Notes from the operating room of AI

Oct 7, 2026·12 min

AI Tools Statistics 2026: 10,066 Products Measured

We swept 10,066 live AI tools in October 2026: 54% serve llms.txt, 4% have a Wikidata item, 535 block AI training crawlers and 75 block AI answers.

AI Visibility Tools: How to Choose Among 176 Trackers

176 AI visibility trackers, median entry price $49 a month. Ten established ones compared on engines, answers per month, citations and price, read 2026-10-07.

Generative Engine Optimization Tools, Sorted by the Job They Do

Ten GEO tools that change what AI engines can fetch, read and quote, sorted by job: crawler access, structure, content and audit. Prices read 2026-10-07.

Oct 7, 2026·14 min

Make vs Zapier: What the Same Workflow Costs on Each

Make bills credits per module run, Zapier bills tasks per successful action. One branching workflow priced on both at three volumes, from pages read 2026-10-07.

n8n Alternatives: Match the Tool to the Reason You Are Leaving

Why teams leave n8n (licence, self-hosting, per-run pricing, the learning curve) and which alternative fixes each reason, with prices read on 2026-10-07.

Oct 7, 2026·12 min

n8n vs Make: How Each One Counts a Run, and What That Costs

n8n bills a whole workflow run as one execution; Make bills every module action as a credit. Two workflows priced from both pricing pages, read 2026-10-07.

Oct 7, 2026·15 min

n8n vs Zapier: Which Costs Less, and Who Has to Run It

n8n bills per workflow run, Zapier per action. The same 6-step workflow priced at three volumes from both pricing pages, plus what self-hosting really costs.

Zapier Alternatives: Pick by the Reason You Are Leaving

Why teams leave Zapier, which alternative fixes each reason, and one five-step workflow priced on Zapier, Make, n8n and Activepieces, read on 2026-10-07.

Oct 1, 2026·3 min

Track: The Baseline Is the Only Thing That Makes After Mean Anything

Without a before number signed by the operator, the after number is a claim. Track produces the one measurement the whole engagement is judged against.

Sep 17, 2026·3 min

Hold: Why the Process Gets Frozen Before It Gets Measured

Measuring a process that is still being changed produces a baseline nobody can defend. Hold stops the drift for long enough to get a real number.

Sep 16, 2026·7 min

Vibe Coding Cleanup: What Breaks After the Demo

An app built in Lovable or Claude Code works on the day it is shown. What fails later is everything around the code, and a cleanup is the wrong unit to buy.

Sep 10, 2026·3 min

A Small Automation Firm or a Large Consultancy

The difference is not size or price. It is who carries the scoping decision, who writes the code, and what you hold afterwards.

Sep 8, 2026·3 min

Choosing Between Two Automation Vendors Who Both Said Yes

Both quotes look reasonable and neither is comparable. Five questions that separate them, none of which is about price.

Sep 3, 2026·3 min

What the Break Map Shows, and What It Decides

A process map is not the deliverable. The cost-of-chaos number and the priority matrix are, and they are what let a diagnostic end in no.

Sep 1, 2026·3 min

The Automation Brief an Operator Should Write First

Seven things written down before the first vendor call. They make three proposals comparable and cost an afternoon to produce.

Aug 28, 2026·4 min

When the Integration Changes Underneath You

A workflow that ran for months rarely breaks loudly. Something it depends on changes shape, and it keeps running on wrong data.

Aug 24, 2026·5 min

How to Read an Automation Quote Before You Sign It

The shape of a quote tells you more than the total. What is missing from the page is usually what you will be arguing about in month three.

Aug 23, 2026·5 min

AI Visibility for Operators, Measured on Our Own Site

Our site scores 90 out of 100 on AI readiness and 25 on AI visibility. That gap is the whole subject, and it is not fixed by more technical work.

Aug 22, 2026·5 min

Why Your Last Chatbot Failed, and It Was Not the Model

Most bad chatbots were measured on deflection rate, which pays them to avoid handing over. The metric caused the experience your customers hated.

Aug 21, 2026·4 min

Your Developer Can Build It. That Is Not the Question

An internal build competes with your roadmap rather than with a deadline, and that difference decides more of these projects than any technical comparison.

Aug 20, 2026·5 min

Why Most Process Maps Are Useless, and What Fixes Them

A process map drawn from interviews describes the process as designed. The money is in the difference between that and how the work is actually done.

Case Study

Aug 19, 2026·5 min

The Decision That Shaped a Three-Week Rental Build

Three plausible options were on the table and two of them would have taken months. What made the difference was deciding where the truth lived.

