OpenAI - ChatGPT - Copilot - what is your best choice?
The year 2024 will be dominated by artificial intelligence. There are now announcements from ChatGPT maker OpenAI, Microsoft with Copilots and Google with Gemini. Whether they are called chatbots or Copilots, the purpose of this smart software is to provide support in our work or search for information. So how do you make these AT tools work for your organization?
Copilots and AI assistants
This is the first part in a series of blogs in which we take a global look at applications of Copilots and AI assistants. In follow-up articles, we will take a closer look at all the Copilots that are and will be available.
AI sub-areas
Roughly speaking, we can use AI for 3 sub-areas:
- To support employees (internally)
- For the purpose of customer relations (external)
- Private assistance
In this blog, we mainly look at developments for the first 2 sub-areas and security aspects that many organizations struggle with. By private assistance, we mean navigation and applications such as Siri and home automation such as Amazon’s Alexia.
AI to support employees
Here we can already see various applications. For example, search engines such as Bing are getting smarter and smarter by not only showing a list of references after entering keywords, but also providing immediate answers to questions. Microsoft has framed this as Microsoft Copilot.
This Copilot appears in the browser and can provide assistance in searching for information. It can also be downloaded as a stand-alone app for Apple or Android phones. It is a form of AI that uses publicly available information on the Internet to come up with answers.
This year, Windows PC users will be able to call Copilot directly from their desktop, with the Copilot able to perform basic tasks on a PC. This is already available in certain countries, but there are still some EU privacy issues that need to be worked out before the launch later this spring. Functionalities such as directly translating documents that are currently being worked on or adjusting general settings such as changing your network settings based on your location will then become available.
Where in the beginning of the internet age it was mainly about Information at your fingertips, now AI is at your fingertips.
This Copilot appears in the browser and can provide assistance in searching for information. It can also be downloaded as a stand-alone app for Apple or Android phones. It is a form of AI that uses publicly available information on the Internet to come up with answers.
This year, Windows PC users will be able to call Copilot directly from their desktop, with the Copilot able to perform basic tasks on a PC. This is already available in certain countries, but there are still some EU privacy issues that need to be worked out before the launch later this spring. Functionalities such as directly translating documents that are currently being worked on or adjusting general settings such as changing your network settings based on your location will then become available.
Where in the beginning of the internet age it was mainly about Information at your fingertips, now AI is at your fingertips.
Security is key
An important aspect here is security. How do you ensure that information you share with an AI system is not used and spontaneously appears in other sessions with other colleagues or competitors?
If you are logged in with a company account, OpenAI (Team subscription) or Microsoft (Bing Chat Enterprise) ensures limited access to the data for colleagues only. In this way, information can be “securely” shared with an AI model to combine with information on the Internet to create new content or find solutions to complex issues. Data shared with the AI system will not be used to train the language model (GPT) itself.
Such systems become even more powerful when they take on a personal aspect and thus know and store the preferences of a user. In today’s Cloud Computing era, it’s easy to use your personal assistant anywhere (mobile, at home, on your wrist or in the car). Apple will also not let the interest in AI pass by and is a player that is close to the users with its devices. We are waiting for a Siri successor that can do more than just voice recognition. The latest phones and laptops are becoming more and more powerful and can process data independently (without an internet connection) through the use of an NPU (Neural Processing Units). This NPU relieves the CPU and ensures that things like targeted recognition or real-time translations of audio calls can take place without delay.
AI to enhance productivity
With the arrival of Microsoft 365 Copilot this week , you give the AI access to company information. Not only documents or spreadsheets, but also the minutes of meetings and calendar appointments are combined to get answers, generate new content or set out actions.
Companies that are already using Microsoft 365 Business (premium/standard) are ready to get started. In particular, the security controls in Business Premium ensure that only the data that a user has access to and is allowed to label information will be used. There is an interplay of Office applications (Outlook/Excel/Word), Internet search results (Bing), business information (MS Graph) and language models such as GPT. The outcome predicts a boost in productivity that costs an average investment of around 30 euros per user per month. It will not be a brainteaser for many organizations to calculate that with a small time saving of 1 hour per month, the ROI is already positive. Let alone when you can save 15-20% time per employee.
