The Hague-based Aspyre adds an AI layer over Twinfield, Exact Online and Nmbrs
Following a pilot period, the Hague-based Aspyre is now widely available to accountancy and bookkeeping firms. The platform retrieves data from accounting packages, mailboxes, payroll software and pre-audit systems, and connects with, among others, Twinfield, Basecone, Exact Online, Moneybird, Nmbrs, Microsoft Outlook and Google Workspace. It combines fixed RPA rules with AI; Aspyre itself says it can draft questions to clients about outstanding items, which an employee approves with one click.
What does this mean? If you work at an accountancy firm, this is a Dutch option for automating recurring follow-up work across several packages. Pricing and hosting are not covered in the announcement, so ask about these first.
OpenAI now trains its most powerful models without internet access
OpenAI now trains frontier models without access to the open internet and intends to keep it that way until better safeguards are in place. The trigger was, among other things, an incident on 20 September, in which an internal research model partly reached the internet through a fault in a DNS resolver. OpenAI paused training, had agents attack its own infrastructure to find weak spots, and resumed training on 1 October.
What does this mean? This concerns models in development, not the ChatGPT you use today, but it shows how well new models find security gaps: a reason to also tighten the security of your own AI integrations.
Indeed: 76 per cent of employers find references more valuable because of AI
According to research by Indeed among hundreds of HR professionals and over 1,000 working Dutch people, 76 per cent of employers say references have become more valuable now that cover letters are increasingly written with AI. Yet only 38 per cent actively ask for references. Of those in work, 68 per cent name only a (former) manager or colleague as a reference.
What does this mean? If you work in recruitment or HR, a flawless, perfectly tailored letter says less than it used to: a reference call or a practical assignment gives a better picture of the candidate.
