Fulcrum Digital launches Ryze GenAI platform

Fulcrum Digital organized the company’s annual event Tandem, in Pune. There, Fulcrum delivered Ryze, its advanced Generative AI platform. Rajesh Sinha, Founder and Chairman, Fulcrum Digital, talked about the future of our planet. He elaborated on how it is going to be a confluence of human and AI.

Everyone is talking about generative language. That got simplified from spoken language to personal language. AI started from narrow intelligence, to general intelligence, on to super intelligence. You can now dress the fabric as per your industry to serve the purpose. Super intelligence is perhaps, 10 years ahead of us. 

For example, if you have an AI sales person, it can negotiate with a customer. Today, narrow sales intelligence is happening, and that will evolve in the years ahead. The AI bot will become super intelligent in the future. We will have AI as a copilot. Data is going into our brains, in the form of structured and unstructured data. We are now at a stage where we need to understand the sixth sense. AI is now coexisting with you! We have to accept AI, and use it to provide acceleration in lifestyle of all individuals.

Develop an AI mindset

Sinha added that today, we are thinking about an AI operating model. We need to develop an AI mindset. Things move very fast in the AI world. Having a clear mindset is important to develop your AI journey. Achieving levels 1-5 are crucial. 

For readers' information, the levels are: 

Level 1: Technology awareness

Level 2: Experimental activation of AI

Level 3: The operationalization of AI to innovate processes or products

Level 4: A systems approach and the creation of new business models

Level 5: The future of artificial intelligence.

He added that AI revolution is the mantra for the future -- digital, intelligence, experience, and purpose or DIEP. AI is all about data. It also has to be in form of relational database (RDBMS). Vector database gives you narrow adjacent relationship with another data. Today, open source LLMs are very close to ChatGPT. There are several open source models that are very close to ChatGPT. 

For those keen, a large language model (LLM) is a type of AI model designed to understand and generate human-like text. And, some alternatives to ChatGPT are Google Bard, GitHub Copilot, OpenAI Playground, Microsoft Copilot, Amazon CodeWhisperer, Vertex AI, etc.

Sinha said that we are now solving two problems. One, employee horizontal, and two, customer vertical. We need to add more business-related information to make it vertical AI. Improving employee productivity is horizontal AI. We can make more client-centric AI solutions as we grow. 

Today, LLMs have evolved into two camps. For those interested, transformer LLMs are capable of unsupervised training, although, a more precise explanation is that transformers perform self-learning.

Sinha stated that LLMs are now analyzing many parameters. We need GPUs to process them. Nvidia is a leading player today in GPUs. We further have the topic of ethical AI. Humans also develop governance models, especially for aircrafts. The future is unleashing the potential of both worlds.

Changing world

Sachin S. Panicker, Chief AI Officer, Fulcrum Digital, said the world had started changing post WW2. Perceptron was initially built. The first hardware implementation was Mark I Perceptron machine in 1957 at the Cornell Aeronautical Laboratory. Post 2000, Google BERT, XLNet, DistilBert, RoBERTa, etc., also came up. GPT1 happened in 2018 by OpenAI, followed by GPT3 in 2020.

In Nov. 2022, GPT3.5 was released and that woke up everyone from their slumber. In 2023, GPT4, came up. Now, 2023 has been the year of GenAI and LLMs. The AI market surged by $30 billion, fueled by investments and enterprise implementations. There has been widespread adoption. 

According to Statista, the AI market is projected to reach $305.90 billion in 2024. The market size is expected to show CAGR 2024-2030 of 15.83%, resulting in a market volume of $738.80 billion by 2030. Globally, the largest market size will be United States  worth $106.50 billion in 2024.

Panicker added that in 2024, emerging companies and tech (open source) will reign. We are seeing Perplexity AI, Lumiere, HuggingChat, Year of Robotics (VLA), extended reality and quantum computing, GENOME, NOIR, WRAP, etc. All this has levelled the open field for anyone. We will see more of open source LLMs in 2024.

Perplexity AI is one platform to watch out for. We have now conquered GenAI text. We have also conquered image. Next, we have to conquer video. We have a lot of work to catch up. It is an incredibly difficult feed to do right now. Today, we can also feed data to a robot, instructing it to pick up a bottle from a table.

We are now reducing the gap between human-machine interaction or HMI. You can have a knowledge world of the entire world in your hands. We are seeing new renaissance -- text, art, audio, video -- providing new definition of generation. You can optimize business operations with simple solutions. You can tackle issues without giving up on efficiencies. 

Robust yielding zestful enterpise

We are introducing RYZE, an enterprise GenAI accelerator platform, or Robust Yielding Zestful Enterpise. Ryze is designed to enhance business problem solving, offering solutions that are quicker and more effective.Enterprises are currently strugging to get all their structured data in place.

Ryze is a one-stop shop for all your enterprise AI needs. It is an intelligent enterprise assistant. You can upload your knowledge base, and query in natural language using text or speech. There will be questions, answering, summarization of entire documents and creation of new documents. You can now tap into every single byte of enterprise data. You can also synthesize data and download, and further do sentiment analysis.

We also have the Ryze reader or an intelligent document processor. You can now do data visualization, and predicive analysis using LLMs. We are still doing data visualization. You can also do video understanding and description. You can have the AI agents to automate business processes, and solve complex business problems. We will start seeing videos as part of AI by end of this year. 

We are now infusing AI into businesses to augment your capabilities. It can increase your productivity. We need to make it a non-biased tool for the world. It is a complex problem to address. We also have AI agents.

