Steven Jordan

Everybody wants to do RAG... ...but nobody wants to give the RAG Triad any love. Well...that changes today, mostly, I think. We talked about tool call accuracy for Agentic workflows last week and how it can help diagnose issues in agentic workflows. Well this week I wanted to take a look at how we might apply this to an Agentic RAG flow. In this vid you'll learn a little about: 1. Evaluating RAG workflows using the RAG Triad 2. How to do it using the Agentic Evaluator 3. Integrating it with Langgraph for in pipeline evaluation Let me know what you think? Should we dig deeper into why my responses were a little sketchy? Bring in the LLM Judge?

507

15 kommentarer

my new favorite flow as product owner of multiple ventures is basically an always-on insight machine. saving hundreds of hours (and details). here’s how it works: every week, each venture (currently 7) gets at least two real user interviews. on top of that, we pull in synthetic user feedback to simulate hundreds more voices. every conversation is recorded, transcribed, and ready to be processed. then the magic happens: → the transcripts and synthetic feedback go straight into a product owner agent powered by claude mcp → the agent extracts key insights, clusters feedback, and identifies unmet needs → it writes feature requests, user stories, and even pulls direct user quotes for context → every piece of work is translated into dev-ready kanban cards → we also have memory from there, claude mcp integrates with notion and automatically populates the weekly sprint boards for each venture. the only part still 100% human? the conversations themselves + double-checking cards. i spend my time talking to real people, building relationships, and keeping a pulse on our community. the rest= summarizing, prioritizing, and formatting...is fully automated. before this setup, product ownership was hours of manual backlog grooming, reformatting notes, and chasing clarity between meetings. now? it’s instant, consistent, and scalable across ventures. this is the future of product work: → humans for connection, trust, and judgment → agents for synthesis, organization, and execution tools in this stack: - syntheticusers → synthetic + real feedback at scale - granola → instant meeting transcripts - claude mcp → product owner automation - rag → create memory and consistency - notion → sprint + kanban management product ownership is becoming less about moving cards around, and more about steering the ship with the clearest, freshest signals possible. exciting times...

123

46 kommentarer

Today, we’ve released our biggest improvement to the way data teams work with the launch of Secoda AI Agents. Secoda AI Agents make data automation simple and are now available in all Secoda workspaces. Instead of doing routine tasks by hand, you just trigger agents in plain language. The Agent takes care of the task and brings back the results. That means data teams can automate jobs like data tagging, documentation, monitoring, or creating glossary terms. Business teams can set up reports to run on a cadence and run Analysis agents to report on the business on a continuous basis. With Secoda AI Agents, you spend less time on manual work and more time using your data. #secoda #aiadoption #aiagents #enterpriseai

55

5 kommentarer

AI agents are completely reshaping what’s possible in SaaS. We’ve captured what we’re learning in a new whitepaper. There’s a ton of great info in there, but if I can leave you with one thing: start small, but start now. The space is moving incredibly fast, and waiting for perfection is riskier than thoughtful experimentation. If you’re not leaning in, you’re going to get left behind. Want to dig deeper? Check out our detailed guidance here: https://lnkd.in/gxhKvCyQ

498

8 kommentarer

More data is not the answer. More good data is. A model underperforms? "We need more data!" Analytics seem off? "Let’s ingest more sources!" Executives demand better predictions? "Just get more data!" No. Instead ask for : ➤ Data Contracts to define clear expectations for schema, freshness, and accuracy before data even enters your pipelines. ➤ Observability & Monitoring so we can track data drift, schema changes, and anomalies in real-time to prevent silent failures. Invest in processes that prioritise quality over volume. How is your organization ensuring data quality before data quantity?

99

5 kommentarer

⚡️ Real-time analytics isn’t just for dashboards anymore. It powers customer experiences like: * Profile views in LinkedIn * Live ETAs in Uber Eats * Merchant cash-flow in Stripe But legacy stacks (OLTP + Flink + Redis / KV stores …) are brittle, expensive, and hard to evolve: ❌ Data bloat from pre-aggregations ❌ Complex ETL pipelines ❌ Painful schema/metric changes ❌ Skyrocketing infra costs In this new blog post: https://lnkd.in/gR6-SxeA we describe how Apache Pinot™ + StarTree Cloud collapse the stack. Companies like LinkedIn, Uber and Stripe have significantly slashed their infra costs by adopting Apache Pinot. The capability to do high concurrency, low latency queries enable such companies to consolidate realtime analytics and serving insights in the same platform.

44

2 kommentarer

So many companies say they’re 𝗱𝗮𝘁𝗮-𝗱𝗿𝗶𝘃𝗲𝗻. Yet most are just 𝗱𝗮𝘁𝗮-𝗱𝗿𝗼𝘄𝗻𝗲𝗱. They invest in tools, dashboards, and AI models… But still struggle to make decisions faster or better. Because data without clarity is noise. And consultants, the good ones, are trained to turn that noise into narrative. When I was at BCG, I once asked a leadership team to pick their three most important KPIs. Nobody spoke for 30 seconds. Here’s how the best consultants turn data into clarity: 𝟭. 𝗦𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲 𝗯𝗲𝗳𝗼𝗿𝗲 𝘃𝗼𝗹𝘂𝗺𝗲 Don’t start with “what data do we have?” Start with “What decisions do we need to make?” Then reverse-engineer what data actually supports those. 𝟮. 𝗩𝗶𝘀𝘂𝗮𝗹𝗶𝘇𝗲, 𝘁𝗵𝗲𝗻 𝘃𝗲𝗿𝗯𝗮𝗹𝗶𝘇𝗲 A chart helps people see. A story helps them care. Pair visuals with a narrative → cause, effect, implication. Help your audience feel the insight, not just read it. 𝟯. 𝗨𝘀𝗲 𝗔𝗜 𝗮𝘀 𝗮 𝗰𝗹𝗮𝗿𝗶𝘁𝘆 𝗮𝗺𝗽𝗹𝗶𝗳𝗶𝗲𝗿, 𝗻𝗼𝘁 𝗮 𝗰𝗿𝘂𝘁𝗰𝗵 AI can clean, cluster, and summarize your data. But only humans can frame meaning. Know how to explain the story behind the numbers. The best data presentations don’t look like dashboards. They look like stories that make executives nod and act. Be the consultant in the room, even if it’s not your title. What’s one technique you use to make data understandable? Great visual by Leonardo Freixas — 📌 Save this for later. ♻️ Share this with your analytics team, they’ll get it instantly ;) 🔔 Follow Till Schmid for more on leadership, AI, and consulting.

391

21 kommentarer

camelAI's most popular feature is Chat with CSVs. Upload dozens and ask questions across all of them at once. Plot twist: After talking to 300+ executives, I've learned that despite having every modern BI tool imaginable, most companies still run their analytics on... CSVs and Excel files. I have a feeling this doesn't surprise you either. 📊 I recorded a ~really~ long 14 minute video explaining: - Chat with Postgres setup - Chat with CSV setup - Federated Queries - Fine Tuning Camel's Multi-Agent Data Analyst - Knowledge Base - Reference Queries - Thread Grader - Dynamic Memory Check out the youtube video for the full details!

42

6 kommentarer