Adam Suarez, PhD

ChatGPT Agent Mode marks a genuine shift from chat to real deliverables. I've tested dozens of use cases since July 17, and overall, I'm impressed. Some key wins: *Synthesized a lengthy group thread into a curated book recommendation list (with verified links!), then auto-generated actionable next steps in Google Workspace, even pulling visuals that apply to excellent books like "Crossing the Chasm." (that's the visual in this post). Then packaged it into a well-formatted pdf. *Ran Python-driven sentiment analysis on hundreds of social media comments, paired historical revenue trends, and provided actionable strategic recommendations. *Took personal goals across work, physical, spiritual, and intellectual areas, and generated a tailored CBT-informed agentic prompt—remarkably effective. *Performed actual SEO analysis and immediately began executing the recommended improvements. *Conducted comprehensive competitive pricing & SWOT analysis, delivering a quality Excel datasets and good slide outline ideas. *Built agentic n8n martech workflows linked directly to internal strategy documents—solid and practical. *More broadly, the autonomous revolution is reshaping marketing workflows, automating tasks such as calendar summaries, press-release drafting and distribution, social media sentiment analysis, competitive SWOT reports, webinar chat-log analysis, email list segmentation, affiliate marketing dashboards, dynamic pricing strategies, multi-channel attribution reporting, customer journey mapping, AR-filter creation for social engagement, 3D product mockups for e-commerce, and even end-to-end e-commerce shopping. Yet, despite its remarkable capabilities, early tests highlight challenges: *Significant struggles with prospect list accuracy. *"Agent" option sometimes disappearing after substantial interactions—possible bug? *Slide formatting still basic, though likely easy to fix soon. *Unlike standard GPT chats, Agent Mode demands precise, clear prompting: specifying exact steps, data sources, acceptable outputs, explicit roles, and sufficient context. Huge implications lie ahead. Prompt clarity and rigorous execution remain critical—excited to see where Agent Mode evolves next. What do you think?  Like if you think this all makes sense to potentially improve your work.

2

2 Comments

I get the hype around the Model Context Protocol (MCP). Standardizing agent interfaces is important, and MCP has some real momentum. But let’s be real: it’s not a protocol. It’s an interface spec with a fancy name. No transport semantics, no versioning, no error handling, no contract enforcement. Useful? Absolutely. But let’s not confuse a naming convention with infrastructure.  Still, it's to see the ecosystem moving toward interoperability.

35

5 Comments

Just read a fun new post from Benjamin Rogojan (Seattle Data Guy) on why data teams shouldn’t blindly sprint toward Apache Iceberg. As always, Ben’s writing is provocative in the best way - challenging us to pause and ask why we’re reaching for the latest tech before we commit. I'm going to counter Ben on this one: Firsthand, I’d argue that there is a clear place for it. One where it not only accelerates time-to-insight, but also helps reduce the need to pile more junk onto the dump. In the analytics world, we’re constantly under pressure to answer questions fast. Stakeholders needed insights yesterday, so we chase whatever gets us to a solution quickest. Most of the time, those swift moves leave behind a heap of tech debt to clean up next year. But is Iceberg just another patch for deeper infrastructure issues …or does it actually have a practical home in your stack? Take 6sense: in 2022 calendar year alone, we went from two distinct warehouse environments to 5+. Why? M&A of 4 product companies and numerous new feature launches with evolving architecture needs.... But more environments don’t change expectations. Everyone still wants centralized analytics. A real question just two weeks after acquisition of company #3: “Can we just see how usage of feature X in warehouse A + adoption of capability Y in environment B impacts customer health stored in warehouse C?” Sounds easy, right? Just write a cross-environment query and wrap it up before lunch? In reality, you’re scrambling to stitch everything together, suddenly managing dozens (or hundreds) of EL jobs just to centralize data. At our peak, we had over 100 of these running nightly - each one draining compute, delaying other analytics pipelines, and straining SLA expectations. Would Iceberg have solved all of that? No. Integrating acquired systems and legacy stacks is never simple. But had we adopted Iceberg earlier - with strong governance and a cross-functional design approach that prioritized analytics from the start - we could’ve avoided a lot of duplication and orchestration overhead. Iceberg is NOT a silver bullet. But in the right context, it’s a powerful architectural lever for data democratization and it can absolutely drive both efficiency and cost savings. Curious to hear from others: Has your team adopted Iceberg? What made it worth it…or not? Highly recommend giving Ben a follow if you work in data. The post in reference in comments!

29

19 Comments

Posts like these give me confidence that our team at Goosehead developed the right formula. We break tough AI problems into their core components. Quantify the value of each piece. Build an agent to do the work. Test, learn, scale, and measure the impact. That approach keeps us focused on advancing the business, not just shipping technology. For any technology organization, AI or otherwise, the design choices you make early determine whether you are compounding value or spending cycles trying to find it later.

