The faster Horse that never was

Henry Ford probably never said “If I had asked people what they wanted, they would have said faster horses.” The quote first surfaced in a 2001 marketing magazine letter, 54 years after Ford died. No historian has found it in his autobiography, his letters, or the Ford Museum’s archive of 200 verified quotes.

Which makes the irony perfect. The most famous quote about not thinking big enough is itself a myth we repeat without questioning. We do not even check the source of the story we use to warn people about not checking their assumptions.

That is exactly what is happening with AI right now.

Most companies are building faster horses. They take an existing process, bolt AI onto it, and call it transformation. Customer service gets a chatbot. Legal gets a contract summarizer. Marketing gets a content generator. The process stays the same. It just runs quicker.

McKinsey’s 2025 State of AI report puts a number on this. Only 55% of AI high performers actually redesign their workflows. The rest accelerate what already exists. MIT Sloan found the bottleneck is not model quality or data. It is organizational design. The technology is ready. The thinking is not.

Ford’s actual contribution was not speed. He did not make carriages faster. He made the carriage irrelevant. The assembly line was not an improvement to an existing production method. It was a fundamentally different way of organizing work, materials, and people. The resistance was not technical. It was conceptual. People could not see past the cart.

The pattern repeats with every general-purpose technology. Early automobiles looked like horse carriages with engines strapped on. Early television was filmed radio. Early websites were digital brochures. In every case, the first instinct was to pour the new capability into the old container. Real value only emerged when someone asked a different question. What can we do now that was impossible before?

A telecom cutomer of ours did not use AI to make customer service faster. They restructured it. Their AI assistant handles the workload equivalent of 50 agents, resolving issues in 45 minutes instead of 7 hours. The difference is not speed. It is architecture. They did not optimize the old model. They replaced it.

Three questions for any AI project. Are we making the existing process faster, or are we questioning whether the process should exist? Are we measuring speed improvement, or capability expansion? Would Henry Ford, the real one, recognize what we are doing as a faster horse?

If the answer to that last question is yes, you are building the wrong thing.

—
Sources:
– Quote Investigator (2011), Snopes (2025), HBR (2011)
– McKinsey, “The State of AI in 2025”
– HBR, “The Last Mile Problem Slowing AI Transformation” (2026)

Posted in Artifical Intelligence, Business, Think Different | Comments Off on The faster Horse that never was

Think Different about AI

Many organizations treat AI like a coachman treats a combustion engine. They strap it to the old cart. These series of posts each expose a different version of that fallacy, building from the individual mental model to the organizational operating model.

Posted in Artifical Intelligence, Business, Think Different | Comments Off on Think Different about AI

Groundhog Day, actually Year

Here we go again. The local journalistic farce just got an extension. The article in German for the 2026 edition of the Groundhog Day.

Posted in Uncategorized | Comments Off on Groundhog Day, actually Year

Elections 2026 – a short review and a promise

I’m member of the council of my village Zollikon. In German a short review and a promise for the next four years ahead of the elections in March 2026 – German only.

Posted in Uncategorized | Comments Off on Elections 2026 – a short review and a promise

A little local journalistic farce

A local news blog published an online slander for no reason. It’s all a farce. In German only.

Posted in Uncategorized | Comments Off on A little local journalistic farce

The State of Play in RAG and Enhanced RAG

In the rapidly evolving landscape of artificial intelligence, Retrieval-Augmented Generation (RAG) and its enhanced versions are gaining significant traction. These technologies are pivotal in advancing the capabilities of generative AI by integrating retrieval mechanisms with generative models, thereby improving the accuracy and relevance of AI-generated content.

RAG: A Brief Overview

RAG combines the strengths of retrieval systems and generative models. It retrieves relevant documents or data points from a vast corpus and uses this information to generate more accurate and contextually relevant responses. This approach is particularly beneficial in scenarios where the generative model alone might lack the necessary context or factual accuracy.

Enhanced RAG: The Next Step

Enhanced RAG takes this concept further by incorporating additional layers of intelligence and scalability. For instance, Squirro’s Enhanced RAG platform integrates a knowledge graph to improve retrieval accuracy and contextual understanding. This integration allows for a more nuanced understanding of the data, enabling the system to provide more precise and contextually aware responses.

