Are companies becoming more efficient by buying tools—or silently losing engineering strength?
Why companies buy tools
- Speed: Subscribe to an observability platform instead of building one. Ship product features now rather than creating infrastructure.
- Lower upfront cost: Use an identity platform rather than staffing a team to maintain security patches, token rotation and multi-factor authentication.
- Advanced capability: Use AI coding assistants where building an in-house model would be impractical.
- Focus on the core product: Use established analytics and alerting platforms so engineers can concentrate on customer-facing value.
The hidden risks
- Skill erosion: A team may become comfortable reading dashboards but lose the ability to investigate the underlying platform directly.
- Vendor lock-in: Product or licensing changes can force costly, lengthy migrations.
- Cost explosion: A reasonable initial subscription can grow dramatically as logs, metrics, users or data volume increase.
- Security exposure: Proprietary logic or customer information may be processed by third-party services.
- Tool sprawl: Multiple overlapping tools make it unclear where the source of truth lives.
The real problem
Tools should assist engineers, not replace engineering judgement.
If a team cannot operate—even slowly—without its tooling, the dependency may be unhealthy.
A quick health check
Can we debug production without our APM tool?
Healthy: Yes, using logs and metrics directly.
Dangerous: We
would not know where to begin.
Can we ship code if the AI coding assistant is unavailable?
Healthy: Yes, although work may be slower.
Dangerous:
Engineers struggle to produce routine code without it.
Can we monitor infrastructure without the vendor?
Healthy: Yes, we retain knowledge of the underlying cloud and monitoring
fundamentals.
Dangerous: We would be blind.
Do we understand our own architecture?
Healthy: Yes, it is documented, discussed and reviewed.
Dangerous:
Only diagrams generated by a tool explain it.
A balanced approach
- Non-core → Buy: Use SaaS for commodity concerns such as monitoring, CI, alerting and documentation.
- Core → Build: Keep competitive-advantage logic in-house, including domain logic, pricing, order orchestration and ranking.
- AI tooling → Accelerate: Use AI for boilerplate, while engineers review, understand and own every line.
- Knowledge → Maintain: Keep architecture decisions human-led and practise incident response without depending entirely on dashboards.
The long-term outlook
Tools + strong engineering culture + controlled dependency
Creates an
organisation that is resilient, cost-efficient and adaptable.
Buy everything + skip understanding + automate blindly
Creates high costs,
skill gaps, vendor lock-in and technical debt.
Buy tools to increase capability, not to replace competence. The real advantage is the balance: smart automation paired with deep engineering ownership.