Home / Intelligent Automation /
Ever since Alan Turing asked the question: “Can a machine think?”, people have been developing new technologies to improve the ways computers can impersonate a human in a written conversation. With steady evolutions happening in computing machinery and intelligence over the past 70 years, we’re now entering a new era in AI, where impersonation and imitation are just the tip of the iceberg in terms of what AI can now deliver.
Who’s who and what’s what in the AI space?
The term ‘Artificial Intelligence’ was first used in 1956 at the Dartmouth Summer Research Project. In these sessions, the team were working to the idea or theory that ‘every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it’.
AI is now the umbrella term for all systems that can simulate human-like cognitive functions.
Machine Learning (ML) is a subset of AI. It automates a computer’s ability to learn from data and improve the way it does things. Generative AI (GAI) is then a type of ML that specialises in generating content, such as text, images, video, and code in response to prompts.
Large Language Models (LLMs) are a type of GAI since they generate new combinations of text. They operate within vast neural networks that are trained on equally vast pools of data. It’s this ‘data pool’ that creates a virtual ‘font of learning’ allowing LLMs to be trained on large quantities of unlabelled text using self-supervised or semi-supervised learning. This is how they learn to understand and generate human-like text.
The Generative Pre-trained Transformer (GPT) is one series of LLMs, created by the company OpenAI. It’s been in development since 2018 and its most recent release is GPT-4 Turbo.
The leading organisation driving innovation in AI is OpenAI, an AI research and deployment company that’s headquartered in San Francisco. In 2019, OpenAI (which is non-profit) established OpenAI LP (which is a ‘capped profit’ subsidiary). This allows it to increase investments while delivering a ‘capped profit’ to its employees and investors if it succeeds at its mission, which is to ‘ensure that artificial general intelligence – AI systems that are generally smarter than humans – benefits all of humanity.’
Microsoft is the main investor in OpenAI LP1. In 2019 Microsoft placed a billion US dollars in the partnership to enable an underlying infrastructure for the training of GPT models, via its Microsoft AI Azure supercomputing technology. In 2022, this partnership was further strengthened with an additional $10 billion investment2.
OpenAI is delivering many of the products, platforms, and services that we’re all now using on a daily basis.
ChatGPT is a conversational AI platform, developed by OpenAI, that was released in November 2022. It belongs to a family of machine learning models called Transformers, which are designed to process and generate human-like text based on large datasets.
Because GPT models can perform various natural language processing tasks, such as text completion, translation, and summarisation, they can quickly and accurately provide informative and relevant responses to the questions (or prompts) people give it.
ChatGPT was previously trained using a dataset that featured information up to 2021. But the latest version (GPT-4 Turbo) now has knowledge of events up to April 2023, and has access to real-time information from the internet at large. An article in Computerworld4 explains that, thanks to an integration with Microsoft’s Bing search engine, the platform is now able to look up the latest available information on any topic. This means users have up to date information about recent events, developments, or changes in the world.
Also developed by OpenAI are the complementary technologies: DALL.E – an AI system that can create realistic images and art from a description in natural language; CODEX – an AI system that translates natural language to code; and WHISPER – a versatile speech recognition model that can transcribe, identify, and translate multiple languages.
Azure OpenAI Service and ChatGPT Enterprise
Microsoft has been offering its Azure customers access to the same LLM that ChatGPT uses, via its Azure OpenAI Service, since November 2021. This was intended to help businesses use large-scale generative AI models with the built-in security, reliability, compliance, and data privacy of the Azure environment.
Then in August 2023, OpenAI launched ChatGPT Enterprise. Delivering ‘enterprise-grade security and privacy, unlimited higher-speed GPT-4 access, longer context windows for processing longer inputs, advanced data analysis capabilities, customisation options, and much more5,’ ChatGPT Enterprise is an overlapping service to Azure OpenAI.
