Question 1

Did you actually replace yourself as a teacher with an AI clone?

Accepted Answer

_A teaching experiment reveals AI's limits in education._

I built one of the first online certificates, Designing and Building AI Solutions, where I replaced myself with an AI clone. The clone is public. It's now a no-code course where students can imagine business ideas and have the AI clone write the necessary code. We have a 24/7 teaching assistant. But did this mean I replaced myself as a teacher? Not even close. AI can summarize information, but it can't predict future trends. Teaching is more than just delivering knowledge. It's about understanding and evaluating AI trends to find the best ways to use AI for business.

— [Is AI Replacing Us — Good News For Knowledge Workers](https://www.forbes.com/sites/lutzfinger/2025/01/22/is-ai-replacing-us-good-news-for-knowledge-workers/) · _Forbes_

Question 2

How are AI agents different from RPA (robotic process automation)?

Accepted Answer

_Agents add reasoning where RPA's rule engine stops._

RPA is a rule engine. It automates structured, low-skilled tasks where every decision is essentially zero or one, which is why UiPath and Automation Anywhere built real businesses but hit a ceiling: a script either breaks or runs, and it cannot handle uncertainty or unstructured data. Agents add a language model on top, so they can read an email's style, cope with ambiguity, and self-correct. Cherry's portfolio company Octomind does this for front-end testing with a self-healing loop that notices when a test goes wrong and tries another path, something a rule engine cannot do. The catch: you still have to give an agent clear boundaries, or it goes rogue.

— [The Edge: “AI Agents: The Future of Work?”](https://www.lutzfinger.com/transcript/2024-12-12-ai-agents-the-future-of-work/) · _Podcast_

Question 3

How can AI support frontline workers who aren't sitting at computers all day?

Accepted Answer

_Think voice-first interfaces and knowledge retrieval, not just desktop tools._

Frontline workers move around and might not have time to sit at a laptop, but they could have a quick second to speak or type a question. You could support them with an AI interface that answers questions in their language, translates into technical knowledge, and searches knowledge databases. The key is making sure answers are right. Don't just put ChatGPT on your FAQs. You need human in the loop at the beginning, testing a lot to ensure frontline workers are happy with the output before rolling it out widely.

— [You asked, we answered](https://www.lutzfinger.com/transcript/2023-07-28-you-asked-we-answered/) · _The Edge_

Question 4

How did LinkedIn use knowledge graphs to solve the job title problem?

Accepted Answer

_Structuring 150 million job titles required more than simple categorization._

LinkedIn had over 150 million unique job title strings, which was impossible to model effectively. We decomposed titles into four facets: seniority, employment status, role, and specialty. For example, Senior Software Engineer contains seniority information, a role, and a specialty. We then built knowledge graphs for each dimension to show how values interrelated. This helped us understand completeness and vagueness of titles. A VP at a bank means something completely different than a VP at a tech company. Taxonomies alone were not expressive enough to capture these nuances.

— [The synergy between LLMs and knowledge graphs](https://www.lutzfinger.com/transcript/2023-08-24-the-synergy-between-llms-and-knowledge-graphs/) · _The Edge_

Question 5

How did OCR technology affect the number of lawyers?

Accepted Answer

_OCR reduced costs but increased demand for lawyers overall._

Before OCR, legal teams relied on staff to manually read, annotate, and organize documents, a slow and expensive process. With OCR, lawyers could search and analyze text instantly, dramatically reducing costs. By lowering costs, OCR democratized access to legal services, increasing the overall demand for lawyers. Today, AI is once again reshaping legal work. FLANK.AI automates legal document summarization. Instead of spending hours reviewing paperwork, lawyers can now focus on higher-value tasks like strategy, case analysis, and client advocacy.

— [Is AI Replacing Us — Good News For Knowledge Workers](https://www.forbes.com/sites/lutzfinger/2025/01/22/is-ai-replacing-us-good-news-for-knowledge-workers/) · _Forbes_

Question 6

How did Varo use AI to reduce operational costs while improving customer service?

Accepted Answer

_An AI chatbot cut costs 60% while handling most inquiries._

We started by providing voice support for everybody, but it got very expensive. We moved to live person chat, then built an AI chatbot. That chatbot now contains over 55% of our inbound inquiries and allowed us to reduce operational costs by about 60%. This is part of solving the high-cost problem that made serving underserved populations uneconomical for traditional banks. The AI handles routine questions efficiently while maintaining service quality for customers who value access and responsiveness.

— [Disrupting Traditional Banking, with Colin Walsh](https://www.lutzfinger.com/transcript/2025-05-22-disrupting-traditional-banking-with-colin-walsh/) · _The Keynote on AI_

Question 7

How do chained AI models create better workflows than simple generative AI tools?

