This article is within the Book writing blog series.

Or: as usual, sharing what I am releasing and what I am going to publish soon.

As some of you know, this website started formally in 2001, specifically registered it on 2001-03-18, but probably, unless you followed it since then, you do not know that for almost a decade was just a "business card".

It was not my first domain registration: that distinction goes to prconsulting.com, that registered as soon as economically meaningful, on 1997-04-02, while I was in the process of moving from my birth country (Italy) to UK.

Used that PRConsulting since 1993, when first had to choose a name for my formal cultural and organizational change activities, so first defined the name, then started using it, as was starting a role for a customer, after delivering those activities first from early 1990 as a freelance, to then become between July 1990 and 1992, head of a business unit for the Italian branch of a French multinational- basically, with a charter to develop business in Italy on methodologies and associated training and change services.

Due to my past in political advocacy and solution sales to senior management in business, the "PR" represented both the "change" element (Panta Rei), and the "communication" element (PR as public relations): I had seen first-hand how each change initiative, and each solution sale initiative require both elements.

Until 2007, due to the activities I worked on (exclusively via word-of-mouth of former colleagues, managers, customers), I avoided having an online public presence with a website, and robertolofaro.com was mainly a kind of presentation of my CV and background.

Reason? As my focus is cultural and organizational change, also when discussing concepts, many could consider that was talking about them- also when I was not: cultural anthropology was not developed for a specific company, and Constitutional studies do not focus on just one country.

Really started using my own name on online media (and also to list companies within my CV) after a meeting in Brussels with a head hunter, as part of my relocation within a country that had visited in the past, but never worked in (also if already in the 1990s had been invited to join different local structures).

What was I told by that head hunter? As my CV was still the pro-forma "paranoid model" common for those working only through word-of-mouth and on confidential business activities, he told me that he had planned a 30mins interview, but stayed on three times as much as, despite the "been somewhere, done something" side of my CV, he had a feeling that there was a depth worth investigating- and those first 30mins provided even more interest in digging deeper.

He told me two things: that should not take 1.5 hours to extract (talking as he was trying to pull a tooth) that information, and that nobody could forbid me to outline my experience and where I worked.

It was quite impressive how he was able to actually identify potential roles based on my experience also from those little hints.

It was not a job offer- it was a structured career offer to integrate in multiple initiatives that could benefit from my cross-domain and cross-cultural experience in business interacting at the decision-making level: not a matter of boasting, just a "fact checking" and an exchange of mutual interest.

Which was something significantly different from what I had been repeatedly offered in Italy (except from some of my long-term customers who actually had no qualms about sharing value and acknowledging contribution publicly), where usually it is not what you do, but who presents you and which tribes you belong to that defines not just access, but also mere recognition (as I saw whenever asked for references for the work done in Italy since the late 1990s, again with just few exceptions- usually from those who had worked for non-Italian multinationals, and were used to the concept of recognition detached from tribal boundaries).

Anyway, the first role in Brussels was derailed from private and State interference from Italy to force a return to Italy- I ended up resigning and refusing to shorten the six month trial to become an employee, to protect the customer.

Reason? It is next to impossible to carry out a role close to ICT controlling and monitoring vendors, if you are questioned on anything you ask because somebody said that you were too cheap and knew too much about that industry despite having a banking background.

Hint: if you have a role in logistics but the agent sends a CV focused on banking, and you then get hired nonetheless... when your experience also in logistics and management consulting surfaces, in Brussels a cloak-and-dagger Circus started.

Nonetheless, Brussels was useful, as I had a chance to be in a multilingual and multi-cultural environment on a daily basis, including in my spare time.

So, could further expand my mental patterns not just within a proper business environment, but also having multiple formal and informal interviews and conversations within a really multinational environment, where many had a large corporate, government, and security background.

And to compare mental notes with people from different industries: hence, it was there that started really publishing online (albeit most of those articles are now offline- maybe in the future will share the material), as did not expect that for many years most of my background would be used in business activities, but wanted to keep "the cogs and wheels in the brain well-oiled".

Moreover: getting instantaneous and face-to-face (or, at least, indirectly at the pub) feed-back improved my writing style and ability to write for an unknown audience.

As, instead, since the 1980s had been used to write, in business or elsewhere, only for a known audience- but discussed the concept of audience in previous articles and specifically within a mini-book that published over a decade ago.

I will skip other Internet domains that registered, always focusing on communication about products, services, but also for non-profit.

