Atrium Academy

GTM Data Architecture: Foundation, Enrichment, & Action

Manny Adelstein

GTM Engineering Internally at Clay

Set up data to flow cleanly across platforms, use custom objects for signals and intent, and build systems that scale

How to Architect the Foundation

Use enrichment to fill in data gaps, score prospects and customers, & build re-enrichment automations to keep data fresh

How to Build the Enrichment Layer

Clean data informs downstream workflows like orchestrating outreach to warm prospects or ranking top accounts to engage

Why Data Quality is Crucial for GTM Success

The Problem: Scaling organizational expertise with generative AI&nbsp;captures the basics but misses the distinctive thinking patterns and cultural fingerprint that differentiate your organization. The Solution: Capture, automate and reimagine your distinctive methodology with agentic systems using frameworks emerging from years of strategic consulting and building AI tools for Fortune 500 companies and think tanks. We provide the strategic insight and hands-on capabilities to integrate generative AI strategically—in a way that preserves your competitive advantage while unlocking new dimensions of interactivity and personalization. Effective AI integration requires new collaboration frameworks between domain experts and technical teams — we teach you how to bridge this gap so expert knowledge translates into reliable, personalized AI outputs. You will master the frameworks we have been applying for half a decade for global consultancies and companies, walking away with a repeatable blueprint for scaling generative AI across the enterprise. Atop the theoretical bedrock, we bring heavy focus on hands-on practice doing unglamorous but essential work for mastering generative tool deployment: data plumbing, iterative evaluation and stress-testing, collaborative prompt design and de-risking roll-outs through phased deployment. This is not just a course. This is a commitment of experts with a decade of combined experience to transform your workflows and multiply your organization's output.- Unlimited one-on-one consultation opportunities during the course- Proprietary chatbots trained on the learning content- A resource hub 3 years in the making- Takeaway: The extent of your learning is bordered only by the time you are willing to commit.

Transform Your Organization's Expertise Into AI-Powered Competitive Advantage

Course syllabus

Synchronous zoom lecture on Tuesday, focusing on the conceptual backbone of AI agent functionality and applications. Recordings of class will be available soon after its completion.

Tuesday

An async walkthrough will guide you in building functional prototypes of AI Agents. You will get to shape the project deliverable. Lecturers will be available by email, during office hours and via one-on-one sessions throughout the week for questions and support.

Weekly projects (async)

Optional office hours where you can ask questions and receive personal guidance on developing your AI applications.

Tuesday & Thursday

Schedule 1:1 coaching time with instructors for individualized help with your hands-on projects, privately thinking through a use-case, or whatever else you need.

1:1 Coaching

Course schedule

Executive Leaders who need to scale organizational capabilities with AI while preserving competitive differentiation

Project/Product Managers leading cross-functional AI integration initiatives bridging domain expertise with technical implementation

Domain Specialists who want to enhance their analytical workflow through custom AI applications without having to code

Who is this course for

What you’ll get out of this course

Master the end-to-end methodology from initial requirements to production launch, including risk assessment, phased implementation, and organizational adoption strategies that minimize disruption while maximizing competitive advantage

Strategic AI Integration Frameworks

Transform unstructured institutional knowledge like meeting transcripts, reports, and backlogs into agentic systems that prevent your expertise from walking out the door with the next turnover

Culture Capture From Unstructured Data

Design and implement human-in-the-loop evaluation systems and testing methodologies that ensure trust, relevance, and reliability

Quality Assurance and Evaluation Frameworks

Navigate technical tradeoffs across compliance, latency, and cost vectors while managing non-technical stakeholders through AI development cycles and scope definition

Strategic Decision-Making and Project Management

Update your awareness with strategic foresight and thorough ecosystem analysis of AI affordances in specific industries, domains and modalities, enabling informed investment decisions and long-term positioning

AI Ecosystem Intelligence and Future Planning

Automate and amplify institutional knowledge through AI agents that embody your cultural DNA