Current Developments and Use Cases

Recent developments in Enhanced RAG focus on scalability and enterprise-level applications. Squirro has demonstrated the implementation of a permission-enabled enterprise RAG setup at the European Central Bank, showcasing the platform’s ability to handle complex, large-scale data environments. Additionally, the integration of out-of-the-box features and automation capabilities is redefining use cases, making Enhanced RAG a versatile tool for various industries.

Conclusion

The evolution from basic RAG to Enhanced RAG represents a significant leap in the capabilities of generative AI. By leveraging advanced retrieval techniques and integrating knowledge graphs, Enhanced RAG platforms are setting new standards for accuracy and contextual relevance in AI-generated content. As these technologies continue to develop, they promise to unlock new possibilities across different sectors, from finance to healthcare.

Posted in Uncategorized | Comments Off on The State of Play in RAG and Enhanced RAG

AI is like WD40: it permeates everything, including productivity

Reply to the articla “KI beflügelt nicht primär die Produktivität, sondern mehr noch die Bürokratie“

In his article, Mathias Binswanger argues that artificial intelligence (AI) currently does not primarily increase productivity, but instead promotes bureaucracy. While there are legitimate concerns that AI applications could create new bureaucratic processes, this viewpoint neglects an important historical perspective and underestimates the transformative potential of AI.

More in German

Posted in Uncategorized | Comments Off on AI is like WD40: it permeates everything, including productivity

What Grok* thinks about Squirro

I asked X Grok (*Grok is Twitter’s, oh, X’s alternative to ChatGPT) about Squirro and if I should buy it (of course – I already did ;-). Here’s Grok’s response:

Ah, Squirro AI, the Swiss Army Knife of business intelligence, or as I like to call it, the “Squirro-nado” of data insights! Here’s why you might want to dive into the Squirro pool:

What is Squirro AI?

  • It’s an AI-powered business intelligence platform. Imagine if your company’s data could not only talk but also sing opera. That’s Squirro. It uses AI to analyze, understand, and visualize your data in ways that make your spreadsheets look like they’re from the stone age.
  • Natural Language Processing (NLP): You can ask Squirro questions like you’re chatting with a friend. “Hey Squirro, what’s up with our sales in Q3?” and it’ll give you the rundown without making you feel like you’re back in math class.
  • Data Integration: It’s like the ultimate party host for your data. It integrates data from all corners of your business, making sure everyone’s invited to the data party.
  • Predictive Analytics: Squirro doesn’t just tell you what happened; it’s like a fortune teller for your business, predicting trends and outcomes based on current data.

Why Should You Buy It?

  1. Save Time, Save Money: Time is money, and Squirro helps you spend less of both. It automates data analysis, so your team can focus on strategy rather than number-crunching.
  2. Better Decision Making: With insights that are more ‘Sherlock Holmes’ than ‘Scooby-Doo’, you’ll make decisions based on deep, AI-driven insights rather than gut feelings or guesswork.
  3. Competitive Edge: While your competitors are still figuring out Excel, you’ll be using AI to predict market trends, customer behavior, and more. It’s like having a crystal ball, but one that’s actually useful.
  4. Scalability: Whether you’re a startup or a multinational, Squirro grows with you. It’s like buying a pair of shoes that never wear out and always fit perfectly.
  5. User-Friendly: Even if your tech skills are stuck in the ’90s, Squirro’s interface is designed to be intuitive. It’s like having a personal data assistant who’s also your best friend.
  6. Security: With data breaches being the new ‘plague’, Squirro takes security seriously, ensuring your data is safer than a dragon’s treasure.
  7. Customization: Tailor it to your business needs. It’s like having a suit made just for you, but instead of fabric, it’s data.
  8. Innovation: By adopting Squirro, you’re not just buying a tool; you’re embracing innovation. It’s like being the first on your block to get a smartphone when everyone else is still using flip phones.

In essence, Squirro AI is like upgrading from a bicycle to a jetpack in the world of data analysis. If your business thrives on data-driven decisions (and let’s be honest, which one doesn’t?), Squirro could be the superhero cape you didn’t know you needed. Just remember, with great data comes great responsibility!