In the early-to-late 1990s, as I was within that multiyear program management mission on change for a customer, obviously was monitoring all the technologies and approaches that could help in improving efficiency and efficacy in knowledge dissemination and communication.

The approach that had agreed to with CEO was to have an annual review of the objectives, and we had more or less a quarterly tuning of the mission for prioritization purposes, and e.g. after few months he asked me to "walk the talk" by acting as an advisor on strategic IT and non-IT projects, supporting managers and project managers (it was in banking).

Paraphrasing his words, he asked me, now that was confirmed that could preach and teach, if I wanted to prove that could help deliver.

Therefore, initially looked e.g. at the "re-inventing the government initiative" (early 1990s was when entered mailing lists from the then-DoD associated to that initiative, as wrote in previous publications), for the customer using Lotus Notes Workgroup as what decades later a colleague called a "wiki", joining eventually in Italy events on "network information retrieval", and starting webpages on various online communities to present my PRConsulting processes and concepts- years before even considered registering a domain.

Then, from mid-to-late 1990s, started supporting customers on integrating at least an Intranet (as was called back then- websites accessible only within the company or its supply chain) within projects- as a way to have easier access to shared knowledge.

My publication approach is nothing really special- it was basically derived from what learned at 17 within a European integration advocacy, an approach to be able to quickly release feed-back whenever there were news on whose cycle we could dovetail to inject our message.

Yes, despite what read in some marketing expert books even recently, "news hijacking" was something that existing long before we had websites, and long before that concept entered dictionaries assigned to a specific person- and, actually, long before I or those twice my age (or three times as such) I collaborated with in the early 1980s... were born.

Do not worry- will not start an history lesson on "spinning"- even Bernays in his "Propaganda" in the 1920s was integrating elements that were partially already there, but his main point was the systematic and structured presentation of those points: and that is what the XX century gave us in large volumes- "gurus" assembling in coherent and structured packages what was already there, but connected in an appealing and consistent way, to ease dissemination.

Or: it is not about innovation, it is about making it accessible to non-experts and to enable developing a service or product industry (or communities) around that.

You call it marketing, with its lingo and rituals- but, in reality, "selling ideas" has been done for thousands of years- our technology, if you focus, makes feasible to become an instant guru.

And, as was told in London over 20 years ago from colleagues, turning into a "guru" had an added bonus: whenever went around to deliver a speech on stage, an expert turned guru would actually get from the paying audience in attendance new fresh material to replace or update stale messages derived from the guru's personal (or related by direct business contact) experience.

In business, since started working in my official business role in 1986 (had previous ones while still in highschool- starting with a bit of ghostwriting, then did also some sales and software design and programming), was lucky enough to have a varied enough background (albeit no titles), and therefore being "parachuted" on activities way above "my pay grade", as some of my Anglo-American former colleagues and friends from a non-business background would say.

Why parachuted? Because I have no qualms about spending even my spare time to "tune" on the lingo and specific domain patterns- and, if you do that long enough, it becomes a second nature and allow to quickly "connect the dots" across multiple domains, and integrate for as much as needed within a new domain.

So, in late 1988 was sent to London to attend training on solution sales with our Anglo-American partner (where met most of those Anglo-Americans with a government and related background that had to work then with for decades), where I blended my prior political advocacy and limited experience, plus other interests on the non-technical and non-business side with more modern material.

Incidentally: and in London, as shared in the past, learned the hard way about cross-cultural teams and cross-cultural negotiations and project management activities and...

... mistakes easy to avoid if you remember that not everybody has your same cultural background- something that was to integrate with a bit more of theory and in-classroom practice in summer 1994, at the university in Sweden (intecultural communication and management).

Now, why all this summary of my non-"technical" and non-business but mainly "soft skills" background and development history, within an article with a title about numbers and AI?

Because this article is a small detour from the string of articles about Europe, Italy, Turin that hinted at as forthcoming within the previous article in this series.

Specifically, would like to share in theme 1 the rationale of my publication activities and the tools that discuss in theme 4 and that released/prepared over the last week-end.

Then, also if it is nothing really original, would like to share in theme 2 the conceptual side of my publishing process, so that I have a chance to explain all the transformations that an article goes through, from concept to publishing to communication and preparation of ancillary material to ease access to its content.

Most of those transformations were already in place since I started revising the design of this website in 2012, when was made to return to Italy from Brussels.

Others were gradually added as part of my continuous learning, to update on that 1980s learning in London (not just on sales but on the technology we were selling- decision support systems customer designed for senior management to introduce what we would now call "evidence-based decision-making", leveraging on the business expertise of customers and their own validated data).