Lev Gorelov

Jeremy Kirshbaum

Cohort 14

Cohort 15

Transform your retrieval from "good enough" to "mission-critical" in weeks, not months Most RAG implementations get stuck in prototype purgatory. They work well for simple cases but fail on complex queries—leading to frustrated users, lost trust, and wasted engineering time. The difference between a prototype and a production-ready system isn't just better technology, it's a fundamentally different mindset. The RAG Implementation RealityWhat you're experiencing right now: ❌ Your RAG demo impressed stakeholders, but real users encounter hallucinations when they need accuracy most❌ Engineers spend countless hours tweaking prompts with minimal improvement❌ Colleagues report finding information manually that your system failed to retrieve❌ You keep making changes but have no way to measure if they're actually helping❌ Every improvement feels like guesswork instead of systematic progress❌ You're unsure which 10% of possible enhancements will deliver 90% of the value What your RAG system could be:With the RAG Flywheel methodology, you'll build a system that: ✅ Retrieves the right information even for complex, ambiguous queries✅ Continuously improves with each user interaction✅ Provides clear metrics to demonstrate ROI to stakeholders✅ Allows your team to make data-driven decisions about improvements✅ Adapts to different content types with specialized capabilities✅ Creates value that compounds over time instead of degrading What Makes This Course DifferentUnlike courses that focus solely on technical implementation, this program gives you the systematic, data-driven approach used by companies to transform prototypes into production systems that deliver real business value: ✅ The Improvement Flywheel: Build synthetic evaluation data that identifies exactly what's failing in your system—even before you have users✅ Fine-tuning Framework: Create custom embedding models with minimal data (as few as 6,000 examples)✅ Feedback Acceleration: Design interfaces that collect 5x more high-quality feedback without annoying users✅ Segmentation System: Analyze user queries to identify which segments need specialized retrievers for 20-40% accuracy gains✅ Multimodal Architecture: Implement specialized indices for different content types (documents, images, tables)✅ Query Routing: Create a unified system that intelligently selects the right retriever for each query The Complete RAG Implementation Framework Week 1: Evaluation SystemsBuild synthetic datasets that pinpoint RAG failures instead of relying on subjective assessments BEFORE: "We need to make the AI better, but we don't know where to start."AFTER: "We know exactly which query types are failing and by how much." Week 2: Fine-tune EmbeddingsCustomize models for 20-40% accuracy gains with minimal examples BEFORE: "Generic embeddings don't understand our domain terminology."AFTER: "Our embedding models understand exactly what 'similar' means in our business context." Week 3: Feedback SystemsDesign interfaces that collect 5x more feedback without annoying users BEFORE: "Users get frustrated waiting for responses and rarely tell us what's wrong."AFTER: "Every interaction provides signals that strengthen our system." Week 4: Query SegmentationIdentify high-impact improvements and prioritize engineering resources BEFORE: "We don't know which features would deliver the most value."AFTER: "We have a clear roadmap based on actual usage patterns and economic impact." Week 5: Specialized SearchBuild specialized indices for different content types that improve retrieval BEFORE: "Our system struggles with anything beyond basic text documents."AFTER: "We can retrieve information from tables, images, and complex documents with high precision." Week 6: Query RoutingImplement intelligent routing that selects optimal retrievers automatically BEFORE: "Different content requires different interfaces, creating a fragmented experience."AFTER: "Users have a seamless experience while the system intelligently routes to specialized components." Real-world Impact From Implementation﻿✅ 85% blueprint image recall: Construction company using visual LLM captioning✅ 90% research report retrieval: Through better text preprocessing techniques✅ $50M revenue increase: Retail company enhancing product search with embedding fine-tuning✅ +14% accuracy boost: Fine-tuning cross-encoders with minimal examples✅ +20% response accuracy: Using re-ranking techniques✅ -30% irrelevant documents: Through improved query segmentation Join 400+ engineers who've transformed their RAG systems with this methodology Your Instructor Jason Liu has built AI systems across diverse domains—from computer vision at the University of Waterloo to content policy at Facebook to recommendation systems at Stitch Fix that boosted revenue by $50 million. His background in managing large-scale data curation, designing multimodal retrieval models, and processing hundreds of millions of recommendations weekly has directly informed his consulting work with companies implementing RAG systems.

Stop building RAG systems that impress in demos but disappoint in production

A product leader, engineer, or data scientist looking to move beyond ad-hoc RAG prototypes into scalable, production-grade AI solutions.

A professional who understands LLM basics but wants a repeatable, data-driven methodology to improve retrieval relevance, latency, and user

Eager to create feedback loops that continuously refine and enhance the quality of RAG applications as models, data, and user needs evolve.