Accepted Answer

_Building dynamic workflows by connecting AI models in sequence._

Chained AI models connect multiple models together so the output of one becomes the input for the next, creating dynamic workflows rather than fixed steps. For an email assistant, the system first analyzes the email and figures out needed steps, then checks the calendar, looks up past emails, and predicts likelihood of attending. It creates three draft replies with different options and presents them to the user. This approach enables system integration, context search through RAG, predictions using traditional AI, and user control through multiple options.

— [AI Agents In 2025: What Enterprise Leaders Need To Know](https://www.forbes.com/sites/lutzfinger/2025/01/05/ai-agents-in-2025-what-enterprise-leaders-need-to-know/) · _Forbes_

Question 8

How do venture firms use AI in their own operations?

Accepted Answer

_AI in sourcing, diligence, portfolio construction, and founder support._

> AI — it's who controls the data and who builds products that businesses and people actually use.

— [Stargate: The $500 Billion Message To China — The Future Of AI Regulation](https://www.forbes.com/sites/lutzfinger/2025/01/22/stargate-the-500b-message-to-china---the-future-of-ai-regulation/) · _Forbes_

Question 9

How does accountability work when AI agents are doing tasks for us?

Accepted Answer

_You remain accountable for your AI agents, like a CEO for employees._

If people write an email and that email is wrong, who is accountable for it? Should I say it was my AI agent? Not quite. It was me who trained and set up and focused an AI agent. We talk about the new role that all of us become managers or CEOs of many agents. If that description is true, then the accountability lies in you, the CEO. When something happens at a traditional corporate, the CEO might lose their job if somebody of their employees did not do the right thing, because ultimately they are responsible.

— [From Google to Cornell: Rethinking AI Workflows | Lutz Finger | Build AI Podcast 17](https://www.lutzfinger.com/transcript/2026-04-16-from-google-to-cornell-rethinking-ai-workflows-lutz-finger-build-ai-podcast-17/) · _Build AI_

Question 10

How does Catalog ensure certain workflows always follow the same steps when using LLMs?

Accepted Answer

_A rules engine ensures critical workflows always execute consistently._

We built a workflow engine underneath our LLMs. The LLM generates a workflow on the fly every time, but users can design deterministic journeys inside our system using workflows (like Zapier but internal). If a request matches a workflow definition, it triggers that workflow every time instead of a dynamically generated journey. It's no code, so users can do it themselves. First, the LLM does a workflow lookup to see if there's something you always want to do, then for the rest it goes into free flow logic.

— [Large Language Models: one interface to rule them all?](https://www.lutzfinger.com/transcript/2023-08-15-large-language-models-one-interface-to-rule-them-all/) · _The Edge_

Question 11

How does Catalog use large language models to simplify enterprise software access?

Accepted Answer

_One interface to access all your enterprise tools at once._

Catalog uses LLMs to create a unified interface for all enterprise software. Instead of navigating multiple systems, employees enter a single prompt. The LLM understands the query, identifies which data repository holds the answer (like Workday for payslips), converts the request into an API call, performs deterministic retrieval, and returns the result. We don't use LLMs for data retrieval itself, only for reasoning and synthesis. This compresses workflows from multiple surfaces and ten to fifteen clicks down to one query and two or three clicks.

— [Large Language Models: one interface to rule them all?](https://www.lutzfinger.com/transcript/2023-08-15-large-language-models-one-interface-to-rule-them-all/) · _The Edge_

Question 12

How does Zenith OS reduce costs in neuropsychiatry treatment?

Accepted Answer

_AI cuts neuropsychiatry costs by reducing required therapist headcount per session._

Neuropsychiatry treatments like ketamine and psilocybin for PTSD, depression, and anxiety are 30 to 40 percent more effective than SSRIs but remain inaccessible due to high human capital costs. Currently, two therapists must be present during treatment because patients are in a vulnerable state and there are sexual harassment concerns. Each therapist charges $100 to $150 per hour. Zenith OS uses AI to augment administrative tasks and help with patient recommendations, allowing clinics to operate with one therapist instead of two, significantly reducing costs and expanding access to these groundbreaking treatments.

— [Founder Coaching: Tips for your AI Journey](https://www.lutzfinger.com/transcript/2024-07-26-founder-coaching-tips-for-your-ai-journey/) · _The Edge_

Question 13

How is AI changing the role of software developers?

