The platform under the agents.Reference architectures for organisations shipping more than one.
A single agent is a project. A portfolio of agents is a platform. We design and stand up the foundational architecture — shared MCP servers, shared observability, shared eval infrastructure, shared cost attribution — so your second, third, and fifteenth agent are cheaper, faster, and more consistent to ship than the first.
Most organisations build their second agent from scratch because the first was a one-off. The platform layer is the difference between linear and sub-linear cost-to-ship.
~50%
faster on the second agent vs. the first
~30%
cheaper steady-state operating cost from shared infra
0
duplicate MCP servers in maintenance after migration
Case studies
How recent engagements actually shipped
IT Services · 6 weeks discovery → handoff
PR review pipeline cuts senior-engineer time 4×
Mid-market IT services firm · Ahmedabad · 180 engineers
Problem
Senior engineers were spending 8–12 hours per week each on first-pass PR review across a 6-team monorepo. Junior PRs waited 2+ days for sign-off; velocity stalled; the highest-judgement people were doing the lowest-judgement work.
Solution
A multi-agent CI workflow triggered on every PR open. Three specialist agents run in parallel — a reviewer (Claude Sonnet 4.6) for code-correctness and convention, a security agent for risk patterns, and a test-generator agent for coverage gaps. Outputs are consolidated into a single PR comment within 90 seconds. Humans review the agent's synthesis, not the raw diff.
Claude Sonnet 4.6 (reviewCustom MCP server: GitHub APIGitHub ActionsLangfuse traces
~36 hrs/wk
senior engineer time reclaimed across the team
< 3 days
payback period at loaded-cost rate
4×
review throughput per senior engineer
0
production regressions traced to AI-passed reviews in 90 days
Customer support backlog had grown to ~340 open tickets. Level-1 triage took 12–20 minutes per ticket on average, and 35% of tickets were misrouted on first pass — every misroute became a customer-facing escalation churn.
Solution
A supervisor-pattern agent that ingests email and form submissions, classifies the issue, queries the customer's Odoo instance for context (open invoices, recent modules, last login, current contracts), drafts a Level-1 response with the right module screenshots inline, and routes complex tickets to the right consultant with a pre-filled handoff brief.
Claude Sonnet 4.6 (drafting)Custom MCP server: Odoo (read-only customer / order / invoice scope)Supervisor patternPydantic schemas
340 → 18
open L1 backlog within 6 weeks of go-live
~60%
L1 staffing reduction on agent-eligible categories
Audit-grade compliance review ships under multi-layer guardrails
Regulated financial-services intermediary · India · 95 employees
Problem
Manual compliance review of vendor and onboarding documents was the bottleneck for new-customer activation. Every traffic spike threatened SLA breach. Reviewer fatigue led to inconsistent flagging — some weeks too strict, some weeks too loose, with no defensible pattern.
Solution
A single-agent system wrapped in four guardrail layers: an input filter that detects and redacts PII / strips prompt-injection patterns; a versioned policy registry the agent must cite by clause ID for every conclusion; output validators (schema + LLM-as-judge cross-check); and a human-in-the-loop gate on anything scored above a defined risk threshold. Every decision is appended to an immutable audit log.
Custom detectorsClaude Opus 4.7 (final ruling)Versioned in repoPydantic v2