Salesforce is opinionated PaaS for CRM and business apps. AWS / Azure / GCP are general-purpose IaaS+PaaS.
Salesforce strengths:
- Pre-built CRM data model.
- Multi-tenancy with strict governor limits.
- Declarative tools.
- Built-in security and identity.
- Lower setup cost for CRM use cases.
Salesforce weaknesses:
- Restricted compute / storage.
- Higher per-user cost at scale.
- Limited control over infrastructure.
Cloud (AWS / Azure / GCP) strengths:
- General-purpose; build anything.
- Vast service catalog.
- Pay-as-you-go.
- Total infrastructure control.
- Can scale to extreme volumes.
Cloud weaknesses:
- More setup work.
- Need expertise across domains.
- Security, identity, scaling all on you.
- Operational complexity.
Hybrid is common:
- Salesforce as system of engagement (CRM).
- AWS / Azure / GCP for compute-heavy or specialised infrastructure.
- Connected via APIs / Mulesoft.
Architecture decisions:
- CRM-shaped problems -> Salesforce.
- Custom apps with non-CRM data -> AWS / Azure / GCP or Heroku.
- Heavy compute / ML / streaming -> cloud.
- Mobile / web apps -> can be on either.
Data lakes / warehouses typically on cloud (Snowflake on AWS, BigQuery on GCP).
AI / ML services typically on cloud (AWS Bedrock, Azure OpenAI).
Architects often need familiarity with both Salesforce AND a major cloud platform. Hybrid architectures dominate enterprise.
The senior framing: don't try to put everything on Salesforce or everything on cloud. Match each capability to the right platform.