Accepted Answer

_Coding goes away, but software remains. Developers become technical architects._

Software developers are taking on a much more abstract role, similar to technical product managers. Many AI-first companies claim they're writing 80 to 90 percent of their code via AI, and a replacement event is beginning to occur. While coding as an activity is probably going away, software itself is still needed. Developers will need to be quite specific about how they want code to be stitched together and architected. They're essentially moving to a higher level of abstraction, just like when we went from punch cards to compilers to higher order languages.

— [Reshaping Investing With AI, with Lila Tretikov](https://www.lutzfinger.com/transcript/2026-02-17-reshaping-investing-with-ai-with-lila-tretikov/) · _The Keynote on AI_

Question 14

What are some examples of how healthcare providers are using LLMs with patients?

Accepted Answer

_Chatbots and knowledge systems deliver personalized, educational patient engagement._

Healthcare providers are using AI-powered chatbots from Value-Based Care providers that send tailored advice based on a user's personal preferences. They're also using knowledge base systems that explain a potential diagnosis and educate patients on pathways. By combining AI technology with traditional methods such as care management programs or wellness initiatives, healthcare providers have an enhanced opportunity for achieving better outcomes at lower costs while providing personalized, human-centric care.

— [Large Language Models & AI In Healthcare](https://www.forbes.com/sites/lutzfinger/2023/03/27/large-language-models--ai-in-healthcare/) · _Forbes_

Question 15

What are the main problems with using simple GPT wrappers for enterprise AI applications?

Accepted Answer

_Why basic ChatGPT integrations fail in business settings._

GPT wrappers have several critical shortcomings that prevent them from working well in enterprise settings. They lack system integration, so they cannot check calendars or access other tools. They have no context about user preferences or past interactions. Security is missing, which means they could accidentally reveal private details. They lack guardrails to handle controversial questions appropriately. They offer no user control over outputs. And they suffer from hallucinations, where the AI sometimes makes things up. These issues make simple wrappers unsuitable for real enterprise deployment.

— [AI Agents In 2025: What Enterprise Leaders Need To Know](https://www.forbes.com/sites/lutzfinger/2025/01/05/ai-agents-in-2025-what-enterprise-leaders-need-to-know/) · _Forbes_

Question 16

What did you get wrong about multimodal AI interfaces?

Accepted Answer

_Right about multimodal potential, but overly optimistic on timing and control._

We were right that models can encode anything into a latent representation, like Esperanto for computers, and switch back and forth between text, voice, and images. We saw this with clip, voice to voice, encoding from one language into voice representation of another. But we missed the timeline. It's not as strong as we hoped. The challenge is LLMs are only probability distributions. They're not controllable. If you have a machine that's not controllable, it sucks as a machine. That's okay for your friend at the bar at 2 a.m., not your computer.

— [Did our predictions hold true?](https://www.lutzfinger.com/transcript/2024-03-27-did-our-predictions-hold-true/) · _The Edge_

Question 17

What do you mean when you say winners won't be those who use AI, but those who redesign work around it?

Accepted Answer

_Business needs remain constant, workflows must evolve._

Our business hasn't changed. We still have needs, we still travel, we still consume, we still build. The only question is how can I do this business more effectively? Yes, there will be people who use ChatGPT to write an email, but the winners will be those who either build the platforms or rearrange the workflows in order to be effective. The focus is on the workflow, the surrounding, the setup. How do we empower you in your environment, not can you use it?

— [From Google to Cornell: Rethinking AI Workflows | Lutz Finger | Build AI Podcast 17](https://www.lutzfinger.com/transcript/2026-04-16-from-google-to-cornell-rethinking-ai-workflows-lutz-finger-build-ai-podcast-17/) · _Build AI_

Question 18

What does 'AI as an operating model' mean?

Accepted Answer

_Why the winners redesign work around AI rather than just adopting tools._

> AI — it's who controls the data and who builds products that businesses and people actually use.

— [Stargate: The $500 Billion Message To China — The Future Of AI Regulation](https://www.forbes.com/sites/lutzfinger/2025/01/22/stargate-the-500b-message-to-china---the-future-of-ai-regulation/) · _Forbes_

Question 19

What is a reward engineer and why does Richard predict this will be an important new job type?

Accepted Answer

_Defining AI success requires a new kind of manager and skill set._

A reward engineer is someone who defines what success looks like when delegating tasks to AI. Richard predicts this will be a main new job type because AI can be very smart but needs unambiguous instructions. For example, if you tell an AI to increase customer satisfaction scores, it might create bots to spam five-star ratings or give thousand-dollar gift certificates for ten-dollar problems. Reward engineers will need to think through all the ways to properly define success, refine rewards, and become better at delegation and managing AI with appropriate trust and verification.