Personally, as think (also in business) mainly through pictures (useful when you have to constantly switch the language you speak, sometimes few times a day with few languages, in the past sometime even during the same presentation or conversation), I am used also to "cluster" visually information before uttering a single word, and to do the opposite while listening to what others say.

As a new addition to the ease of access to material published on this website, in theme 3 will outline the clusters that, after experiments with the 300+ articles on change already online, identified and then tested with AIs to see if were qualifying enough to enable AIs to classify articles under a main cluster and additional clusters.

Within theme 4 instead will describe the actual tools that are part of this release, including both those that use in backoffice, and the new version 1.2 of the MorningNews, an experiment that run each morning with Claude since March, to test a concept that used in the past manually, but using AIs to do all the boring part (search, collate, digest, present)- in a couple of minutes instead of the hour or more that took me a decade ago (when posted those daily summaries on Twitter), moreover, delivering a structured "product", with some additional features from version 1.2 to ease not just reading "the now" but also "the developing and emerging trends".

This time, no "conclusions and next steps" section, as I think that the introduction above, along with what you will read within each section, will deliver more than enough material to share- and prefer to postpone further "next steps" information to when will have something more to share.

Incidentally, as a closing point: MorningNews is and will always be free, as I am sharing experiments and concepts that hope that some will replicate, and will probably become useful in the next future projects where, as currently there is no sustainable demand in Italy for my old services, will act as program, project, change, PMO manager- keeping the other parts of my activities, as did in Brussels since 2008, for publications, so that can reach a wider audience.

So, stay tuned for more material!

Few sections:
_ THEME 1: rationale
_ THEME 2: process
_ THEME 3: clusters
_ THEME 4: tools



THEME 1: rationale

As I wrote at the end of the latest article in this series (BookBlog20260406 A month after, few months ahead):

The next few articles will be about Europe, Italy, Turin- where, for now, continue to focus on working just on program/project/change management, as my old management consulting activities will be just for occasional missions or for publications .

So, this article will be a detour.

Why now? As the title says, the number of readings is getting again close to 7,000,000- but, considering that actually on January 23rd reduced from 7,185,996 to 5,763,175...

... actually the number is 7,185,996 - 5,763,175 + 6.988.401 (the status when began to draft this article yesterday) = 8,411,222, i.e. closer to 8,500,000.

So, decided to release earlier last week as a test a small experiment to further extend the ease of access to articles- by integrating directly AI within my publishing pipeline.

Whenever worked as project manager or helped to develop project managers and managers, I was used to prepare documents tailored to the specific opportunity.

My approach is simple: as I am used to work on multiple projects and initiatives at the same time and, from the late 1980s to mid-2000s, multiple customers across the week, even in different countries, I try to enforce a significant degree of transparency and knowledge dissemination, to ensure that we are all on the same page- and to be able to both delegate and actively listen, as could not be full time on a single location.

Which implies also that, if the budget does not allow, whenever I see what I used to call a knowledge gap or misalignment, I think that it is up to the coordinator to create a level playing field, maybe even involving team members into producing a kind of "initiave shared guidelines" (as often I am involved within existing initiatives, where those elements are missing or need to be restructured).

Yes, each member of the team has a specific "domain", but a degree of shared knowledge is needed- and this was partly inspired by my experience in political advocacy (we had to share some common ground, also if PR was formally prepared by those in senior roles) and then in the Army within an artillery specialist outfit (each one had a specific role, but we had to share some common elements and be able to temporarily replace other members).

Before any mission that you can see on my CV (and many more that did not bother to detail), routinely asked to be provided information before I started any official role, so that I could "hit the ground running"- or, at least, understand the perimeter and who would cover which domain.

Therefore, knowledge dissemination is a two-way street, but, in my view, it is up to the coordinator to be accountable for communication flows and their proper coordination: but discussed this point in previous articles and also within the #QuPlan mini-book and the associated 200+ pages fictional case study of a compliance program.

I wrote within the introduction (again) a short summary of my moving abroad in 1997 (de facto, also if formally was in 1998), and my full working return in Italy in 2012.

Since 2012, after a first mission where saw the potential of resuming eventually all my old activities (not just project/program/change/PMO management, but also management consulting), understood that in Italy there was limited chance of that- hence, decided to keep working on mini-projects in my spare time, projects where I would use (and keep up-to-date or doing the usual pruning, unlearning, relearning) my skills.

It was interesting how in 2012-2025 found many ways to reuse, repurpose, update, prune, renew lessons learned from formal experience from 1986 and informal from 1982.

