Sophie Gonzalez
By Sophie Aug 21, 2023

Tech Report – 4 Artificial Intelligences to Maximise your Productivity

In the professional world, opportunities and challenges are multiplying at breakneck speed. So it’s essential to master the tools we need to improve our efficiency at work. And that’s where artificial intelligence comes in! Sophie, our data scientist has put together a list of valuable AI tools that will boost your productivity at work and undoubtedly become your new allies.

Important note
The use of artificial intelligence (AI) requires vigilance and absolute compliance with company regulations. Safeguarding sensitive data is essential. Make sure to consult your superiors and adopt full transparency with regard to consent for the responsible use of AI.

1. Meeting assistants

At a time when remote working is becoming the norm, online meetings are ubiquitous. Note-taking is often unavoidable, but can be tedious when you’re chairing a meeting or speaking several times. As a result, crucial information can sometimes be forgotten.

Today, several AI tools enable automatic note-taking by offering a written transcription of ‘who said what’ during a virtual meeting. These assistants go even further, offering an abstract summary of the meeting, statistics on participants’ speaking times, screenshots of what was presented or a list of actions to be taken.

This is the promise of Microsoft Copilot for Teams, but until it is (finally) available, there are other multi-platform possibilities such as Airgram, Firefly and Otter.

Microsoft Copilot

Source: Microsoft

How does it work? It’s very simple: just register and add the assistant to the list of guests when you plan a meeting! However, it should be noted that the relevance of the results depends largely on the acuity of the transcription. For the record, Firefly has so far been better than the others!

2. Developer assistants

Another example of a relevant assistant, but this time aimed at developers: the Copilot feature on Github and Gitlab. These real-time assistants offer a number of advantages, such as:

  • Automatic completion of lines of code: this efficiently completes lines of code currently being written.
  • Integration with Visual Studio Code: integration with this popular code editor makes it easy to use. What’s more, it supports several programming languages.
  • Conception assistance: it can suggest functionalities or transform natural language comments into code, making it easier to translate requirements into programmable instructions.
Github Copilot 2

  • Documentation support: it automatically completes comments, speeding up the development process.

However, there are a few points of restraint to consider:

  • In Gitlab, the buffer is limited to a few lines, forcing you to click continuously to be able to see what comes next.
  • This can interrupt the workflow, as you may spend a fair amount of time pausing to read, accept or ignore suggestions, which can sometimes be frustrating and waste more time than it saves.

3. GPT plugins

One of the main complaints that keeps cropping up about GPT models is that the training data only goes up to 2021. However, for a number of applications, we need to interact with more recent or more private information. To compensate for this, there are plugins (both free and paid) that can be used to enhance GPT4.

For a simple, detailed explanation of how to activate plugins, see this link. Then check out the list of plugins available directly from the store. Here are a few examples of plugins that could be useful in your day-to-day work, whatever your profession is:

Use the power of GPT for your PDFs:

  • AskYourPdf is ideal for using the power of GPT4 on a PDF, especially if you have several “small” files that you want to compare, summarise or differentiate. For example, you could upload a candidate’s CV and the job description for the position your company is trying to fill, and then ask them how good a candidate they think they’d be.
  • If you have larger files, such as manuals, ChatPDF is an excellent alternative.

GPT on the World Wide Web:

  • VoxScripts allows you to browse data from the web beyond the GPT training date.

Specifically, you can use it to generate a list of YouTube videos from well-known channels on a given topic (it will search directly in the video transcripts for more accurate results). You can also ask it to provide you with a list of web pages on a specific topic, or to obtain financial data on a listed company. In all these cases, VoxScripts will go directly to the web for information, so there are no limits on the ‘freshness’ of its results.

  • The World News plugin specialises in press searches. For example, in a pre-sales context, you could ask it:

“I have a meeting with a potential client: Canadian Solar. Can you give me the latest news about this company? I’m particularly looking for information on their business strategy and development. List the newspapers you use as sources and the dates/periods of publication.”

World News

You could also ask: “Thank you. Could you give me a summary of the evolution of its stock market indicators over the last few weeks?” or finally (my favourite): “What is Canadian Solar’s commitment to research and development, and are there any specific AI-related initiatives?”

  • Finally, if, like me, you prefer to learn through a written medium rather than watching a video, why not use the Summerize Anything plugin to get a direct summary of the course, interview or demonstration.

GPT Code Interpreter: the analyst’s good plan

This is one of the latest features offered by OpenAI! A code interpreter has been integrated directly into its ChatGPT and, in addition to simply suggesting code, this feature allows code to be executed and tested directly in the interface.

This allows you to load raw data files (CSV, Excel, etc.) and request processing on this data (cleaning, aggregation, etc.) or even order a graphical display of the results (histogram, interactive maps, etc.). It’s ultra-easy and extremely practical for analysts. A big time-saver on the horizon!

4. (Bonus) Deploying a private GPT easily

If your employer is against the use of public artificial intelligence software because of confidentiality and data sovereignty concerns, you can always ask the AI or IT departments to deploy private LLMs (Large Language Modele).

For example, Azure OpenAI now allows you to take advantage of the Azure backbone in terms of encryption, confidentiality and physical location of data (and the service can be deployed in Canada).

As a result, you could see the power of ChatGPT arrive on company documents and databases directly via Teams (for example), while respecting internal authentication policies.

OpenAI

We hope these tools will free up your time and boost your productivity! And if you’d like to delve even deeper into the world of artificial intelligence, our experts are on hand to discuss the possibilities it can offer your company.

Recommended Articles
Published on April 4, 2023

Tech Report - Artificial Intelligence: A Closer Look at ChatGPT with Sophie! - Part 1

Sophie shares her initial thoughts on this revolutionary chatbot while also discussing its limitations. As a bonus, she also provides a wealth of information on artificial intelligence in general.

Read more
Published on June 27, 2023

Tech Report - Our developers' top 10 resources!

Tech, data, programming and AI are your favourite subjects? Would you like to keep up to date in these areas and stay on top of the latest trends? Our developers have put together a list of their favourite resources for you.

Read more
Share on