— [The Future of AI, with Richard Socher](https://www.lutzfinger.com/transcript/2026-02-05-the-future-of-ai-with-richard-socher/) · _The Keynote on AI_

Question 20

What is an AI agent, really?

Accepted Answer

_Not a sci-fi character, a workflow that chains steps._

An agent is not a sci-fi character and it is not one giant model. It is a piece of code that takes a defined task, breaks it into steps, and lets you act on software through language. The core idea is chaining: instead of asking a model to "improve revenue," you break the goal into step one, step two, step three, exactly the way a manager hands work to a junior employee. Tools like LangChain popularized this. So an agent is less a breakthrough model than a workflow that combines models, tools, and sometimes plain deterministic code, each doing one step.

— [The Edge: “AI Agents: The Future of Work?”](https://www.lutzfinger.com/transcript/2024-12-12-ai-agents-the-future-of-work/) · _Podcast_

Question 21

What is the augmentation thesis for AI in workflows?

Accepted Answer

_AI can transform repetitive workflows into seamless, conversational interactions._

Augmentation is about making everyday workflows better through AI. Many workflows involve humans taking information, remodeling it, and placing it again, like looking up documents or writing summaries. AI can make these interfaces easier. For example, instead of selecting dates in a booking service, you could talk in normal language. Or when requesting vacation, instead of checking your ERP system, emailing your team, and setting an out-of-office reminder separately, you could just say you want vacation and the computer takes over everything. It's like robotic process automation but at the next level, because large language models understand language and can help prepare decisions.

— [Cherry's Investment Thesis on AI](https://www.lutzfinger.com/transcript/2024-07-19-cherrys-investment-thesis-on-ai/) · _The Edge_

Question 22

What is the BizML framework and why is it needed for AI projects?

Accepted Answer

_*A business framework for predictive AI, not just technical tools*_

BizML is a six-step framework I created that ushers predictive AI projects successfully from inception to deployment. It's not like MLOps or any technical solution. It's an organizational procedure, a management protocol that bridges the tech-biz divide. It brings together business-side stakeholders with data scientists through semi-technical collaboration. The framework addresses what's predicted, how well, and what's done about it. Without this kind of practice or framework, most business people don't realize you need a very specialized approach to get these projects successfully through to deployment.

— [The Value Translation Gap: AI's Deployment Problem](https://www.lutzfinger.com/transcript/2024-11-28-the-value-translation-gap-ais-deployment-problem/) · _The Edge_

Question 23

What is the most important format for presenting dealership content to AI models?

Accepted Answer

_Q&A format reduces cognitive load and increases AI model pickup._

Package your content in Q&A format. When someone asks a question, the model looks for the least cognitive load, very much like humans. If it finds the exact question already formatted with an answer, it just picks that up rather than reading a sentence and having to think about extracting the answer. That costs money for the model. So take your newsletters from the last 10 years, use an AI tool to generate Q&As from that authentic content, and put it on your website. The content should be human generated, ideally based on your own customer data.

— [Lutz Finger | From SEO to GEO: How AI is Changing How Car Buyers Find You](https://www.lutzfinger.com/transcript/2026-05-22-lutz-finger-from-seo-to-geo-how-ai-is-changing-how-car-buyers-find-you/) · _Three Key Insights_

Question 24

What is the practical framework for implementing AI agents in enterprise settings?

Accepted Answer

_Five steps to deploy AI agents that create business value._

The framework I recommend has five essential steps. First, define the business objective to ensure you are solving a real problem. Second, collect and clean the data needed to power the workflow. Third, develop the agentic workflow by chaining AI models appropriately. Fourth, test with actual users to validate the solution works. Fifth, create a feedback loop for continuous improvement. Most importantly, the focus must remain on value creation throughout the entire process, not just on implementing technology for its own sake.

— [AI Agents In 2025: What Enterprise Leaders Need To Know](https://www.forbes.com/sites/lutzfinger/2025/01/05/ai-agents-in-2025-what-enterprise-leaders-need-to-know/) · _Forbes_

Question 25

What kind of AI agents will actually enter the workforce in 2025?

Accepted Answer

_Specialized workflow agents, not AGI, will join teams in 2025_

In 2025, we'll see more AI agents entering the workforce, transforming workflows by simplifying, enhancing, and automating tasks across industries. These won't be all encompassing AGI models but smaller, specialized models designed for dedicated workflows. AI will scale and improve processes one step at a time, combining traditional AI, context retrieval called RAG, and robust user design. Soon, there will be job websites with posts that only AI agents can apply for. 2025 will be the year AI colleagues enter our lives.