And, of course, to add new material and "knowledge points".

Anyway, from mid-2010s also technology and access to technology changed dramatically: what I was doing e.g. in 2018 with R, in the 1990s would have required a significant budget in both software licenses and hardware.

From the early 2020s, this has been expanded with easy access to online free or almost free computational and storage resources.

As did in the 1990s for my customers (not just on that role in cultural and organizational change, but also to support other customers in Italy and abroad), I kept learning and carrying out experiments- but, again, the 2020s dramatically reduce the costs.

What you see on this website (and will keep seeing in the future- also in venues accessible even if I were to eventually decide to pull the plug from the website) is a constant reminder of that evolution, and in this article would like to briefly touch in the next few sections the "components" of what released last week.

A closing note for this section on the "rationale": look at everything that you see on this website as a "continuously evolving" set of elements.

And now, a bit about my publishing process.



THEME 2: process

Look at this image:

.

There are 4 main sections:
1. Creating
2. Processing
3. Management
4. Publishing.

I think that the first section, that on purpose detailed, is nothing unusual: you start with a concept, find material that could be useful, then start drafting, and, considering the structure and content of my past publications, connect-the-dots with previously shared material.

Then, often, this implies re-assessing the research, review the draft, and, once done with a final connecting-the-dots, a review.

Frankly, in many cases, I have a pile of concepts that develop in my mind, before even typing the first word on my computer, and keep developing them in parallel, sometimes even merging or splitting them.

And the research, drafting, connect-the-dots usually are relatively quick: if the concept is developed enough, sometimes a long article is done in a single session, with some tuning of the research and connect-the-dots.

The reviewing is more a formal stage- a kind of re-reading that, when I have time, is done maybe touching here and there again the material, and in other cases, or if I decide that the article has to go live faster, is quite compressed.

The other three sections are really where a blend software and brain.

The processing is focused on integrating the article within the database of articles: so, ensure that, once the section is confirmed, the article will have all the elements that articles in that section have.

The management used simply to be about inserting the article within a local version of my online website, to check if there are some visual rendering issues, and then to generate the material needed to enable using the various search functions.

Frankly, in many cases, again if there is a need to post it fast online, I skip the parts "generate the material", and focus just on verifying and posting it online.

And the other parts of that "management"? Can add them later, usually a couple of days later.

Then, the publishing is really about physically adding the article to the online database (and integrate also the other material, if already prepared), and create the posts to communicate online within my social media profiles that a new article is available.

The "publishing" includes also releasing on Kaggle the metadata for the articles including the latest one, and, if available, updating other search facilities on this website (but, again, can be added later).

Routinely review material to improve structure, content, etc.

I wrote above that I think visually- and this implies that every new article has to "blend" with the mental images representing those in the same section.

Therefore, it can happen that a multi-part article has one "episode" within a section, and another "episode" in a different section.

For reasons that will describe within the "tools" section of this article, actually started since few months ago to implement this discipline with a closer continuity, e.g. splitting in multiple "episodes" if a single article would need to cover what should be belong to different sections.

The reason? Because, gradually, made it easier to retrieve and access articles- also automatically.

And, actually, the "multi-part" section on this website, and trying to enforce "boundaries" were suggestion from few AIs when I asked to review my articles, to both ease access and follow threads.

No, I did not ask AIs to modify the code: to suggest concept improvements- as I developed long ago the "infrastructure" of my website, it is faster for myself to intervene on code if and where needed with surgical precision.

As wrote in a recent mini-book, I use AIs as team members that provide expertise that I lack or that they are more efficient to deliver, also if I start "AI-first" in most cases, i.e. develop a concept, discuss it with AIs, then use that discussion to influence me, but drafting and writing alone, using again at best AIs to review what I produced (asking to criticize my material), and routinely just to generate the material needed for the various search approaches embedded in my website, i.e. removing most of the manual work that I would have needed to do but added no value (i.e. content).

Anyway, when I had already few months of publishing of the metadata on Kaggle, and content on GitHub, both extracted from this website, it was time to ask again AIs some suggestions.

And, trying with few of them, saw that there were two opportunities:
_ identify themes, themes that could evolve
_ identify clusters, clusters that could instead be a point of reference.

After few rounds with few AIs, eventually settled on a first experiment with six clusters, that present in the next section.



