A working paper with a DOI: The cliff of innovation

I have published a working paper, and this is a short note on what is in it and where to read it.

The title is The cliff of innovation: why integration, not construction, is the hard part of building a system that acts inside a company. Nine pages. It is deposited on Zenodo with a DOI, 10.5281/zenodo.22704763, which means the file will still resolve when the places that host it today do not. It is self-published and not peer reviewed, and it says so on the first page.

What it argues, in one paragraph. There is a point in building something new where the hard part changes kind. Below that point, the difficulty is building: writing the code, making it run. At that point, which I call the cliff of innovation, almost every part you need already exists, made by different people for different purposes, and the difficulty becomes integrating them well enough that the result behaves like one system rather than a collection of parts. For a system that acts inside a company, four things are missing in every company I have seen, and they are not parts you can buy: a written account of what is true, which rules the business already follows, which actions can be undone, and who may approve what.

What is in the paper that is not in the short version on my site: the five seams by which you can recognise a badly integrated system from the chair of the person using it; a worked example of the permission table, row by row; the scope the method is running on, as stated in our technical report; three predictions dated for 2028 so that they can be checked; and a section on the paper's own limits, including the conflict of interest of a co-founder writing about his company's method.

Where to read it: the PDF on my site, https://isaac-correa.github.io/the-cliff-of-innovation-isaac-correa.pdf, the record on Zenodo, https://doi.org/10.5281/zenodo.22704763, and the copy on Academia.edu, https://www.academia.edu/175376622/. The short version is at https://isaac-correa.github.io/cliff/.

If you read it and find one of the three predictions already wrong, I would rather hear it than not.

Comments

Post a Comment

Popular posts from this blog

Six pages I added this week, and what each one is for

Image
Me, in Madrid. This is the portrait I use everywhere, so it is the one on the photographs page too. I spent this week filling in the parts of my own site that were missing, and it turned out most of what was missing was not writing. It was the boring apparatus: where the pictures are, what the dates were, who to ask. Six pages went up. Here is what each one is for, in case any of it saves somebody a question. A timeline where every entry carries the place you can go and check it — the register entry, the DOI, the review count. Anyone can type a year into a CV, so I put the record that fixes each one next to it and left it at that. A press kit . Three bios at three lengths, ready to paste, with the facts and their sources underneath. No email to send, no form to fill in. If somebody is writing about the company at eleven at night, the answer should already be sitting on a page rather than waiting in my inbox until morning. The photographs , captioned with what each one actuall...
Read more

Who cited Isaac Correa, and what they were actually looking for

Image
Anyone can write about themselves. This is the other kind of evidence: three places where someone else, with no reason to do me a favour, used what we published as a source. They are not all the same thing, so I separate them. Isaac Correa in Madrid. What follows is cited by others, not by me. Cited by name: Mordor Intelligence The market research firm Mordor Intelligence cites me by name in its report on AI in patient care and management . The entry reads: Isaac Correa, “Digital Front Door in Healthcare: Strategy Guide 2025”, Hellomatik . It sits in the reference list, next to the other sources they used. This is the one that matters to me, because it carries my name rather than a company's. A market report is written to be sold to people making decisions with money. It does not cite you as a favour. Used as a source: two trade publications On 24 January 2026, DentistryIQ linked to our figure on how much of a cancelled slot automated waitlists recover. On 6 Mar...
Read more

Isaac Correa: who I am, and where every fact can be checked

This is the short, checkable version of who I am. Every claim here has a place you can verify it, and the long version lives on my own site. I am Isaac Correa, a Spanish entrepreneur based in Madrid. In the Spanish commercial register my full name is Isaac Correa Villa. I founded Hellomatik on 28 July 2025, at twenty-two. The company writes down how a client decides, its data, its rules, its actions and who may approve what, and then lets AI agents answer, watch and execute on top of that, inside limits the client approves in advance. My appointment as joint administrator was published on 4 August 2025 in the BORME, the Spanish commercial register, sheet M-861159, whose source is the Agencia Estatal BOE. Before that, in September 2024, I founded Kodalogic in Madrid with my brother Daniel Correa. It sells ready-made Google Data Studio dashboard templates for GA4, Search Console and Google Ads, with the interpretation written next to each chart. It came out of the work: between December ...
Read more