The difference is that ChatGPT Enterprise subscribers don’t need to subscribe to Azure, meaning this new service from OpenAI opens up the availability of its services to organisations outside the Microsoft Azure environment.
Leveraging OpenAI and ChatGPT in your organisation
With ChatGPT now available through the OpenAI Enterprise and Azure OpenAI Service, organisations can choose how they integrate AI into their business.
If you’re an existing Microsoft Azure customer, you already have REST API access to GPT, DALL.E, Codex and other LLMs with the security and enterprise benefits of Microsoft Azure.
The service is deployed within your existing Azure subscription with encryption of data at rest and data privacy governed by Microsoft's Responsible AI principles.
Using an LLM with your own data
If your data is already within the Microsoft ecosystem (i.e. you’re already on Azure) Microsoft Copilot is your most direct route to using an LLM with your own data. Microsoft 365 Copilot leverages the power of LLMs and combines this with user-specific data within the Microsoft Graph, such as calendar, emails, chats, documents, meetings, and more.
It can perform tasks such as drafting emails, creating content based on user prompts, and assisting with other productivity tasks within the Microsoft 365 or Windows environments.
For more information on Microsoft 365 Copilot, read our latest blog on this.
Using LLMs if your data is located elsewhere
Many Tecala clients have their data outside Azure. It could be stored in a database, in a file store, or even a custom or 3rd party application. If this is the case, you can still harness the power of an LLM with your data, and the way you’re going to do this is through plugins.
Plugins serve as the "eyes and ears" for LLMs because they give them access to external data sources, systems, and services. Quoting ChatGPT directly here, “they extend the model's capabilities to gather real-time information and interact with databases, APIs, or websites."
“These plugins can process and interpret data from various sources, allowing LLMs to provide up-to-date and contextually relevant information. Essentially, plugins serve as the sensory inputs, providing LLMs with the ability to perceive and respond to the ever-changing world, enhancing their usefulness in tasks like research, data retrieval, and dynamic content generation," ChatGPT explains.
Combining the power of Generative AI with your own business data
Realising that many clients need a quick and effective way to leverage the power of Generative AI and LLMs, Tecala has developed ‘Tecala GPT’.
Tecala GPT acts as an interface between your users and your data. From SharePoint and One Drive to Azure Blob Storage and even 3rd Party Apps, Tecala GPT covers a wide spectrum of data sources. This means there are no data silos, and every piece of information you need is within reach.
Going beyond simple keyword searches, it interprets and understands the context of questions based on its knowledge of the user's industry, and fetches the most relevant data from all available sources, to deliver the most relevant response. Because it also provides citations with the response, users can do their own further reading to improve their own understanding of the topic. But what Tecala GPT always delivers, off the bat, is the most accurate and nuanced response, in the most efficient way.
Whether you're on Microsoft Teams, Web Chat, or on your mobile, Tecala GPT enables you to delve into your organisation's data to get the information you need in relation to any situation. No need to juggle between apps or platforms, the answers you need are always just there, in one convenient interface.
4. Computerworld – Oct. 20, 2023
5. OpenAI Website – Aug. 28, 2023
Tecala is leading the way in intelligent automation in the mid-market space. Most of the leadership teams we’re talking to are truly committed to ensuring that technology does, in all honesty, empower people and enrich their daily lives.
We believe that at the intersection of automation, data, and AI, there are solutions for mid-market organisations that truly amplify human potential. Using intelligent automation, AI, and high-integrity data, we are helping our clients drive efficiency, performance, and profitability at every level of their operations.
Tecala's Data Landscape and Maturity Assessment (DLMA) is the first step in ensuring effective use of your data. In a three-step approach we assess the current state of your data, formulate the future state based on your goals, and provide a gap analysis between where you are now and where you need to go.
With the promise of unleashing creativity and productivity, we ensure the most effective route to a truly empowering and game-changing experience.
In this blog we put a spotlight on the importance of high quality, processed data to be the foundational resource for your automation and AI initiatives.