This Course Is For You If You Are

By the end of this course, participants will be able to

Implement the Data and Evals Flywheel approach to continuously develop and improve RAG applications—breaking free from guesswork and relying on measurable, iterative enhancements.

Adopt a Systematic, Data-First Methodology

Quickly assess your retrieval systems using precision and recall metrics, identify bottlenecks, and confidently validate changes without sinking into endless trial-and-error cycles.

Run Fast, Unit-Test-Like Evaluations

Generate and utilize synthetic data sets to speed up experimentation, enabling you to test new approaches, embeddings, and architectures before committing full resources.

Leverage Synthetic Data for Rapid Iteration

Apply fine-tuning strategies for embedding models to boost search relevance and explore hard negative examples to further sharpen retrieval performance.

Master Fine-Tuning & Hard Negative Mining

Use query classification and segmentation techniques to pinpoint exactly where your RAG system falls short—whether it’s due to limited inventory or insufficient capabilities.

Classify Queries & Identify Bottlenecks

Create tailored indices for documents, images, tables, SQL databases, and more. Learn when and how to fuse or layer these indices to handle complex retrieval tasks elegantly.

Design Specialized Indices for Multiple Modalities

Implement synthetic text chunk generation and strategic summarization methods to improve retrieval results, ensuring end-users get clear, concise, and contextually rich answers.

Enhance Retrieval with Summarization & Chunking

Develop systems that intelligently route queries to the right indices, tools, or pipelines. Blend and fuse indices effectively to handle nuanced, multi-step queries.

Implement Query Routing & Index Fusion

Evaluate the performance of your routing logic and each individual index separately. Gain the nuance to fine-tune global system performance and local retrieval accuracy in tandem.

Optimize Both Global & Local Performance

Design explicit and implicit feedback mechanisms—capturing user signals, automating re-labeling, and applying improvements in real-time to keep your RAG systems on an upward trajectory.

Integrate Feedback Loops for Continuous Improvement

Follow a repeatable process to continually evaluate and improve your RAG application

Jason Liu

Cohort 3

Cohort 4

“Thats not what I wanted”“But thats what you asked for!” If this sounds familiar, then chances are you or your team may be struggling with some or all of the following: 🚩 Your stakeholders and your data team seem to be speaking VERY different languages.&nbsp; 🚩 Gathering data requirements takes too long.&nbsp; 🚩 The process of gathering data requirements is not fun for you or your stakeholders.&nbsp; 🚩 Your stakeholders needs and your data requirements are full of ambiguity (no matter how hard you try).&nbsp; 🚩 Your data team cannot work out what to build from the data requirements you capture. 🚩 Your data requirements are full of the what, but not the why.&nbsp; 🚩 You capture screeds of detailed data requirements that are never used.&nbsp; What if I told you, that in 1 x 5-hour masterclass, we'd be able to help you transform the way you operate, so that you: ✨ Become the data universal translator 🖖🏽 in your organisation, consistently applying a shared language that everyone understands, regardless of their role or technical fluency. ✨ Capture stakeholder needs and data requirements clearly, concisely and in a way that garners instant stakeholder feedback through awesome templates. ✨ Empower data teams with everything that they need to build; without drowning them in detail that that distracts them into oblivion. ✨ Create trust within your stakeholders as someone who can capture data requirements that enable critical business decisions. People know that you're excellent at identifying what's important and being able to figure out what is just noise. In this masterclass you'll learn learn how to solve these problems by using the Information Product Canvas pattern to gather data requirements in 30 minutes or less, using a shared language. Enrol now and get ready to make some magic. ✨ ___________________________________ You can see the cohort times for your timezone in the big green section titled "Join an upcoming cohort" below. There are cohorts scheduled each month for: UK / EUBST - 10 am - 3 pm (UK)CST - 11 am - 4 pm (EU) USPDT - 12 pm - 5 pm (US)EDT - 3 pm - 8 pm (US) NZ / AUNZT - 10 am - 3 pm (NZ)AST - 8 am - 1 pm (AU) ___________________________________ Or use this url: https://maven.com/agiledata/info-prod?selecting=true

How many times have you heard this.... 

This is a live course. Over 50% of the course is hands on team work going through a case study to "learn by doing".

You need to attend the live session in person to participate

You need to attend the live session in person to participate. Access to the content is available on the companion website after the course.