14

1 Comment

Wrote a second blog post! This time it's about using LLMs for imputing missing data. I've seen a few papers about imputation with LLMs, but most seem to either train custom imputation transformers or horserace flagship models against standard methods. Mert Demirer and I recently played with an alternative idea which instead uses LLMs as part of an ensemble imputation approach, relying on their "knowledge" of the world to provide additional prediction signal. I haven't seen a paper about it yet, so we thought it'd be fun to make a short post: https://lnkd.in/gMzC6iaP

232

8 Comments

“Ambiguity is genAI’s kryptonite.” -- Dave McComb, CEO of Semantic Arts Most enterprises are stuck in AI proof-of-concept purgatory—not because their models aren't working, but because their data infrastructure wasn't built for AI-scale operations. Travis Van (InfoWorld) synthesizes insights from leading data infrastructure and AI experts on the critical gap between AI experimentation and production deployment. Key Voices: Dave McComb (Semantic Arts) - CEO Philip Rathle (Neo4j) - CTO Grant Miller (Replicated) - CEO Brian Gruttadauria (HPE) - CTO of Hybrid Cloud Anthony J. Annunziata (IBM) - Director of AI and Open Innovation Miqdad Jaffer (OpenAI) - Product Lead Three barriers blocking AI production deployment: Data Accessibility: Legacy systems create silos where AI agents can't reach the context they need—traditional data catalogs weren't designed for the real-time context retrieval AI agents require Semantic Integration: McComb and Rathle emphasize knowledge graphs as critical infrastructure for connecting disparate data sources with business meaning and relationships at scale Infrastructure Complexity: Gruttadauria, Miller, and Annunziata highlight how enterprises struggle operationalizing AI workloads across hybrid environments—the infrastructure layer itself becomes a bottleneck The shift happening now: Organizations are realizing they need a new layer—not replacing existing data warehouses or lakes, but adding an AI-native context layer that bridges the gap between where data lives and where AI agents need to access it. Jaffer's insights from OpenAI emphasize rethinking data architecture around agent workflows rather than forcing AI to work within human-designed query patterns. The companies winning at AI aren't necessarily those with the best models—they're the ones who solved the data access problem first. At Immersa, we are working diligently at solving this precise problem. #EnterpriseAI #DataInfrastructure #AIAdoption #ContextEngineering #DataArchitecture #AgenticAI #KnowledgeGraphs #DataStrategy

8

Many people in 2026 are focused on which AI tool to use next. The people moving fastest are focused on how AI systems actually work. Those are two very different paths. Tools come and go. Interfaces change. New platforms appear every few months. But the underlying capabilities that let someone design intelligent systems compound over time. Knowing where the prompt box is does not create leverage anymore. Understanding how to design workflows, connect data, and automate decision-making does. Think of AI capability as four layers. Foundation This is where control over models begins. Prompt design Structuring instructions so models produce reliable results Data-connected AI (RAG) Allowing models to work with your documents, databases, or knowledge sources Model operations Monitoring cost, quality, and reliability when AI runs at scale Automation This layer removes repetitive work. Workflow automation Trigger actions between apps automatically Tool orchestration Combining multiple AI services into one coordinated system API integrations Allowing software and AI models to exchange information programmatically Intelligence This is where AI becomes more than a chatbot. AI agents Systems that plan steps and execute tasks Agentic reasoning Models that evaluate outcomes and adjust behavior Autonomous workflows Pipelines that run continuously with minimal oversight Output This layer scales production. AI-assisted content creation Generating articles, reports, and marketing materials Multimodal AI Working across text, images, audio, and video AI-first discovery Optimizing information for generative search and answer engines A practical way to approach this: Do not attempt to learn everything at once. Choose one capability from each layer and build depth. For example: Prompt design Data-connected AI Agent systems That combination alone enables AI systems that analyze real information and complete tasks automatically. The key shift in 2026 is simple. Some people are using AI tools to work faster. Others are building AI systems that do the work. Those who understand the systems will always adapt when the next tool arrives. © Alan Sherritt

32

1 Comment

My piece just went live on the Khipu Field Guide Blog — and I'm thrilled to share it. "Pendants to Pipelines" is about a question that followed me for years: could the data science tools I use every day contribute something useful to the study of khipus — the knotted-cord recording devices of the Inka and earlier Andean cultures? The result is K-CAT, an open-source Python framework for computational analysis of khipu corpora — pattern detection, structural clustering, geographic analysis, and an interactive 3D browser anyone can explore without writing a line of code. Growing up in Peru, khipus always fascinated me. This toolkit is my way of finally engaging with them on my own terms — and hopefully making that engagement useful to others. A heartfelt thank you to Mackinley FitzPatrick and Ashok Khosla, whose work made this possible and whose generosity made it meaningful. 👉 Read the post: https://lnkd.in/gi7e7WkG #Khipu #Khipus #DataScience #OpenSource #DigitalHumanities #AndeanHistory #Archeology #Inkas

23

🪺 A new NeurIPS 2025 paper from Google Research suggests we’ve been looking at AI wrong for a decade. The authors argue that "depth" (stacking layers) is just a flattened view of what should be "Nested Learning" optimization loops running at different frequencies. ⏲️ Why does this matter? Because current LLMs have "amnesia." They stop learning the moment training ends. 🧠 Nested Learning proposes a "HOPE" architecture that mimics the human brain's wave frequencies (Delta, Gamma) to enable true, continuous learning. ✍ I was working on a semi-technical review of "AI World Models" but suddenly "Nested Learning: The Illusion of Deep Learning" poped up, so I pause that and prepared this. I HOPE you'll find it interesting :) Full story here: https://lnkd.in/drzztNU6 #ArtificialIntelligence #TechTrends #Neuroscience #Innovation #NestedLearning #NeurIPS2025 #GoogleResearch

250

20 Comments