AI Tools
Enterprise AI platforms — capabilities, cost models, governance, and adoption sequences.
Working references for the managed AI platforms that sit next to enterprise data — what each product actually is underneath the marketing, where the money goes, what the governance story is, and a defensible order to adopt things in.
Every page here is written for the same reader: someone who has to pick a platform, estimate a bill, and answer a security review — not someone shopping for a demo. Two of those three jobs are the same whichever platform wins, so they have their own pages: estimating a bill and what a security review asks. The platform pages below cover the third.
Estimating an LLM bill
18 d agoThe token arithmetic every platform bills on — where the money actually goes, and which levers move it.
What a security review asks
18 d agoThe questions every enterprise AI platform gets asked, what a good answer looks like, and which ones are genuinely hard.
AWS Bedrock AI Strategy
1 mo agoConsolidated overview of Bedrock capabilities, use cases, architecture patterns, governance, and a sequenced project portfolio for a wealth management firm.
AWS Bedrock Use Cases for RIAs
1 mo agoA catalog of 112 concrete Bedrock applications for a wealth management firm, grouped by function — document intelligence, data engineering, CRM, compliance, analytics, backfills, and agents.
Snowflake Cortex AI: The Big Picture
1 mo agoA non-technical orientation to Cortex AI — what it is, why it matters strategically, where the real value is, what it costs, and what to be skeptical about.
Snowflake Cortex AI: A Working Engineer's Guide
18 d agoImplementation-level Cortex reference — product map, setup, AISQL functions, Cortex Search, the token cost model, monitoring queries, governance, and gotchas.
The through-line
Both platforms sell the same core promise — models that run next to governed data, under existing IAM, without shipping records to a third-party endpoint. They differ in where the gravity is:
| AWS Bedrock | Snowflake Cortex | |
|---|---|---|
| Center of gravity | AWS services, S3, documents, APIs | Data already in Snowflake |
| Interface | SDK, agents, flows | SQL |
| Strongest at | Application reasoning and orchestration | In-warehouse transformation and retrieval |
| Main cost trap | Idle agent/knowledge-base infrastructure | Model choice, and Search indexes billed while idle |
A hybrid is usually correct: the warehouse stays the governed data platform, the cloud platform provides application reasoning. Decide per workload rather than declaring one universal AI platform.
What goes here
- Platform capability maps and honest comparisons
- Cost models and the levers that actually move the bill
- Governance artifacts, evaluation measures, and adoption sequences
Related: LLM Wiki on building a self-maintaining knowledge base, and RAG for transit planning on grounding answers in an organization's own documents.