The course is not recorded

Any automated AI transcription bot will not be accepted into the session.

No "AI Bots"

Business Analysts who are gathering data requirements on behalf of their data teams

Data Product Managerswho need to define and prioritise the data work coming up

Data Analysts who want to agree upfront the outcome of the work they are about to do for their stakeholder

Data Engineers who are sick of getting data requirements they can’t build from

Data Leaders who want to try the pattern themselves before recommending it to their data teams

Agile Coaches who want to add a new pattern to their coaching toolkit

Designed for people in the data domain

What you’ll learn from this course

Which your&nbsp;data team&nbsp;can&nbsp;deliver&nbsp;in&nbsp;days&nbsp;or weeks.

The literal blueprint for capturing data requirements from your stakeholders in 30 minutes

No more trying to convert “I want to understand the total value of orders last year” into “Dim Order, Dim Time, Daily Order Value Fact” in your head or writing two different requirements docs.

Learn to talk a shared data requirements language that both your stakeholders and your data teams can understand

Learn how to fill out the Information Product Canvas&nbsp;in front&nbsp;of your&nbsp;stakeholders, so they can&nbsp;validate&nbsp;their&nbsp;requirements&nbsp;as you capture them. You'll never need to send a document that sits in their inbox and only gets reviewed at the last minute (or not at all) ever again.

Visually collaborate with your stakeholders, validating their requirements as you capture them

Create data requirements&nbsp;stakeholders&nbsp;can&nbsp;understand&nbsp;at a&nbsp;glance&nbsp;and use to&nbsp;compare&nbsp;the different&nbsp;investment&nbsp;options, before they&nbsp;prioritise&nbsp;them. Stronger, transparent evaluation criteria = data-based prioritisation decisions.

Help prioritise what's important - in a way that stakeholders can understand and evaluate at a glance 

Learn how to elicit and capture the&nbsp;actions&nbsp;and&nbsp;outcomes&nbsp;the&nbsp;stakeholders&nbsp;want to&nbsp;achieve&nbsp;with the data they are requesting. Move from being a data "order taker" to becoming a trusted analyst, data product owner or data leader who can help achieve meaningful business outcomes.

Capture both the what and the why

Don’t continue to be the bottleneck in the process, or continuously working months in advance of your team so you aren’t. Working together to solve problems is so much more fun!

Learn a pattern you can teach the rest of your data team, so they can do the data work with you.
 

Learn a&nbsp;great way&nbsp;to gather data requirements by&nbsp;defining&nbsp;them in&nbsp;small chunks, then&nbsp;delivering&nbsp;the&nbsp;Information&nbsp;Products to your&nbsp;consumers&nbsp;as you go. Rinse and repeat!﻿No more putting in the hard effort upfront and never getting to see the results.

Feel the rush of delivering value often and early

Stop boiling the ocean, capture requirements in a repeatable way, so your stakeholders love and data teams can build from

Shane Gibson 

Agile Data Coach | Author of Agile Data Guides | Podcast Host | Co-Founder

This cohort is suitable for people in UK and EU timezones

UK / EU August 29th  (10am - 3pm BST / 11am - 4pm CST)

This cohort is suitable for people in US timezones

US September 12th (12pm - 5pm PDT / 3pm - 8pm EDT)

