Blog

Analysis, practice, opinion.

Short pieces on what we are seeing in real Salesforce, data and AI projects.

Lakehouse Is Not a Silver Bullet: When Plain Warehouse Still Wins

Read →

Data observability: catching pipeline failures before stakeholders do

Read →

AI FinOps: how to charge LLM inference back to internal customers without fighting IT

Read →

When NOT to use Salesforce: 4 scenarios where license cost beats ROI

Read →

When an agent is the answer — and when it's an escape from a poorly modeled problem

Read →

Data Cloud is no longer a CDP — it's the central nervous system of Salesforce

Read →

Modern Data Stack in 2026: what survived, what died, what became commodity

Read →

Tableau as an executive language: killing the vanity dashboard

Read →

Salesforce Partner Program: what the client never tells you

Read →

Databricks vs Snowflake vs BigQuery: lock-in, exit costs and what the official partner doesn't say

Read →

AI for HR: a practical case of the internal triage agent

Read →

Churn analysis: the mistake of defining "loss" before strategy

Read →

Multi-agent in production: what we learned running 5 agents for 90 days

Read →

Salesforce Industries (Vlocity): worth it for Brazilian mid-market companies?

Read →

Open source vs proprietary LLMs: honest criteria to choose without ideology

Read →

Does dimensional modeling still matter in 2026? Yes — a defense against lakehouse-for-everything

Read →

CPQ in B2B SaaS: the difference between a quote and an actual proposal

Read →

Generative AI in sales: beyond "ChatGPT for replies" — where it makes revenue

Read →

Prompt engineering for analytics: the forgotten pipeline between data and report

Read →

Marketing Cloud + Data Cloud: the stack that should have been born together

Read →

Data privacy in LLMs: the governance checklist missing from the pilot

Read →

The Data Catalog nobody uses: a symptom of the real problem (not the tool)

Read →

Sandbox strategy: how to avoid "the last sandbox refresh was 4 months ago"

Read →

Real inference cost: how to avoid the end-of-month USD surprise

Read →

Self-service BI: why every department has its own "final draft"

Read →

Salesforce ↔ ERP integration: projects stall on the contract, not the architecture

Read →

Multi-agent systems: when to orchestrate vs. consolidate into one agent

Read →

Product metrics: why north star becomes "north dust" in 6 months

Read →

Salesforce Flow vs Apex: when code is worth more than clicks

Read →

Vector databases compared: Pinecone, Weaviate, pgvector — when each one fits

Read →

Data contracts: the least painful way to not break production

Read →

Agentforce in customer service: what to automate and what NOT to

Read →

Fine-tuning vs RAG vs prompt engineering: how to choose without burning cash

Read →

Snowflake vs BigQuery vs Databricks: an honest mid-market comparison

Read →

Customer 360 vs CDP: differences that change the data roadmap

Read →

Agent evaluation: the metric nobody wants to publish

Read →

ELT vs ETL: why the fashion changed and what actually matters

Read →

Service Cloud: SLA isn't decoration — measure capacity before promising

Read →

LLM as internal agent: three cases where it works, two where it fails

Read →

dbt in practice: documentation is the real win, not the model

Read →

Sales Cloud: five antipatterns that separate costly rollouts from rollouts that pay back

Read →

RAG in practice: retrieval is the bottleneck, not the LLM

Read →

Clean data is a myth: living with imperfect quality

Read →

Map the process before Salesforce: the checklist that saves months

Read →