In 2023, Evolution AI produced numerous resources to navigate the increasingly complex world of AI. In this post, we’ll briefly summarise our five most clicked articles.
Our most popular posts tend to be practical, focusing on ‘how-to’ skills that enable readers to utilise AI tools. Have you read them all?
In this article, we explain how GPT can be harnessed to perform simple data extraction from PDFs. Though GPT is constantly changing, this article remains a suitable starting point for using popular large language model (LLM) ChatGPT to extract data from documents.
The IDC predicts that, by 2025, 80% of all data will be unstructured (unstructured meaning that that the data doesn’t conform to any predetermined model). Every enterprise should include a strategy for managing unstructured information in their growth plans. This article examines how generative AI can be deployed to capture unstructured data from PDF files.
Following the success of our ChatGPT article, we penned a follow-up post detailing how to summarise and extract data from PDFs using OpenAI’s newest iteration, GPT-4. Though GPT-4 offers a marked advantage over its predecessor by being able to directly upload files to its interface, there are nuances to its ability to capture data. In this article, we examine how to automate extraction and summarisation from GPT-4, and its suitability for batch file processing.
A question that prospective clients often ask us is ‘Can your generative AI extract from complex tables?’. We decided to frame our response into a full article, exploring how cutting-edge AI models process and extract the required data points from complex tables.
.png)
Last but not least, our article ‘How to Extract Data from Financial Statements’ immediately began attracting clicks upon publication. Automating the extraction of data from financial statements, including balance sheets, income statements and statements of cash flows, has historically been a challenge in financial services. Now that there are generative AI solutions commercially available, we explore how to harness these technologies to produce structured data from financial statements in seconds.
Thoughts or comments about last year’s content? Email or Tweet us.