Aug 18, 2026·5 min

Six Hours to Eight Minutes: Inquiries at a Brokerage

A property inquiry is not one lead. It is a person, a listing and a routing rule, and getting that triple wrong is what makes brokerages slow.

Aug 17, 2026·4 min

What to Automate First, and Why It Is Not the Worst Job

The instinct is to start with the process everyone hates. Blast radius is the better criterion, and it usually points somewhere else entirely.

Aug 16, 2026·5 min

Five Signs Your Company Should Not Automate Yet

Size is the wrong question. Five conditions decide whether automation pays now, and failing any one of them is usually a reason to wait a quarter.

Aug 15, 2026·4 min

What the Amazon vs Perplexity Ruling Changed

The Ninth Circuit vacated Amazon's injunction against Perplexity's Comet browser in August 2026. It did not say agents are welcome.

Aug 14, 2026·4 min

Where four weeks comes from, and when you will not get it

Why every vendor quotes two to four weeks, what that number actually rests on, and the signs that your own project is not going to fit inside it.

Aug 13, 2026·4 min

Agency or Freelancer for an AI Automation Build

A freelancer is often the better choice for a first automation. The decision turns on what happens in month four, not on the day rate you compare first.

Case Study

Aug 12, 2026·5 min

The Week Before: What We Measured at a Rental Firm

Before touching anything at an equipment rental firm we spent a week measuring. Two of the four numbers we expected to matter turned out not to.

Aug 11, 2026·7 min

Where the Four Hours Go in a Rental Booking

Booking to dispatch took four hours at one rental firm and now takes three minutes. The four hours were never one task, and most of the fix was not AI.

Aug 10, 2026·5 min

The AI Checkout Got Cancelled. The AI Customer Did Not.

OpenAI switched off in-chat checkout in March 2026 after five months. The protocol survived, the buying behavior survived, and the work moved back to you.

Aug 3, 2026·5 min

What Your AI Is Not Allowed to Decide

Safe AI in a real deployment is four operating rules about where the human sits. Here they are, with what each one costs and where the line gets drawn.

Jul 27, 2026·6 min

Four Questions That Break Most Automation ROI Numbers

Any vendor can produce a payback number. Four questions decide whether it survives contact with your actual operation - including when the vendor is us.

Jul 20, 2026·5 min

We Built the Same Product Twice. Only 6% of It Carried Over.

Our second product reused 6% of the first one's business logic and still shipped far faster. The savings came from somewhere most people never budget.

Jul 7, 2026·4 min

Shipping an MCP Server for Your SaaS: A 90-Day Plan

AI assistants went from a curiosity to 97 million monthly connections in 18 months. A plain 90-day plan to let Claude and ChatGPT safely use your SaaS.

Jun 22, 2026·10 min

Process Audit Before Automation: When Not to Build

An engagement can end at the diagnostic with the deposit refunded. The audit playbook, the four rejection patterns, and the ROI math we sign on day one.

May 25, 2026·19 min

INITE Protocol: 6 Stages Applied to a Real Q2-2026 Deployment

Break, Hold, Track, Cut, Cast, Form - the six stages walked through a Q2-2026 B2B SME deployment, and the artifact each stage owes the next one.

Architecture

May 18, 2026·4 min

One core, eighteen entries in the ledger and one in production

What our industry products share, what has to be written from scratch every time, and why we publish a ledger of statuses instead of a count of products.

May 11, 2026·6 min

Browser-Agent-Ready SaaS: Making Your App Usable by Operator and Claude

Operator, ChatGPT Agent, Claude Computer Use can log in and check out on your site - if you let them. 2026 audit for selectors, forms, ARIA, captchas.

May 10, 2026·7 min

AI Crawler Allowlist 2026: Which Bots to Let In, Block, or Ignore

GPTBot, ClaudeBot, PerplexityBot, Google-Extended, Bytespider - 28+ AI crawlers want your content. A robots.txt + Cloudflare allowlist that keeps citations.

May 2, 2026·8 min

Real History of AI Part 5: From the Transformer to ChatGPT (2017-2022)

ChatGPT is not the arrival of AI. It is UX on top of 5 years of scaling: BERT, GPT-1, GPT-2, GPT-3, InstructGPT - and a 2019 GPT-2 commercial case study.

May 1, 2026·8 min

The Real History of AI, Part 4: The Deep-Learning Big Bang (2012–2017)

On 30 Sep 2012 deep learning stopped being academic. AlexNet won ImageNet by a record margin. Five years - word2vec, AlphaGo, GANs, the Transformer.