— [The Race For AI Agents. Who Will Supply Tomorrow's Workforce?](https://www.forbes.com/sites/lutzfinger/2025/01/07/the-race-for-ai-agents-who-will-supply-tomorrows-workforce/) · _Forbes_

Question 26

What makes autonomous AI agents different from traditional automation tools like RPA?

Accepted Answer

_How AI agents differ from traditional automation like RPA_

Traditional RPA is basically scripting with rules the machine follows, without much deviation except for human input. It's a structural, human-made approach. Autonomous agents are different because they use large language models as an interface that understands human language instead of requiring rigid API calls or workflows. This means I don't need to design specific engineering interfaces. The language model can interact with tools like calculators, internet searches, or order engines using natural language, breaking down tasks iteratively rather than following predetermined scripts.

— [Connecting LLMs to the world: The possibilities of the feedback loop](https://www.lutzfinger.com/transcript/2023-05-02-connecting-llms-to-the-world-the-possibilities-of-the-feedback-loop/) · _The Edge_

Question 27

What makes the grid so complex and difficult to manage compared to other infrastructure?

Accepted Answer

_The grid is an electron highway with no traffic lights or routing._

The grid is mankind's largest and most complex machine. It is a continuous supply and demand balancing mechanism connecting generation with demand across various voltage levels. Electrons follow the laws of physics, from high potential to low potential, like a river system. There are no stop signs, and we cannot assign electrons an IP address and route them easily. The grid is also very opaque, with not enough sensors and monitors. Most of it was built during the post World War II period. Now it is becoming a two way system with distributed solar and wind, which becomes extremely complicated to manage.

— [AI Needs Power, the Grid Needs AI, with Josh Wong](https://www.lutzfinger.com/transcript/2026-04-02-ai-needs-power-the-grid-needs-ai-with-josh-wong/) · _The Keynote on AI_

Question 28

What patterns make a workflow 'agentic' rather than just a wrapper around an LLM?

Accepted Answer

_Reflection, self-correction, knowledge expansion, coordination._

A few recurring patterns. Reflection: the system plans, asks "is this approach good, does this make sense, can you double-check," and only then executes, the way a reasoning model breaks a request into several approaches, picks one, and verifies the answer. Self-correction: it looks at the result, notices it went wrong, and tries a different route. Knowledge expansion: when the answer is not in its memory, it searches the web for current information. Coordination: it splits a job into sub-tasks and routes each to the right tool, for example handing a math question to a calculator instead of guessing. Wrapping a chat box around an LLM is none of these.

— [The Edge: “AI Agents: The Future of Work?”](https://www.lutzfinger.com/transcript/2024-12-12-ai-agents-the-future-of-work/) · _Podcast_

Question 29

What practical business use cases exist for language model chaining beyond autonomous agents?

Accepted Answer

_Real business applications from databases to booking to summarization_

There are many business cases beyond the autonomous agent hype. Question answering over a database is one: use a language interface to access the database, figure out the information, distill it, and report back. Chatbots interacting with booking systems need to understand how to access the system and remember personal preferences. Document summarization is another: if I take all seven Harry Potter books, I can't paste them into OpenAI because it's a finite machine. Instead, summarize page by page, store that in memory, then do iterative summarization to get the final result. These applications are very practical.

— [Connecting LLMs to the world: The possibilities of the feedback loop](https://www.lutzfinger.com/transcript/2023-05-02-connecting-llms-to-the-world-the-possibilities-of-the-feedback-loop/) · _The Edge_

Question 30

What problem did transformers solve for Ultimate AI's multi-language support?

Accepted Answer

_One transformer model replaced 100 language-specific models per customer_

Before transformers, we had different models for different languages. Supporting 100 languages meant running 100 different models per customer, which was painful for model operations. When we switched to transformers like BART about three and a half years ago, we could use one polyglot model for all languages. You could train the bot in English and ask questions in Finnish, and it worked reasonably well. It also helped our clustering pipeline show customers all their questions across languages semantically grouped together. You could see Finnish, Spanish, German, and English examples all meaning the same thing without worrying about the language.

— [How Ultimate evolved from hundreds of supervised models to UltimateGPT](https://www.lutzfinger.com/transcript/2023-05-12-how-ultimate-evolved-from-hundreds-of-supervised-models-to-ultimategpt/) · _The Edge_

Question 31

What was Ultimate AI's key insight that made customer onboarding faster than competitors?

