Uploaded June 2026 | Updated September 2026, 2 weeks ago
What if asking a predictive question was as easy as writing a prompt? In this
hands-on webinar, Andreea Turcu (Head of Global Training at H2O.ai) introduces
tabular foundation models and demonstrates TabH2O live - going from a raw
spreadsheet to predictions in seconds, with no model training, no code, and no
infrastructure.
You'll see the full workflow inside Excel: install the add-in, load a dataset,
and predict missing values at the click of a button. Plus a candid look at when
TabH2O is the right tool — and when traditional machine learning still wins.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🔗 LINKS
→ Try TabH2O (free): tabh2o.h2oai.com
→ Docs & manifest.xml download: tabh2o.h2oai.com/docs
→ Sample dataset (opportunity value): github.com/h2oai/h2o-university-resources/blob/main/demo-datasets/sample_customer_churn.csv
→ Excel Online: https://excel.cloud.microsoft/
→ Free courses & certifications: h2o.ai/university
→ Enterprise limits & use cases: h2oai.com/demo
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
⏱️ CHAPTERS
0:00 Welcome & the prediction question
2:13 Who we are — a decade of predictive AI
4:22 Today's roadmap
5:20 The shift — why business data stayed hard
6:43 Tabular data: the language of business
8:23 Why traditional ML feels heavy
10:01 The prediction gap
11:56 What is a tabular foundation model?
14:48 Meet TabH2O
17:01 How it works — labeled rows in, predictions out
17:32 Your data stays yours (privacy)
18:52 Live demo — setup
34:50 The predict moment
36:15 What just happened — before vs after
39:09 One workflow, many business questions
40:43 When traditional ML is still the better choice
43:28 Beyond the spreadsheet — agentic & actions
46:03 The future — from creator to forecaster
47:42 Q&A
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
ABOUT TABH2O
TabH2O is a tabular foundation model from H2O.ai.
Just as large language models learn patterns across language, tabular foundation models learn patterns across
tables — so you can bring a spreadsheet, provide a few labeled examples, and
generate predictions without building or training a model. It handles
classification and regression, runs inside Excel via an add-in, and works even
on small datasets. Your data is processed in memory — not stored, not cached,
not used to retrain the model.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
ABOUT H2O.AI
H2O.ai has spent more than a decade helping organizations generate value from
predictive AI — with open-source H2O, AutoML, and Driverless AI used by data
scientists worldwide.
#TabH2O #H2Oai #TabularFoundationModels #PredictiveAI #MachineLearning #ExcelAI #NoCodeML
What if asking a predictive question was as easy as writing a prompt? In this
hands-on webinar, Andreea Turcu (Head of Global Training at H2O.ai) introduces
tabular foundation models and demonstrates TabH2O live - going from a raw
spreadsheet to predictions in seconds, with no model training, no code, and no
infrastructure.
You'll see the full workflow inside Excel: install the add-in, load a dataset,
and predict missing values at the click of a button. Plus a candid look at when
TabH2O is the right tool — and when traditional machine learning still wins.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
🔗 LINKS
→ Try TabH2O (free): tabh2o.h2oai.com
→ Docs & manifest.xml download: tabh2o.h2oai.com/docs
→ Sample dataset (opportunity value): github.com/h2oai/h2o-university-resources/blob/main/demo-datasets/sample_customer_churn.csv
→ Excel Online: https://excel.cloud.microsoft/
→ Free courses & certifications: h2o.ai/university
→ Enterprise limits & use cases: h2oai.com/demo
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
⏱️ CHAPTERS
0:00 Welcome & the prediction question
2:13 Who we are — a decade of predictive AI
4:22 Today's roadmap
5:20 The shift — why business data stayed hard
6:43 Tabular data: the language of business
8:23 Why traditional ML feels heavy
10:01 The prediction gap
11:56 What is a tabular foundation model?
14:48 Meet TabH2O
17:01 How it works — labeled rows in, predictions out
17:32 Your data stays yours (privacy)
18:52 Live demo — setup
34:50 The predict moment
36:15 What just happened — before vs after
39:09 One workflow, many business questions
40:43 When traditional ML is still the better choice
43:28 Beyond the spreadsheet — agentic & actions
46:03 The future — from creator to forecaster
47:42 Q&A
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
ABOUT TABH2O
TabH2O is a tabular foundation model from H2O.ai.
Just as large language models learn patterns across language, tabular foundation models learn patterns across
tables — so you can bring a spreadsheet, provide a few labeled examples, and
generate predictions without building or training a model. It handles
classification and regression, runs inside Excel via an add-in, and works even
on small datasets. Your data is processed in memory — not stored, not cached,
not used to retrain the model.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
ABOUT H2O.AI
H2O.ai has spent more than a decade helping organizations generate value from
predictive AI — with open-source H2O, AutoML, and Driverless AI used by data
scientists worldwide.
#TabH2O #H2Oai #TabularFoundationModels #PredictiveAI #MachineLearning #ExcelAI #NoCodeML