The ChallengeYour days are filled with:🗓️ back-to-back meetings,🚀 sky-high customer or investor expectations,⏱️ wishes that your team could move faster sooner,🔥 wishes that you could move even faster and work on what really matters instead of fighting fires. You start each day with the best intentions, but somehow find yourself constantly reacting instead of leading.&nbsp;Sure, it would be nice to focus on your own goals, but you care about those other people (or at least some of them 😁). They’re colleagues and customers, family and friends... How can you collaborate with them responsively without forsaking your goals and becoming just a (re-)Actor in their story? 🎬 ✨ You wonder if AI can help, you wonder how other leaders are using AI to manage it all and manage their teams--but honestly, who's got the time to figure that out? ⏳ You have a constant, lurking anxiety about what’s falling through the cracks or not getting done, so you work faster. As you try to move faster, people fall short of your expectations and their potential. It causes friction and frustration (even if you hide it well). It's uncomfortable to call out, so you avoid giving feedback or mention it lightly, which makes the situation more uncomfortable. Eventually, you decide to give feedback. Sometimes it goes okay. Often, you try to move through and out of the conversation as quickly as possible because it feels so awkward. Growth and improvements rarely last. You take on more work yourself to get it done. The Solution (Evidence-based. Immediately applicable.)⏱️&nbsp;EVERYTHING YOU READ BELOW WILL BE ORIENTED TO IMMEDIATE APPLICATION DURING THE COURSE ⏱️What if you could shift from reacting to the constant inputs to creating what matters most and leading your team in that direction--everyday? This course will support you, step by step, in building your Operating System to make this shift. Fancy apps aren't enough. Nor are bursts of motivation. A sustainable Operating System has 3 features:🧰 Tools (simple enough to keep using, and amplified by AI)🔁&nbsp;Habits (the tools don't use themselves...except for some of the AI ones we'll build)🧠 Mindsets (you can "productively" spend your life responding to other people's messages and requests--or build something that matters) No, "perfection" isn't possible (or even desirable). But our days--and our results--completely shift when we focus our attention systematically instead of ad-hoc and reactively. And our conversations shift when we stop trying to fix people and start inspiring them to take action and grow. We teach you how to do this by using straightforward approaches that reliably transform defensiveness and friction into self-reflection, connection, and motivation to grow. You will start to look forward to feedback conversations because you will know how to use them as powerful tools for building trust and deepening connection. You will also explore your own mindsets and beliefs around feedback, identifying your own patterns around conflict avoidance, people-pleasing, and controlling behaviors. AI is woven throughout the course, including:1️⃣ building your own Custom GPTs2️⃣&nbsp;mastering your calendar3️⃣ using AI as a practice partner for difficult conversations and coaching skills4️⃣ helping your team use AI, and so much more! Once you've shifted the way you manage yourself and lead your team conversations, applying evidence-based approaches tested with 1000s of managers and amplified with AI, the whole game becomes fun again. Bonus:📝 You will also get 5 step-by-step facilitation guides for sharing and teaching your team members everything we cover in this course. This helps establish shared language and frameworks within your team/organization, and positions you as the leader of change. 📖&nbsp;You'll also get our 20-page Offsite Playbook with step-by-step guidance on how to design and facilitate unforgettable team and company gatherings. 🚀 And you'll get to do this while in a practice lab with other high-growth leaders.

Spend your days fighting fires or each other?

Sundays (+ 1 Thursday)

Weekly Office Hours

High-Growth ManagersReady to move out of IC mode, you want to be the inspiring leader people thank for the rest of their careers.

ExecutivesYou know there's a better way than back-to-back meetings everyday. To scale your impact, you know you must scale yourself first

Startup FoundersYou're expanding your focus from building a product to building a team. And to do so, you need the team to step up fast.

You'll have the Inner and Outer Game in place:<ul><li>A systematic set of habits, tools, and mindsets to define What Matters Most each day and week</li><li>You'll be using your emotions as signals for what needs to happen next</li><li>Your mind will be at ease</li></ul>

Clear Focus & Emotional Balance

Whether you're currently intimidated by AI, curious, or all-in, you will:<ul><li>know exactly how to leverage Claude and ChatGPT to improve your skills and help you prepare for key conversations</li><li>have your own CustomGPT to do a recurring management task FOR you</li></ul>

AI As Your Ally (including your own CustomGPT)

<ul><li>Next Actions template (simplest tool possible to stay organized)</li><li>Weekly Preview Checklist (define what matters most and who should do it)</li><li>Daily Defining Checklist</li></ul>

The Operating System (feel on top of your s***)

<ul><li>Know how to decide WHAT to delegate to increase your leverage</li><li>Know HOW to delegate (to a person or AI) in ways that build confidence and follow-through</li></ul>

The Leverage Framework

<ul><li>GAIN Template for how to structure feedback conversations that inspire response-ability, action, and growth</li><li>how to use AI to prepare and practice</li></ul>

GAIN Feedback Framework

<ul><li>GROW Template for coaching conversations</li><li>The key skill to coach effectively</li><li>Know the difference between a feedback and a coaching conversation</li><li>how to use AI to practice and get feedback on a coaching conversation</li></ul>

GROW Coaching Framework

What if you felt on top of everything and your team were aligning and growing through every conversation with you?

Lightning Lessons

GTM Data Architecture: Foundation, Enrichment, & Action