Ivan Nosov

What nosov.ai is

A working journal on rebuilding knowledge work with AI: the methods, the field notes, and a graded library of the evidence.

Paul Otlet, white-bearded, stands with five colleagues among stacks of papers and card files in his office.
Paul Otlet and his team, March 1937. Otlet set out to index the world’s published knowledge on cards. Collections Mundaneum, Mons. Public domain.

As of : 179 sources kept in the Library, 15 of them graded S, filed over 19 weeks; 62 pieces written.

How it works

The writing is mine and speaks for no employer. It draws on my own work, inside and outside my jobs, on conversations with people solving the same problems in other companies, and on the sources in the Library.

  1. Screen

    Every week an automated screening method I built goes through the stream of AI news and research: papers, institution and consultancy reports, company write-ups, talks and podcasts, in English, Russian, Chinese and other languages. Most of it is noise; the method exists to find what is not. Chinese sources are summarised from translation and marked as such.

  2. Grade

    What passes gets an English digest and a grade, S to C, and goes into the Library. The rest are left out. Every entry links to the original; nothing is rehosted.

  3. Summarise

    Every week a written summary goes up under Weekly: what happened, what it means for work, and what to watch, with the sources linked. A short audio version is added when one is made.

  4. Write

    My own pieces are under Writing; the longer ones under Long reads.

The screening, digests and grades are made by AI models working to a method I designed and keep tuning: which sources to watch, what counts as evidence, how to grade, what to drop. The method improves every week. The pieces under Writing are mine.

Chart 1

What the Library holds, by type and tier

Number of sources. A darker segment is a stronger grade.

Report Report, tier S: 5 Report, tier A: 22 Report, tier B: 23 Report, tier C: 8 58 Paper Paper, tier S: 6 Paper, tier A: 44 Paper, tier B: 3 53 Article Article, tier S: 2 Article, tier A: 8 Article, tier B: 24 Article, tier C: 6 40 Survey Survey, tier S: 2 Survey, tier A: 4 Survey, tier B: 4 10 Talk or podcast Talk or podcast, tier A: 1 Talk or podcast, tier B: 8 Talk or podcast, tier C: 1 10 Other Other, tier A: 1 Other, tier B: 2 Other, tier C: 5 8
  • S · Shapes the field
  • A · Strong evidence
  • B · Useful, limited
  • C · Weak signal
Library at build time, 179 sources. Practitioner posts count as articles; releases and code as other.

How sources are graded

Tier S sources
S Shapes the field · 15 Leading researchers, flagship reports from major institutions and consultancies, essays that changed the debate.
A Strong evidence · 80 Sound method or large documented data; worth reading in full.
B Useful, limited · 64 Narrow sample, companion material, or practitioner experience without measured outcomes.
C Weak signal · 20 Secondhand coverage, sales material, or a source whose content could not be reviewed.

A grade rests on what the screening could read: the full text, the abstract only, a transcript, or only the title and listing when the content was not available. Each source page says which, and gives the reason for its grade in one sentence.

Every week

The Library grows by the week. Each week's sources feed that week's summary, and the summary links back to them, so every claim can be checked against the original.

Latest: Week 41, 3-10 October 2026

Where the sources come from

179 sources from 105 organisations, in 3languages.

By type

  • Report58
  • Paper53
  • Article40
  • Survey10
  • Talk or podcast10
  • Other8

By language

  • English158
  • Russian13
  • Chinese8

Browse all sources

Ivan Nosov explaining with both hands held level, glasses, white shirt, FAIR lanyard.
Speaking at FAIR Barcelona, 1 October 2026. Photo: FAIR Barcelona

Ivan Nosov

I lead AI enablement and transformation: I design the system around the work, set the architecture, and help the people closest to the work become the ones who build. Before AI, fifteen years of transformation inside multinationals of tens of thousands of people: operating models, analytics, workforce planning, integrations after acquisitions. Trained as an automation engineer; I still build. Executive MBA, SDA Bocconi. Based in Milan.

Talks, panels and podcasts

  • ·Panel·The Conference Board, Future: People & AI, Brussels

    AI transformation: strategy, structure and workforce sustainability

    Panel with Adriano Bruno (The Conference Board) and Karlijn van den Berg (ABN AMRO), moderated by Peter Zemsky (INSEAD).

  • ·Workshop·The Conference Board, European CHRO Council, Brussels

    The six percent

    One-hour workshop on AI enablement for the council members.

  • ·Talk·FAIR Barcelona 2026 (Fair Cultures)

    Building the digital twin of the HR function

    Keynote on the Tech & Talent Stage, FAIR Barcelona 2026.

  • ·Podcast·The WorkOps Podcast

    After the AI pilot: building an HR digital twin

    What happens after an AI pilot works: starting with the tools already available, proving demand, and why HR is a useful testbed.

  • ·Podcast·Range Podcast

    Pay equity analysis when you're not a data scientist

    Conversation for compensation and total rewards leaders on running pay equity and pay transparency analysis without a data science team.

  • ·Talk·HR World Forum, Amsterdam

    From ratings to line of sight

    Talk to CHROs and senior HR leaders on performance management as a source of strategy intelligence.

In-house company sessions are private and not listed.

Contact

LinkedIn RSS feed

No sponsors, no affiliate links, no paid placements.