Solutions  |  January 21, 2024

AI tools for knowledge: how WASH AI fits in

In the past 6 months AI has re-invented knowledge acquisition, management and search. From the basic free ChatGPT to advanced, sector-specific AI applications, these tools are reshaping how we access and utilize information. This article places WASH AI, a solution currently being built by Baobab Tech, in context of four leading AI tools - Open AI’s free ChatGPT (3.5), Open AI’s custom GPTs (running GPT-4), Perplexity AI, and - highlighting how WASH AI fits into this competitive landscape.

ChatGPT (free and Custom GPTs)

Custom GPTs by ChatGPT, offer a comprehensive language model but come with constraints. The basic free GPT-3.5, despite its widespread use, is limited by data trained only until September 2021, lacking current and up-to-date information. This limitation is somewhat addressed in GPT-4, which provides more recent insights but still requires a subscription to GPT+, costing $20 per month. OpenAI released the “Custom GPTs” which allow anyone to build their own custom GPT and hook it up with a backend that can retrieve information in one or many database. While this tool can query external sources, its adaptability and customization options are limited to using OpenAI’s workflow, and may impact user experience, particularly for those unable to afford the subscription.

Perplexity AI: The new Google

Perplexity AI steps up as a formidable contender in the AI knowledge tool market. It enhances the traditional internet search process by not only responding to initial queries but also exploring variations and summarizing 20+ resources if one opts for the co-pilot version (not to be confused with Microsoft’s co-pilot). Its strength lies in its ability to stay up-to-date with information indexed by search engines. However, its reliance on these indexes can introduce bias, prioritizing results based on search engine rankings. Unlike WASH AI, which benefits from a curated and custom knowledge base tailored to the WASH) sector, Perplexity AI might overlook critical documents that are less prominent or not indexed online but are vital in the WASH context. That said, this is a brilliant tool for broader topic research and quick exploration. Focused on Published Research offers a specialized focus on published research papers, making it a valuable tool for academic and research-oriented inquiries. Its algorithm effectively navigates the realm of scientific literature to find relevant research papers. However, its exclusive focus on published research narrows its scope, particularly in sectors like WASH where grey literature – non-academic, often unpublished material – plays a crucial role. This singular focus on academic papers can limit the tool's applicability for practitioners seeking comprehensive insights, including practical, on-the-ground knowledge.

WASH AI: A Customized, Inclusive Solution for the WASH Sector

WASH AI emerges as a distinctive player in this field. Designed specifically for the WASH sector, it transcends the limitations of generalist AI tools by offering:

  1. Current and Contextualized Knowledge: Unlike basic models like GPT-3.5, WASH AI is continuously updated with the latest and curated sector-specific information, offering a more accurate and current knowledge base.
  2. Tailored and Inclusive Content: With a focus on both published and grey literature, WASH AI captures a broader spectrum of knowledge, crucial for comprehensive understanding in the WASH sector.
  3. Customized User Interaction: WASH AI's unique ability to infer inquiry intent and provide personalized responses sets it apart, offering a more user-centric experience compared to tools like ChatGPT’s custom GPT.
  4. Languages: WASH AI plans to be available in 20+ languages to ensure knowledge is broadly available irrespective of the source knowledge’s language.
  5. Integration with Diverse Sources: By incorporating varied data sources, WASH AI avoids the biases inherent in tools reliant on search engine indexes, ensuring a more balanced and representative information pool.
  6. Affordability and Accessibility: In contrast to subscription-based models, WASH AI aims to democratize access to information, making it a more viable option for a broader audience always including a generous free-tier for anyone to use.
  7. Multiple channels: WASH AI aims to be available via existing channels such as WhatsApp, Telegram, Signal and can be integrated in organization’s tools such as Teams and Slack.

In conclusion, while each AI tool has its unique strengths and applications, WASH AI particularly stands out in its commitment and specialization for the WASH sector, offering current, comprehensive, and user-tailored information access. This approach not only addresses the limitations found in other leading AI tools but also ensures that WASH AI remains a relevant, inclusive, and effective solution for knowledge management in the WASH landscape.

As the sector continues to evolve at a rapid pace, we are witnessing remarkable developments across the AI spectrum. The PPLX model is poised to support new types of devices like the R1. Google is expanding its Gemini AI into Bard and Google Search, creating a more cohesive and powerful search experience. Apple is developing efficient techniques likely to get efficient models on-device. The rise of open-source LLMs is reshaping how AI agents are utilized, with applications spreading across virtually all software we use. Microsoft’s Co-Pilot is on track to become fully integrated into widely-used applications like Word and Excel, further enhancing productivity.

At Baobab Tech, we are not just keeping pace with these rapid advancements but are actively engaging with them. Our focus extends beyond the current landscape of AI tools to the horizon of possibilities where LLMs are increasingly being harnessed to build practical, impactful applications. WASH AI is a testament to our commitment to staying at the forefront of AI development, ensuring that we leverage these advancements to continually enhance our offerings and better serve the needs of the WASH sector.

For more information on WASH AI and to experience firsthand, read this article or sign-up here and try it out.