Accepted Answer

_Automated clustering replaced manual training data creation for faster setup_

We couldn't afford AI trainers manually creating training data, so we built a clustering pipeline that analyzed historical customer conversations already sitting in CRMs like Salesforce. This let us automatically find the 100 or 200 most common questions and show customers all 30,000 actual ways users asked questions like where's my order. Competitors were using FAQ pages and having people manually write 50 to 100 variations, which didn't reflect real customer language. Our approach gave faster time to value because we used actual conversation data instead of imagined variations, and customers could see their real top questions before even talking to us.

— [How Ultimate evolved from hundreds of supervised models to UltimateGPT](https://www.lutzfinger.com/transcript/2023-05-12-how-ultimate-evolved-from-hundreds-of-supervised-models-to-ultimategpt/) · _The Edge_

Question 32

What will actually happen with AI in 2025 instead of AGI?

Accepted Answer

_Narrow AI chains will excel at tasks without achieving AGI._

In 2025, we'll see more narrow AI solutions integrated into chains similar to OpenAI's o1 approach. These systems will excel at specific tasks and, when combined, will enhance productivity and surpass human performance in many areas. This development will be exciting, but these advancements will not constitute AGI. Instead of Artificial General Intelligence, large language models will find their killer app. The real story in 2025 will be this killer application of large language models.

— [No AGI But A 'Killer App' — 2025 AI Prediction](https://www.forbes.com/sites/lutzfinger/2025/01/03/no-agi-but-a-killer-app2025-ai-prediction-110/) · _Forbes_

Question 33

What's the biggest mistake CEOs make when trying to integrate AI into their companies?

Accepted Answer

_Don't just replace humans, reimagine the entire workflow_

The challenge is not just replacing humans in existing workflows. It's reimagining how the workflow actually would work without the human in the loop. When we had the first wired telephone, they used it to transmit symphonies because they thought that was the killer app. It took quite a while until Spotify actually made the news. Your challenge as a CEO is to not go through every workflow and say how can I replace my customer care agent. That's an old workflow. Maybe build it into the product experience so you don't need a customer AI agent in the first place. That is the change.

— [Deep Dive Into The Current And Future State Of Ai With Lutz Finger](https://www.lutzfinger.com/transcript/2025-04-24-deep-dive-into-the-current-and-future-state-of-ai-with-lutz-finger/) · _The Pulse of AI_

Question 34

Why are AI workflows the real moat?

Accepted Answer

_How redesigning work around AI beats model upgrades._

> SaaS can build the new AI-driven workflows on top of its user data and customer surfaces. But workflows are brittle. One new prompt, one model update, and the whole chain quietly breaks. I watch this every time I get students [coding agents in my Cornell workshops](https://ecornell.cornell.edu/courses/artificial-intelligence/building-enterprise-ready-ai-agents-from-prototypes-to-production/).

— [The Missing Moat In AI: Your Eval Data](https://www.forbes.com/sites/lutzfinger/2026/05/26/the-missing-moat-in-ai-your-eval-data/) · _Forbes_

Question 35

Why are coding applications seeing the biggest success with AI agents right now?

Accepted Answer

_Coding succeeds because compilers provide clear feedback loops_

We see the biggest success in coding because there's a clear feedback loop. When I ask an AI to write code, it can check if the code compiles. If it doesn't work, ChatGPT critiques that and the next step is to rewrite the code including the critique. This iterative loop continues till it actually works. The combination of knowing what was put in previously through memory, plus the ability to talk to a compiler, plus the iteration workflow creates a loop to generate good code. It's more difficult when the task isn't very clear or doesn't have such obvious success criteria.

— [Connecting LLMs to the world: The possibilities of the feedback loop](https://www.lutzfinger.com/transcript/2023-05-02-connecting-llms-to-the-world-the-possibilities-of-the-feedback-loop/) · _The Edge_

Question 36

Why are radically lean teams the future?

Accepted Answer

_How AI compresses the team size needed to ship._

> But what hinders many of the large corporates to adapt them? >>Yeah, well, I think there's a lot of components there but in the end, it comes down to the size, and the politics, and the setup they built. Large companies are usually built to mitigate risk and focus on their core business. If there's a new technology that comes in that is not obviously threatening the core business, or opening a huge opportunity for the core business,…

— [Unit 5 Full Transcript](https://www.lutzfinger.com/ecornell/) · _eCornell course_

Question 37

Why did many generative AI projects fail to scale in 2024?

Accepted Answer

_The two fatal flaws in most 2024 AI implementations._

Many companies rushed to try generative AI in 2024, and some even used chained models. But most projects failed to scale for two main reasons. First, they were test projects not integrated into system infrastructure, meaning they existed in isolation rather than connecting to existing business systems. Second, they were missing guardrails and quality controls necessary for production deployment. Without proper integration and controls, these pilot projects could not move beyond experimentation to deliver real business value at scale.

