The AInora Blog
Insights, guides, and real-world examples of AI replacing administrative overhead in service businesses.
Do Employees Waste 9.3 Hours a Week Searching for Information?
The figure of 9.3 hours a week, universally attributed to McKinsey, appears zero times in the 184-page McKinsey report. We downloaded it and searched it. The report says 8.8 hours, and that number is IDC’s. Every link in the citation chain, traced to the artefact.
EU Data Residency for Internal AI Assistants: 9 Vendors Compared (2026)
Not one of nine internal AI assistant vendors promises that AI inference happens inside your chosen EU region. Storage residency and inference residency are different products, and only one vendor documents a control over the second. Every cell quoted from the vendor.
Why an Internal AI Assistant Surfaces Documents People Were Never Meant to Find
An assistant indexed across your systems surfaces what an employee could technically already reach, which is not the same as what they were meant to see. The oversharing failure mode as Microsoft, Glean, Guru, Dust and Atlassian document it in their own admin pages.
RAG vs Agentic Internal AI Assistant: The Three Questions That Separate Them
A retrieval assistant reads. An agentic one can write into a system of record, which is why it has to ship an approval gate. The three-part test, each part quoted from vendor admin documentation, plus what the benchmarks measure.
Sana Alternatives After the Workday Acquisition (2026)
Workday acquired Sana for approximately $1.1 billion, closed 4 November 2025 and relaunched the product on 17 March 2026. Sana Enterprise is now sold with Workday HCM or Financial Management. What that means if you do not run Workday, and what the alternatives actually publish.
AI in Smaller European Languages: What Breaks, and Why
No open-weight model covers the 24 official EU languages well, and no major vendor publishes a Lithuanian quality figure for any text model. What the published evidence actually shows for Lithuanian and Estonian, what it costs in tokens, when to work in the local language and when to switch, and what to verify. Sources read 5 September 2026.
Why Do AI Rollouts Stall? What the Research Actually Says
AI programmes stall on people and process, not on the model. BCG puts 70 percent of the difficulty there, training dose moves regular usage from 18 to 89 percent, and the widely repeated claim that 95 percent of AI pilots fail comes from a preliminary report whose own pages contradict the headline. Every figure read at source on 5 September 2026.
Shadow AI: What Employees Actually Paste Into AI Tools
Shadow AI is the use of AI tools at work that the organisation has not sanctioned. IBM found one in five studied organisations breached through it, adding as much as USD 670,000 to the average breach cost, with 97 percent lacking AI access controls. What staff paste and why, why bans push usage further underground, and what an approved-tool list changes. Sources read at source on 5 September 2026.
Claude, ChatGPT, Copilot and Gemini: What a Team Learns Differently
A training-perspective comparison of the four assistants European teams actually get issued. The working model each one asks a person to hold, where each is strongest, what new users get wrong first, and what still transfers in prompting in 2026. Every capability claim is sourced to the vendor documentation it came from.
EU AI Act Article 4: What the AI Literacy Rule Actually Requires
Article 4 of the EU AI Act was rewritten on 27 July 2026. The duty is now to support the development of AI literacy, and the Article says expressly that no specific level must be guaranteed for any individual. Article 4 is not in the Article 99(4) fine schedule, though a separate route to a fine at that level exists for operators supervised by the AI Office. The exact old and new wording, who is bound, and what a programme contains. Sources verified 5 September 2026.
Does Your AI Vendor Train on Your Data? Tier-by-Tier, 2026
A primary-source table of what ChatGPT, Claude, Microsoft Copilot and Google Gemini do with your content at every account tier: training defaults, EU data residency, retention and admin controls. Every claim carries the vendor doc it came from, and the page is re-verified quarterly.
Why Did Our AI Phone Agent Book the Wrong Appointment?
A voice agent books the wrong thing when the rule that should have stopped it lived in a prompt instead of in the system.