Here’s your obligatory opening to every AI write-up these days: as you read this, the world of tech is changing. We stand on the cusp of a new era powered by generative AI.

Large Language Models (LLMs), such as OpenAI's GPT-4, have made significant strides in understanding and producing human-like text. Trained on a large corpus of textual data, these models can generate coherent, contextually relevant sentences, paragraphs, and even entire documents.

LLMs can also write structured code from plain language prompts, enabling people without deep technical expertise to tackle technical tasks.

But this is hardly news as all these topics have been repeatedly discussed in the past few months. I’d like, however, to take a look at something I haven't seen anyone talking about yet: how LLMs might reshape the world of SaaS applications—and what role Salto could play in all that.

‍

The power of NaCl

Salto is helping businesses manage their cloud application configurations with our system that is based on a declarative language called NaCl ('Not Another Configuration Language'). Two aspects of NaCl are relevant to this conversation.

First, its unified nature: NaCl has the same syntax across all business applications.

Secondly, NaCl is declarative. Without going deep into the concept of declarative languages, let’s say that NaCl describes the desired end state of the application without specifying how to achieve it. Salto’s SaaS platform and open source project understand how to bring the business application to that desired state described by NaCl.

‍
‍

‍

If we combine NaCl with the text manipulation capabilities of LLMs, a future where AI can assist in analyzing, understanding, and generating configurations is just a hop away. These configurations can then be deployed by Salto across various apps we support, like Salesforce, NetSuite, Jira, Zendesk, Okta, etc.

For many companies, these capabilities are already useful today but could become absolutely essential as business systems transition to natural language interfaces.

‍

Natural language interfaces

In the past decade, cloud applications have evolved towards user-friendly, no-code interfaces. We built Salto to address the "configuration sprawl" that resulted from that new approach.

The next shift we will witness is natural language interfaces created by LLMs. Think of software like Siri, Alexa, or Google Assistant that we can interact with using plain language.

The complexity and volume of configurations underlying natural language interfaces will grow exponentially. This will present a daunting challenge for businesses, increasing the need for transparency and traceability in their systems.

Since conversational interfaces are text-centric, Salto's ability to translate configurations into a structured, human-readable format can bring order to potential chaos.

Democratizing the configuration process

Imagine a world where you can simply describe a desired outcome and AI will translate your requirements into NaCl, which Salto then can deploy directly to your business application.

For example, a marketing manager creates a complex customer journey in Salesforce with just a plain language prompt. AI translates their prompt into a NaCl file, providing a versioned and manageable record of that customization. Using Salto, the manager can also push it to a staging or UAT environment to be reviewed and tested by admins, ensuring enterprise-grade quality across all changes.

This would not only streamline but also democratize the configuration process, making it universally accessible to people without deep technical expertise.

To bring this vision to life here at Salto, we have more ground to cover and a few new concepts to prove. We’ve already started: the first of our many AI features just rolled out in a closed beta. It’s called Explain and you can read about it here.