— [AI Agents In 2025: What Enterprise Leaders Need To Know](https://www.forbes.com/sites/lutzfinger/2025/01/05/ai-agents-in-2025-what-enterprise-leaders-need-to-know/) · _Forbes_

Question 38

Why did Ultimate AI initially focus on helping human agents instead of automating customer conversations directly?

Accepted Answer

_Agent assistance was a learning ground before full automation_

We turned the problem around. Instead of having consumers talk directly to bots, we gave recommendation tools to human agents. Agents had more tolerance for failures because they could ignore incorrect suggestions and still do their jobs, but when recommendations were correct they got value and could train the system. This was less stressful for building products. We could iterate and fail without ruining the customer experience. The technology wasn't good enough yet for full automation, so this approach let us learn gradually. Once models improved, we could then turn it back to direct user interaction.

— [How Ultimate evolved from hundreds of supervised models to UltimateGPT](https://www.lutzfinger.com/transcript/2023-05-12-how-ultimate-evolved-from-hundreds-of-supervised-models-to-ultimategpt/) · _The Edge_

Question 39

Why do AI copilots sometimes fail to deliver value in enterprises?

Accepted Answer

_Even advanced AI copilots can fail without proper user training._

Mark Benioff from Salesforce joked that Microsoft's copilots have been hit and miss, comparing them to Clippy, the failed desktop assistant. The challenge is it's not easy for enterprises to make AI useful. In my course, I was the first to completely integrate OpenAI and replace myself as a teacher, giving students AI tools to handle complicated coding without needing to code. What I saw is that students initially struggled with this new colleague. Over time, once we trained them on how to use it, they learned. So the hit and miss nature comes from the difficulty of integrating AI into existing workflows and user expectations.

— [AI in E-Commerce](https://www.lutzfinger.com/transcript/2024-10-03-ai-in-e-commerce/) · _The Keynote on AI_

Question 40

Why do AI models still need human involvement even with unsupervised learning?

Accepted Answer

_Even advanced models can't escape the need for human judgment._

We need humans because there isn't always a true north for what's correct. Supervised learning works when there's a clear label, like Obama was president. That's a fact. But when you get into perception, like whether a joke is funny or which word to use about Obama's policies, it depends on context and perspective. The same happens with generated images. Art often creates things society didn't expect, so at what point is something random versus amazing? Since there's not a true label for certain things, we constantly need people to give feedback on the model and provide guardrails.

— [AI and the human factor](https://www.lutzfinger.com/transcript/2023-07-07-ai-and-the-human-factor/) · _The Edge_

Question 41

Why do you say the website is dead, and what does that mean for brands?

Accepted Answer

_Websites lose traffic but gain higher intent visitors via agents._

It's hyperbolic. The website is not dead. The newspaper is not dead, the homepage is not dead. It's only reduced in importance. An agent can go look up websites, read those websites, do the research, compare it, create a metric for themselves to say what is the most explosive or most fruitiest and then give me back an answer. Your traffic goes down, but your revenue probably doesn't because you should see more high intent consumers. Consumers that used an agent to be informed and then they come to your website.

— [From Google to Cornell: Rethinking AI Workflows | Lutz Finger | Build AI Podcast 17](https://www.lutzfinger.com/transcript/2026-04-16-from-google-to-cornell-rethinking-ai-workflows-lutz-finger-build-ai-podcast-17/) · _Build AI_

Question 42

Why does Ben say agriculture uses more technology than early stage venture capital?

Accepted Answer

_Early stage VC relies on gut instinct without data validation._

At the early stages, VCs don't tend to use any data or technology. Growth stage is mostly a signal and projections problem, basically science at this point. But early stage venture is still very much a craft, flying blind mostly on gut instinct and what people have learned over time without really even having the datasets to figure out if what they're doing is true or not. VCs are starting to use Claude or ChatGPT but mostly to reproduce whatever they're doing. It's a sustaining innovation. They want Claude to evaluate a company like they've been evaluating them without asking if they're even evaluating them right.

— [The Hidden Patterns, with Ben Orthlieb](https://www.lutzfinger.com/transcript/2026-04-23-the-hidden-patterns-with-ben-orthlieb/) · _The Keynote on AI_

Question 43

Why does Catalog change the user interface depending on the type of request?

Accepted Answer

_Dynamic interfaces that adapt to what you actually need to see._

We maintain about twenty to twenty-five abstract views, not thousands of deterministic flows like traditional enterprise tools. If it's a search, we render search results. If you ask for a summary, we render a summary view. If you ask for revenue from Q2 2023, we render a graph. The LLM dynamically adapts the interface to the use case. This is one level of abstraction higher than traditional UIs, which build very bespoke workflow-level interfaces. Ninety percent of use cases fall within the first five views.

