology evolving incredibly fast, the advice is to be open about exploring different LLMs. Salesforce recognizes the importance of an agnostic approach in their open gateway that gives organizations the choice to ‘bring their own model’. As it stands, which provider will win out the AI war is unclear – and the winners could be different from one use case to the next.
However, models for small tasks exist, and models for larger tasks exist. Being conscious, you wouldn’t choose a large model for a relatively small generative task (that would soak up unnecessary resources, increasing your carbon footprint). With this trial period, the objective will be to find the right model, for the right task, at the right price point. So, while you should experiment, be intentional about your AI model/s of choice.
Other call to actions for organizations is to choose data centers that source clean energy, use efficient hardware, and to customize Salesforce with efficient code.
Most Sustainable Dreamforce Ever
To practice what they preach, Salesforce have made a conscious effort to make Dreamforce, a large-scale event, as environmentally friendly as possible, claiming it to be the most sustainable Dreamforce ever. Check out the impressive statistics Salesforce published in the slide below.
Summary
The tide is turning on the topic of net zero – once a “nice to have”, now becoming a “need to have”.
You could say it’s a ‘carrot and stick’ scenario. While the ‘carrot’ of feel-good corporate responsibility has taken off, current and upcoming regulation is the ‘stick’ that will make even the most reluctant organizations pay attention to their responsibilities in the journey to net zero.
In a rapidly evolving regulatory landscape, efficiency in getting to compliance is healthy for both the bottom-line, and the planet.
It doesn’t stop at tracking and compliance reporting. There are other ways that organizations should be conscious of their environmental impact – and one the most elusive factors could be cloud technology, that while intangible, requires a lot of processing somewhere in the world. With the advent of AI, there are more considerations to take when selecting your AI provider of choice – which could be different according to the task that needs to be performed.