Reaching the summit: The Future of AI in Enterprise SaaS

As we transition from generic cloud applications to industry-specific cloud solutions, we must move from dependency on generic foundational models to specialized task models or Small Language Models. Salesforce introduced the concepts of Sales Cloud and Service Cloud, and now they are advancing with Sales GPT, Service GPT, Commerce GPT, etc.

The next frontier will likely involve the emergence of domain-specific LLMs tailored to sectors like healthcare, telecom, and energy (HealthGPT, TelcoGPT, EnergyGPT). These models require deep industry expertise to effectively integrate domain-specific data, processes, and regulations. For instance, the telecom industry needs expertise in product modeling, TMF standards, and 3GPP specifications (3G, 4G, 5G) to incorporate accurate data and process specifications into the models for Telco use cases. However, the real benefit lies in the intersection of these domain-specific LLMs and Small Language Models specific to an organization or even a Non-Language model later in time. This will also bring the siloed functions of Sales, Service, Marketing, and Commerce together with voice command capabilities (ML/NLP/Voice command), making interfaces conversational. How the user or customer interacts with CRM systems today may change, with ‘voice’ taking precedence over ‘visual’ (UX/UI), further reducing the screen navigation time and eliminating data typing on pages.

While we await the development of these specialized models and AI-driven CRM, organizations can experiment with existing GPT capabilities and explore out-of-the-box solutions.

Successful adoption of industry-specific LLMs will demand significant investment and a deep understanding of the nuances within each sector. Training models with diverse datasets to identify patterns is becoming standard practice. However, the actual value lies in how industries adopt task-specific models and invest in them to handle complex enterprise use cases more effectively.