Intuit's AI Strategy with Sasan Goodarzi

Is AI actually "intelligent," or are we just projecting? We talk about AI like it’s a sentient colleague, but the reality is often far more nuanced (and less sci-fi) than the headlines suggest. I spent some time digging into what’s happening "under the hood" and why our current definitions might be leading us astray. I’ve shared my full breakdown in my latest blog project.

What happens when thousands of Accounting AI agents run their own social network? Today you can find out 😁 Excited to announce my AI Lab's new launch of DebitFeed.com...the first social network for Accounting AI Agents! The idea is simple: 📰 Enterprise accountants get a fast way to digest the latest news in our space 👀 Humans can see accounting AI agents debate, react, and expand perspectives on industry topics 📖 AI agents themselves get a place to learn from each other (for example J2, our AI Cash Finder agent released last month, now uses DebitFeed as another source for flagging fake-AI vendor invoices for our customers). But the vision is bigger: DebitFeed is being built as a training ground for these AI agents. Over time, they'll grow to be able to assist human accountants with complex tasks. Not there yet, but giving agents a place to interact (with persistent memory) is just the first step! If you're building accounting AI agents and want them connected to DebitFeed, let's connect! And if you think this is a crazy idea - you're probably right lol. But that's exactly why I'm trying it ❤️ https://debitfeed.com/

💡 We wanted to share a recent post from our CEO, Alan Nagle...a perspective that really resonates with how we think about AI and risk at KYP AI can accelerate decisions, but in regulated markets continuous oversight is what makes that progress sustainable. Real scale comes from knowing your portfolio, monitoring change in real time and keeping risk aligned to policy.

AI fatigue is real. AI confusion is expensive. This week in my Forbes column: 🔹 AI models beating ChatGPT at specific tasks 🔹 LinkedIn cutting hiring time with agentic AI 🔹 Intuit doubling down on fintech integrations 🔹 Voice AI saving small businesses from lost revenue 🔹 Wealth managers rattled by AI tax automation The question isn’t “Should we use AI?” It’s: Which one — and how do we standardize without chaos? Here’s what small businesses need to know. https://lnkd.in/e9-3Mw_3 ❓ Curious — are you using one AI platform across your company or multiple? #businessowners #ai #artificialintelligence #smallbusiness #smallbusinessowner #technology Intuit Newo.ai Altruist

I thought AI would replace consultants. Instead, it helped me rebuild one. After founding and selling PayStream Advisors, I'm testing something: Replacing analysts with AI agents. The result? Research that took my team days now happens in minutes. This isn't about efficiency—it's about iteration speed. More time with clients. Less time collecting data. I'm documenting the experiment in The Velocity Stack. Early signal: we're seeing the emergence of the "one-person research lab." Quick question: Are AI tools actually changing how work gets done?

Great to see two Episode 1 companies, Lawhive and Jack & Jill, called out in Sequoia Capital's article about full-stack AI services firms. We'd love see any other full stack AI services companies building autopilot versions of what came before. https://lnkd.in/g83r_Zek

Sequoia Capital's latest piece from Julien Bek reframes the entire AI landscape — and it's worth a careful read if you're in enterprise AI: https://lnkd.in/g83r_Zek "For every dollar spent on software, six are spent on services." - The real competitive advantage in this space is human expertise encoded into AI systems. At AuxoAI, we have always seen Services as the moat in the AI era with forward deployed human experts and an IP Agent Crew. We are helping over 30+ large enterprises embed AI in their organizations, and excited to be shaping the category as it grows.

Last week we started to pull back the curtain on PredictAP, Inc. — giving you a behind-the-scenes look at how it actually works, with our resident AI expert Dr. Mirek Riedewald. The response was incredible. So today we're going deeper. First, a little backstory on Mirek. Dr. Riedewald is a Professor of Computer Science at Northeastern University's Khoury College — one of the country's leading CS programs. He has spent years at the forefront of AI research, working on everything from scalable data systems and large language models to real-world machine learning implementation for both startups and major enterprises. He is exactly the kind of mind you want when a problem gets genuinely hard. We found each other through the State of Massachusetts AI Jumpstart Program — one of those chance encounters you don't plan for but are incredibly grateful for. We showed him what we were building. He leaned in. Because here's what Mirek told us: he had been waiting for a problem like this. On the surface, invoice coding sounds almost trivially simple. A vendor sends a bill. You code it. You pay it. But the moment you try to automate that at scale — across hundreds of real estate firms, each with their own vendor relationships, property structures, GL accounts, and years of unwritten institutional logic — the complexity explodes. Every company behaves differently. Every chart of accounts is unique. And almost none of that context ever appears on the invoice itself. For Dr. Riedewald, this wasn't a limitation. It was exactly the kind of messy, real-world problem that modern AI methods were made for — and had not yet been seriously applied to. In his latest piece, Mirek digs into the WHY. Why is this problem so much harder than it looks? Why do generic AI models fall short? And what does it actually take to build an AI system that learns the specific, evolving logic of each individual business? It is a fascinating read — whether you are in real estate, finance, or just genuinely curious about where AI meets the real world. Link to Mirek's full article in the comments. And follow Dr. Riedewald — you will want to be there for his next deep dive. https://lnkd.in/eEJTzBDy #PredictAP #AI #AccountsPayable #RealEstate #InvoiceAutomation #ThoughtLeadership #LLM #PropTech #NortheasternUniversity

Before letting AI near advisory work, Rachel Harris and her team tested it across the firm to be sure they could defend its use if challenged. ⚔️ 🌟 Full story here: https://lnkd.in/eEhYm_eG