Posted in Uncategorized | Comments Off on What Grok* thinks about Squirro

Darwin at his best: the GenAI race

Darwin is renowned for his theory of evolution and natural selection. Looking at the GenAI space I get that same feeling. What looks easy at start is rough in detail and very selective. Let me explain:

Scaling a GenAI solution to large-scale production requires meticulous alignment of various components. Each step is essential for a full production rollout, and without prior experience, each comes with a steep, costly, and time-consuming learning curve. Let’s examine these steps:

  • Data Ingestion at Scale: While a pilot may require only a few hundred or thousand documents to prove the concept, the challenge escalates when moving to production with millions of documents. Ingesting data at scale in environments with high data velocity demands robust infrastructure and excellent operational capabilities.
  • Data Enrichment: Often overlooked in small pilots, data enrichment is crucial for a successful RAG (Retrieval-Augmented Generation) setup. The inbound data must be well-enriched for the retrieval component to yield accurate results. For instance, managing versioned documents and ensuring the latest version is retrieved is a complex task.
  • RAG Search: Sounds simple, is complex. All too often the systems out there produce seemingly good but factually incorrect results. After all search is probabilistic, so are LLMs. Multiplied probabilities don’t yield better results. So, you need to work hard on the retrieval part to really get good results, e.g. by factoring in better enriched data to produce more accurate result sets before exposing them to an LLM to produce a natural language looking result (and by using the LLM what it is good for, that is text comprehension and text generation and not what it is not good for, that is search).
  • Maintaining a Vectorized Index: Building a vectorized index for a proof of concept is relatively straightforward. However, operating and continuously updating a large-scale index, often terabytes in size, is an entirely different challenge. And you need to master the art to at the same time update a large (TBs of data) index while querying that same index.
  • Security: Security is critical in any enterprise setup. Role-based access control, which dictates who can see what, is more complex in a GenAI setup. Ensuring access control for every vectorized chunk and updating the index with every role change is no easy feat.
  • Guardrails: All the above are required steps for a good RAG stack and yet not sufficient to guarantee consistently good, accurate and trustworthy results (people take results as is. Why? That’s for another blog post). It’s wise to integrate guardrails at prompt time (create better quality prompts) and at result time (comprehensive answer validation).
  • Integration with Existing Systems: Integrating the solution into an existing enterprise setup is another significant challenge. Simply adding another dashboard is not useful. The solution must interact with third-party systems, often in a bi-directional manner.
  • Testing: Any such complex beast requires a lot of continuous testing & monitoring of both the actual processing steps and also the result quality. Any subtle change of any of the components (say retrieval adjustments, say new LLM model) will have an impact on result output. You need to measure this continuously and refine your setup accordingly.
  • Operational Maintenance: Finally, operating the entire setup over time and ensuring high availability levels is essential.
  • Measurable ROI: A common oversight is the lack of a clear cut business case detailing the cost-benefit ratio of any such rollout (they are impressive if well done*) By mid-2025, CFOs will start asking tough questions: “Where’s the bang for my buck?”

Delivering reliable, accurate, and transparent AI at scale is not easy, is costly, and yet if well done will yield massive benefits. At Squirro we’ve been doing this for a while – we happily share our experience. Have a look out our Knowledge Hub or simply get in touch!

PS: They are impressive indeed if you go beyond search and see this AI transformation as the biggest opportunity since the dawn of the (commercial) Internet of disintermediation and reintermediation of entire value chains. We will discuss this in a forthcoming blog post.

Posted in Uncategorized | Comments Off on Darwin at his best: the GenAI race

EU accession through the back door

A year ago, Credit Suisse went under. The rescue required 1/3 of Switzerland’s gross national product. CS was about a third as big as UBS. In other words, a new rescue (and most G-SiBs have come knocking in the recent past) would require the Swiss gross national product.

Even our federal government and the SNB cannot afford that. Who can?

  • Chinese National Bank: Probably not
  • Japanese National Bank: We have other challenges
  • Fed: OK, Switzerland is actually a little closer to Washington than Hawaii, but becoming the 51st state of the USA?
  • That leaves the ECB. It could probably handle it. Yet we are not a member (yet). But the constraints described above will bring us closer to Europe…

Posted in Uncategorized | Comments Off on EU accession through the back door