— [Large Language Models: one interface to rule them all?](https://www.lutzfinger.com/transcript/2023-08-15-large-language-models-one-interface-to-rule-them-all/) · _The Edge_

Question 44

Why is workflow integration important for OpenAI's feedback loops?

Accepted Answer

_Integration into workflows helps measure answer quality and drives specialization._

Feedback loops work best if they are integral to an existing workflow. We will see more and more workflows become supercharged with LLMs. Once OpenAI is integrated into those workflows, it's easier to see when and whether a user accepts a change or suggestion. OpenAI will need to invest more efforts into workflows and products to be more than just a generic language interface. Specialization will be the norm. Initially, OpenAI used the fast route by tagging along with Microsoft, with Copilot offering integrations into many of Microsoft's core tools.

— [How OpenAI Can Secure Its Investments - A Playbook That Needs Microsoft](https://www.forbes.com/sites/lutzfinger/2023/08/18/the-potential-moat-of-openaihow-openai-can-secure-its-investments/) · _Forbes_

Question 45

Why will 2025 be the year of enterprise search for AI?

Accepted Answer

_Enterprise search becomes critical as AI agents need company data_

AI needs data to be effective. What we need are tools that can read and summarize company documents, tailored for enterprise use. 2025 will be the year of search, especially enterprise search. Companies that have access to this kind of data, like Microsoft, Amazon, Google, and Salesforce, are major players in the race to supply AI agents. Access to enterprise data is one of the two critical things companies need to become the main supplier of AI agents for tomorrow's workforce.

— [The Race For AI Agents. Who Will Supply Tomorrow's Workforce?](https://www.forbes.com/sites/lutzfinger/2025/01/07/the-race-for-ai-agents-who-will-supply-tomorrows-workforce/) · _Forbes_

Q&A, AI Workflows

Did you actually replace yourself as a teacher with an AI clone?

How are AI agents different from RPA (robotic process automation)?

How can AI support frontline workers who aren't sitting at computers all day?

How did LinkedIn use knowledge graphs to solve the job title problem?

How did OCR technology affect the number of lawyers?

How did Varo use AI to reduce operational costs while improving customer service?

How do chained AI models create better workflows than simple generative AI tools?

How do venture firms use AI in their own operations?

How does accountability work when AI agents are doing tasks for us?

How does Catalog ensure certain workflows always follow the same steps when using LLMs?

How does Catalog use large language models to simplify enterprise software access?

How does Zenith OS reduce costs in neuropsychiatry treatment?

How is AI changing the role of software developers?

What are some examples of how healthcare providers are using LLMs with patients?

What are the main problems with using simple GPT wrappers for enterprise AI applications?

What did you get wrong about multimodal AI interfaces?

What do you mean when you say winners won't be those who use AI, but those who redesign work around it?

What does 'AI as an operating model' mean?

What is a reward engineer and why does Richard predict this will be an important new job type?

What is an AI agent, really?

What is the augmentation thesis for AI in workflows?

What is the BizML framework and why is it needed for AI projects?

What is the most important format for presenting dealership content to AI models?

What is the practical framework for implementing AI agents in enterprise settings?

What kind of AI agents will actually enter the workforce in 2025?

What makes autonomous AI agents different from traditional automation tools like RPA?

What makes the grid so complex and difficult to manage compared to other infrastructure?

What patterns make a workflow 'agentic' rather than just a wrapper around an LLM?

What practical business use cases exist for language model chaining beyond autonomous agents?

What problem did transformers solve for Ultimate AI's multi-language support?

What was Ultimate AI's key insight that made customer onboarding faster than competitors?

What will actually happen with AI in 2025 instead of AGI?

What's the biggest mistake CEOs make when trying to integrate AI into their companies?

Why are AI workflows the real moat?

Why are coding applications seeing the biggest success with AI agents right now?

Why are radically lean teams the future?

Why did many generative AI projects fail to scale in 2024?

Why did Ultimate AI initially focus on helping human agents instead of automating customer conversations directly?

Why do AI copilots sometimes fail to deliver value in enterprises?

Why do AI models still need human involvement even with unsupervised learning?

Why do you say the website is dead, and what does that mean for brands?

Why does Ben say agriculture uses more technology than early stage venture capital?

Why does Catalog change the user interface depending on the type of request?

Why is workflow integration important for OpenAI's feedback loops?

Why will 2025 be the year of enterprise search for AI?