Apr 30, 2026·7 min

Real History of AI Part 3: Recommenders, Vision, Quiet Revolution

By 2010, AI ran inside every service you used. Netflix predicted taste, Last.fm built playlists, Facebook tagged friends - nobody called any of it AI.

Apr 29, 2026·8 min

The Real History of AI, Part 2: Backprop, SVM, and the Second Winter (1980–2000)

In 1986 neural networks got a working algorithm and most of the industry missed it. OCR read your mail, SVMs won benchmarks - all of it 'hidden AI'.

Apr 28, 2026·7 min

Real History of AI Part 1: Perceptrons, Symbolic AI, First Winter

AI did not arrive with ChatGPT. The first mathematical neuron was described in 1943, a working perceptron ran in 1958, and by 1969 neural nets were buried.

Apr 28, 2026·12 min

What Is llms.txt? The Spec, Real Files, and Whether It Works

llms.txt is a Markdown index for AI agents, proposed by Jeremy Howard in 2024. The format, live examples checked today, and what the evidence says it does.

Apr 26, 2026·13 min

Claude Skills vs MCP: Which to Use, and Why a SaaS Ships Both

Skills teach an agent how to do a job; MCP gives it access to a system. Where each runs, what each costs in context (measured), and what a SaaS should ship.

Apr 25, 2026·5 min

Answer Engine Optimization in 2026: An Evidence-Based Guide

A practical, sourced framework for making content discoverable, quotable, and trustworthy in AI-assisted search without invented ranking factors.

Oct 10, 2025·5 min

AI Integration in Business: The 2026 Field Guide That Skips the Hype

AI integration projects fail because the brief is wrong, not because the tech is hard. A 2026 playbook with the framing, numbers, and 73% failure-mode pattern.

Oct 9, 2025·12 min

Business Process Automation Examples, Department by Department

Twelve business process automation examples by department, each with its trigger, steps, systems, the decision a person keeps, and the number to measure.

Oct 8, 2025·6 min

Digital Transformation 2026: Why 70% Fail and What the Other 30% Do Differently

Most digital transformation programs are PowerPoint exercises that burn budget. The 30% that work share five concrete patterns. A practitioner field guide.

Machine Learning

Oct 7, 2025·7 min

Implementing ML Models in Production: The 2026 Reality of MLOps

Most ML models never reach production. The ones that do fail because of monitoring, not training. A field guide to MLOps that actually ships.

Security

Oct 6, 2025·15 min

LLM Security in 2026: The OWASP Top 10 Mapped to Controls and Tests

The OWASP Top 10 for LLM Applications 2026, each risk mapped to what it looks like, the control that limits it, and a test you can run before release.

Analytics

Oct 5, 2025·6 min

AI Data Analytics: From Dashboards to Decisions in 2026

Most analytics dashboards never change a decision. AI-augmented analytics works when it surfaces the question, not just the chart. A practical playbook.

Oct 4, 2025·6 min

AI Chatbots in Customer Service: What Actually Moves CSAT in 2026

Most chatbot deployments lower CSAT. The 30% that work share three patterns: tight scope, sub-30-second escalation, first-turn AI disclosure. Field guide.

Analytics

Oct 3, 2025·6 min

Predictive Analytics That Actually Predicts: A 2026 Practitioner Guide

Most business forecasts are off by 30-50% and nobody acts on them. Predictive analytics works when the model owns one decision, not a dashboard. Field guide.

Oct 2, 2025·7 min

Responsible AI in 2026: The Compliance Reality Behind the Ethics Talk

AI ethics in 2026 is mostly compliance. EU AI Act, NIST AI RMF, ISO 42001 are now enforceable - the gap between principles and audit evidence is the work.

Oct 1, 2025·6 min

RPA in 2026: When Bots Beat AI Agents, and When They Lose

RPA is not dead. For deterministic workflows it beats AI agents on cost (20x) and reliability (99.5% vs 92%). 2026 guide for picking the right tool.

Machine Learning

Sep 30, 2025·6 min

NLP in Business 2026: Six Use Cases That Pay Back in 90 Days

NLP is no longer sentiment analysis on tweets. Six 2026 use cases - contract review, support triage, sales call extraction - shipping in 8-12 weeks.

Sep 29, 2025·7 min

Measuring AI ROI in 2026: The Three Numbers Boards Actually Want

Most AI ROI calculations are vanity metrics. 3 numbers - hours saved per week, cycle time delta, error rate change - tell the truth. Field guide.