Here, Sachin S. Panicker, narrated an incident from last year. Apparently, he spoke about AI agents at a conference. There was muted response. Following that, Sam Altman made an announcement about AI agents a few weeks later. And, things started happening real fast! Panicker started receiving calls and messages from all over the world regarding their AI model.

Ryze has flexibility to choose between OpenAI and other open source LLMs, apart from Fulcrum's very own RyzeLLM/ It is secured, and has high degree of fine-tuning possible. Full customization is also possible. 

Ryze also holds context for an entire chat. Document can now be recalled in a separate window. Industry-specific agents can be integrated with any system. Intelligent bulk document is also possible with Ryze.

Ryze is also future-proofing finance. It has the personalized risk manager, AI customer relationship manager, adaptive fraud detection system, and ethical AI governance monitor. In insurance, we have automated claim executive, personalized policy maker, anti-fraud agent, AI policy recommendation, etc. In education, we have personalized learning assistant, research support executive, AI administrator, campus engagement monitor, etc. 

In retail, it has personalized retail shop assistant, inventory optimizing agent, an intelligent analyst, and in-store manager. We also have the global supply chain monitor, symbiotic food safety inspector, personalized nutrition planner, and culinary trend predictor, for the food industry.

RyzeLLM is at the forefront of FulcrumOne's innovative ecosystem. Born from top open-source LLM fusion, it showcases our pursuit of excellence. Our mission now is to push the GenAI's limits, and uplift creation.

Road ahead

The road ahead includes domain-specific LLMs, video understanding and description, cross model retrieval, autonomous agents, vision language action model, automation testing, audio generation, facial recognition, action prediction for visual clues, and additional modalities -- thermal, inertial, measurement units, touch, smell, etc.

Panicker said that we are transforming lives with our platforms and solutions. Ryze can raise the bar, and help realize potential.

We have an advanced RAG within Ryze. It is making the platform focus on document, and not drop relevant data. We can also introduce multi-modality. For those interested, RAG can be used extensively for banking, legal, and customer service, etc.

Sinha added that we need to query more for achieving an accuate result. Retrieval-augmented generation or RAG removes all the prompt engineering, to the intelligent engineering. We have many powerful tools within RAG. It is going to make LLM very powerful. The model uses up-to-date external data sources to provide responses. He concluded: Be more courteous, and you will get better results!

The Indian Government would do well to get in touch with Fulcrum Digital. 

Guaranteeing scalability

Ryze represents a quantum leap in AI technology, seamlessly integrating generative models, data analytics, user interaction, and customization. This delivers tailor-made content and data solutions that meet the specific demands of businesses and creative endeavors alike. Backed by its flagship platform, FulcrumOne digital accelerator, Ryze guarantees scalability to effortlessly adapt to the evolving needs of customers across various industry sectors.

Positioned as a powerful Software-as-a-Service (SaaS) solution, Ryze empowers businesses to revolutionize content generation, covering textual and visual mediums with unparalleled precision and efficiency. Its integration of state-of-the-art components including large language models, neuro readers, AI computing, and intelligent chatbots heralds a new era in AI-driven productivity and operational excellence. As a SaaS solution, Ryze's cost varies, depending on the number of user licenses, and the scope of implementation.

Key highlights of Ryze:

* Configurable and agnostic components for Language Model Systems (LLMS), accommodating a wide array of preferred models such as OpenAI, LLAMA2, SQL Coder, and GenZ.

* Seamless integration of customizable applications tailored to unique business requirements, enabling users to effortlessly fine-tune models without the need for coding expertise.

* Intelligent enterprise assistant capabilities, leveraging advanced agents to automate critical business processes, tackle complex challenges, and facilitate natural language queries in both textual and verbal formats.

Rajesh Sinha, founder and chairman of Fulcrum Digital, stated: "Ryze signifies a significant stride in our mission to harness AI's transformative potential for businesses worldwide. It leverages the entirety of enterprise data, encompassing text, image, audio, video, and 3D formats sourced from diverse channels including internal messaging platforms, emails, documents, presentations, and spreadsheets. Its key features span data synthesis, sentiment analysis, and the provision for downloading synthesized data for further analysis and utilization. 

"By marrying cutting-edge technology with user-centric design principles, the platform offers unparalleled flexibility, scalability, and efficiency in content generation and data processing. We are now well poised to redefine industry standards and empower businesses to navigate the ever-evolving landscape of digital transformation with confidence."

Sachin Panicker, Chief AI Officer, Fulcrum Digital, added: "With Ryze, we've engineered a transformative platform that embodies the future of AI-driven innovation. Through features like Conversational Interaction, Capable Comprehension, core Business Process Agents, Coding Assistance and Sentiment Analysis we empower businesses to unlock unprecedented insights and efficiencies. 

"Our FulcrumOne, a low-code development accelerator, provides tailored intelligence, enabling precision and context-aware search results, while our Neuro Reader, the powerhouse Intelligent document processor enhances content analysis and understanding better than other general available solutions, reducing manual entry errors. This serves as a predictive powerhouse, enabling users to anticipate trends and make informed decisions. We are optimistic about catalyzing innovation with Ryze, and disrupting in the digital age."

12 use cases of Ryze

* Digitized claims

* Conversational AI

* Fraud detection

* Policy rejection analysis

* Data extraction - loss & run reports

* Repopulate history & assign categories

* Data extraction - PO

* Medical records/prescription - data extraction

* Student transcripts

* Business rules

* Compliance checks for higher education

* Semantic search on catalog - furniture.