NAVIGATING THE SERVICES-AS-AI-SOFTWARE WAVE

At the sixth edition of #Techtonix, we deliberated on "Navigating the Services-as-AI Software Wave" with Manisha Raisinghani, Founder & CEO of SiftHub, and Karan Kirpalani, CPO of Neysa Networks, in conversation with Ashwin Kannan Pandian.

Key takeaways from the panel:

⚓ Embedding AI in Products:

Ashwin: Discussed the importance of integrating AI into products to drive business efficiency and deployment in production.

Karan: Emphasized that successful AI integration depends on specific use cases. AI should solve niche problems rather than be a one-size-fits-all solution.

Manisha: Highlighted the challenges of finding and using information within companies and how AI can streamline this.

🔍 Identifying AI Use Cases:

Karan: Shared that diverse use cases exist, such as in banking and insurance, where AI models like LLMs are being used to solve immediate problems.

Manisha: Advised that understanding how people currently solve problems and making incremental improvements can justify AI adoption.

💡 Developing Valuable AI Products:

Manisha: Emphasized the importance of solving a specific problem comprehensively for one team to drive adoption. Companies are willing to pay for improvements that significantly enhance productivity and management visibility.

Karan: Supported democratizing AI usage within organizations.

🧪 Testing and Experimenting with AI Models:

Manisha: Shared that iteration is key. Initial products may change significantly based on user feedback and testing results.

Karan: Provided examples of how clients are training LLMs on specific datasets, like legal documents, to create valuable AI tools.

⚙️ AI Deployment: On-Premise vs. Cloud:

Manisha: Emphasized the importance of building flexible, decoupled architectures that can adapt to rapid advancements in AI technology.

Karan: Discussed the high costs of hardware and the options between using cloud or on-premise solutions. Early-stage companies need to consider these costs when starting with AI tools.

🔄 Choosing AI Models: Open AI vs. Fine-Tuning:

Manisha: Highlighted the trade-offs between using Open AI models and fine-tuning proprietary models.

Karan: Mentioned that fine-tuning models like Lama 3 could offer cost-effective solutions depending on the requirements.

‍