Part IV (Summary & Navigating the Future) — Beyond the Buzz: Highlighting the Impact of AI in Modernizing Application

In the vast world of Artificial Intelligence, Large Language Models (LLMs) are changing the game, much like how personal computers (PCs) revolutionized the way we interacted with technology. Interest and use of LLMs have skyrocketed. Pre ChatGPT, only a few organizations, regardless of size or industry, had access to AI because it required specialized skills and had high entry barriers, much like how only certain people had access to computers before personal computers (PCs) became widespread.
Nevertheless, the landscape has rapidly evolved, as outlined in the introductory part of this series where we delved into the underlying reasons. Now, let me consolidate our conversation thus far and share my insights on the direction we’re poised to take.

Breaking the Mold: LLMs Beyond FAQs and Chatbots

The usual way organizations often use Generative AI specially LLM is by employing it for tasks like search (FAQ) or chatbot functions. This means they treat it as a distinct or separate initiative, focusing solely on these specific applications rather than exploring its potential for broader uses. However, we explored in this series how we’re breaking free from tradition, showing that LLMs can do much more. We’ve demonstrated how to create advanced conversational apps without needing a hefty budget or a complete overhaul of skills by leveraging the existing backend systems and tools in place. By blending LLMs with existing App devlelopment and modernization efforts, organizations can get the most out of these technologies while keeping costs in check.

What’s on the Horizon: A Glimpse into the Future Landscape

1. Conversational Apps Taking Center Stage:
   In the near future, conversational apps will become the primary focus, surpassing the prominence of the conventional apps we are accustomed to. This signifies a shift towards a “conversational apps first” approach, emphasizing the importance of natural language interaction. However, the evolution from conventional to conversational apps dominating the business application landscape won’t occur swiftly. Instead, we are poised to undergo an extended transition phase. During this period, various functionalities will migrate from conventional apps to conversational ones, resulting in a hybrid end state where the majority of functions will adopt a conversational format. Yet, it’s crucial to acknowledge that business functions that are considered non-compatible or high-risk will continue to be in their conventional form.
   Thinking about a scenario where every task exclusively uses conversation is improbable.
2. Security in a Hybrid World: Given the sensitivity and criticality of enterprise data driving this future state, the preferred deployment model is going to be hybrid with critical components remaining on-premise while others are distributed on cloud and edge depending on the use case.
3. One Platform for All Apps: From the Applicaton development perspective, I foresee a shift coming. Organizations will be gearing up to using a single platform to build and modernize both conventional and conversational apps. The idea is to make things faster to start, safer, and more efficient in delivering apps while maintaining consistency. However, this shift means we’ll be introduced to new technologies, especially for working with Artificial Intelligence (AI). Additionally, there’s a rise in the adoption of IDP (Internal Developer Portal), a special hub for developers and data experts (Engineers and Scientists) within a company. The future of app development seems to be evolving, and these changes are worth keeping an eye on!
4. Smooth Delivery Process: In this shift, both conventional and conversational apps will adhere to a standard delivery process. It’s about having a consistent method for bringing all types of applications to users, ensuring they meet the necessary standards and requirements. Whether it’s a conventional or a conversational app powered by AI, they’ll both go through the same secure and regulated delivery process.
5. Different Flavors of LLM models: we may be soon stepping into a phase with multiple AI models deployed together. We Started with large general purpose LLMs that undergo careful fine-tuning and prompt engineering (RAG) like we did in the POC, we’re now seeing a spotlight on new domain specific small models designed and built for specific industries. Looking ahead, I anticipate a future where even more specialized, compact models possibly tiny in size will come into play. This suggests that in the future, we might need several small models to power conversational apps.
6. Flexibility in the Cloud: I believe these applications will find their place on a hybrid cloud platform once again. The key is to have the flexibility to construct and oversee this hybrid app platform. This flexibility is crucial to adapt to the constantly shifting landscape of compliance, security, and data protection laws. Moreover, it ensures the platform can cater to diverse business requirements and needs effectively.

At a high level, the following is how I see the future enterprise application development and delivery landscape being implemented

Image: Hyrid Application Platform for Conventional and Conversational future

In summary, as part of this series we painted a picture of a dynamic future where LLMs play a key role in changing how organizations approach app development and delivery with a combination of conventional and conversational applications. We also discussed how moving towards conversational apps is not just a tech shift but a strategic move for efficiency and innovation and why it should in conjunction with existing App development and modernization initiatives. Please feel free to share your thoughts and have fun!!!