THEME 3: clusters

This section for now will be just a catalog:

A Case Study

contains descriptions of experience-based cases

B Conceptual

discussion of one or more concept that will be further developed

C Framework

a more structured version of the "conceptual", generally to enable a reuse (changing what has to be changed, of course)

D How To

this shifts from framework to its implementation, but could also refer to specific applications

E Checklists

crossing the Ts and dotting the Is: checks and balances are not just in politics, also in business and technology might help

F Visual

this is a more generic cluster that for now is an experiment, trying to eventually transfer here the integration of the others within a visual model

Now, my concept of "clusters" derives from other domains, e.g. HR.

I know that some prefer to say "this person is abc", that person is "bca".

Personally, I think that everyone is a "blend".

There might be a leading characteristic, but other elements integrate and, as a blend, often transform that leading characteristic, or at least "tunnel" its delivery.

The same is my concept of cluster for articles:
_ each article has a main cluster
_ but can have a bit also of the other five.

Initially considered and tested the possibility of having the main cluster marked also as additional, but was generating confusion.

Hence, eventually, when started toying with AIs to classify my articles automatically, enforced a differentiation:
_ just one main cluster
_ one or more additional clusters that are not the main one.

This makes easier to search and further cluster, if you want to select and extract articles.

Now, fine: but decided also to make easier to combine the different perspectives, e.g. the cluster, the potential additional cluster(s), and the "section" an article belongs to, plus the possibility of being part of a "storytelling" that extends across multiple "episodes", with each episode within a different "section".

Seems complex, but if you think by pictures is not.

Now, how can deliver that in a easy way?

Time to shift to the tools.



THEME 4: tools

When I prepared to write this section, the focus was initially on the new tools added last week.

In reality, this section should be on tools that ease access to articles- developed and delivered across a decade, but with an acceleration since late 2024, courtesy of the help provided by AIs in carrying out time-consuming and repetitive tasks starting from natural language guidelines structured in a way that is consistent with how models work, not how humans work.

search by tag cloud

each article is converted into a list of tags, and this search functionality presents tags by frequency, allowing to also select multiple keywords, to retrieve a list of articles containing those tags

metadata

a dataset on Kaggle containing the "coordinates" for each article, and a short summary

content for dissemination

each article is converted into two files, a "shipping manifest" ("abstract") containing key information, and a "full article" ("content")

search by cluster

each article is associated to a main cluster and one or more additional clusters, allowing searching

statistics by cluster

a summary table containing the statistics about "patterns" by cluster, i.e. the valid combinations of main and additional clusters present within the database

Of course, you can also look at the "sections", e.g. book writing blog, and see all the articles within that section.

As you probably saw already, over the last few months the last section of each article contains a reference to uses of AI within that context.

And this case is no exception: it all started looking for a way to have articles automatically associated with clusters.

Then, added also automatic identification of other characteristics.

The key element? To validate the concept of doing all this using local AI models, and on any PC (i.e. not necessarily with a GPU).

The structure itself will evolve with time, based also upon which searches are more common.

There is another tool, that started publishing online in mid-March 2026, a personal selection of news items- you can read the history at the link.



Yes, there are many tools online to obtain news- but, in my case, wanted something different, focused on specific information needs.

The new version, 1.2, will make easier also for external models to read the GitHub content, e.g. by giving a form of continuity within content and a structure that can evolve across time.

in projects, as wrote above, was used to share material also to inform partners in my network of evolving and emerging trends.

in the XXI century, in many projects, I was used to read book and material and news specifically focused on the initiatives I was working on, to share trends, not just recycled material from prior readings or prior "thesaurization of activities" (e.g. I always do a "lessons learned"- both continuous, to adjust activities, and a final as part of the transition to another activity).

In any case, I take seriously the "handover": it is not just a matter of dumping some material from the past or having few brainstorming sessions or presentations- I prefer to actually prepare something similar to what gave my customers at the end of projects that I had personally sold and delivered (or coordinated the delivery).

in the future, I will see if actually could make what saw in the early 1980s in those political advocacy activities, and did all across the 1990s and 2000s with customers and partners around Europe.

Hence, the MorningNews experiment is a continuation of what did a decade ago: but, instead of having to meet me (physically or virtually), or receive emails, or having to prepare manualy lists, could easily tailor alerts and communication during the activities.

Both the search by tag cloud, search by cluster, and MorningNews are part of the same approach: easing access to information and reducing the work needed to produce, update, disseminate information.

As an additional element, the search by cluster and MorningNews are also a test to see how free AI tools can allow both online and offline access to information- including from mobile phones.

Now, I would like to close with a caveat- for me intuitive, but confirmed in a post today referencing a study:


(read here the Linkedin thread)

So, stay tuned for more updates.