Artificial Intelligence in customer relations
f we look at the first group, we can already see various applications. For example, search engines such as Bing are getting smarter and smarter by not only showing a list of references after entering keywords, but also providing immediate answers to questions. Microsoft has mothed this Microsoft Copilot. This Copilot appears in the browser and can provide assistance in searching for information. It can also be downloaded as a stand-alone app for Apple or Android phones. It is a form of AI that uses publicly available information on the Internet to come up with answers. This year, Windows PC users will be able to call Copilot directly from their desktop, with the Copilot able to perform basic tasks on a PC. This is already available in certain countries, but there are still some EU privacy issues that need to be worked out before the launch later this spring. Functionalities such as directly translating documents that are currently being worked on or adjusting general settings such as changing your network settings based on your location will then become available. Where in the beginning of the internet age it was mainly about Information at your fingertips, now AI is at your fingertips.
AI to support your customers
AI can be used in various applications to enhance customer relationships. For instance, CRM systems such as Dynamics 365 and Salesforce offer Copilots, making it easier for users to generate reports. With Copilot, there is no need for a consultant to build reports in advance using applications like Power BI. Instead, an end user can directly instruct the Copilot to perform tasks such as creating a pivot table of all customers who have shown a turnover increase of more than 20% in the past 6 months compared to the previous year, grouped by product category, and exporting the result to an Excel file.
Data is the new goldmine - Semantic Index
Labeling information in a database with descriptive names for e.g. the columns in that database ensures that an AI language model understands the context and can use associated data to generate high-quality answers. The data sources that can be used are very diverse.
For example, the call recording of an application such as uCallManager can be used to convert audio clips and text. This text can be converted into any desired language by language recognition and serve as a source for a Copilot. In this way, data from SQL, SharePoint, OneDrive, Text files on a network Share, Outlook and CRM is combined. A large index of this data (Semantic Index) ensures that the AI can do its job better.
For example, the call recording of an application such as uCallManager can be used to convert audio clips and text. This text can be converted into any desired language by language recognition and serve as a source for a Copilot. In this way, data from SQL, SharePoint, OneDrive, Text files on a network Share, Outlook and CRM is combined. A large index of this data (Semantic Index) ensures that the AI can do its job better.
The power of AI lies in the data. This is and remains the property of an organization. All efforts to make this data available to the AI in a secure and accessible manner are the task for the coming period. It is not so much that a choice has to be made for products now, but a plan for the coming years to unlock data. It is important to make use of software that can unlock data with, for example, an API. The danger of data being stored in a vault that no one except a few system administrators can access must be avoided.
Data is the new goldmine for an organization. That promise was already made by Business Intelligence, but is now again within the reach of all users in a different form. It’s only in contrast to real gold that you want everyone in your organization to be able to benefit from data without the danger of unauthorized access, stealing it, or worse, destroying it.
Universal's AI pilots
A number of pilot projects that we have done really provide added value for users who, instead of searching for information, can now focus directly on assessing results generated by the Copilots.
The extranets with the aim of acting as a self-service portal for customers can get extra added value with AI. Starting small can be a sensible strategy, feeding online Chatbots with company information or even specific customer information when the identity of the user/visitor can be established. These chatbots are located on a website or are linked to a WhatsApp business number, for example. The combination of conversational AI and knowledge of products and/or services make these chatbots valuable digital colleagues that provide better support to customers. For example, they can help answer questions after hours when live agents are not available.
Don't wait and see but start now
The speed at which AI is currently developing may lead organizations to take a wait-and-see approach. New products follow each other in rapid succession. This is partly due to the further development of language models such as GPT, which will be available this year with version 4.5 or perhaps already with the 5th generation. The power of per-user cloud computing provides an opportunity to start with a small team and assess the value of AI before rolling it out to the rest of the organization. As with the introduction of software such as MS Teams, adoption sessions with users can provide a natural transition from a search-oriented system to assistance by AI Copilot that can be found in all applications.
Security is and remains a hot topic and a challenge for the security officer to stay in control and also to serve the organization with new possibilities. It is Universal’s task to supervise this process and to give advice based on experiences gained in previous projects. We will be happy to guide you in this interesting topic.
In a future blog, we will take a closer look at the various Copilots that are currently available.
Subscribe to our blog and stay tuned…
