L3 is where a single agent (coordinator with a loop, tools, memory) executes multi-step workflows autonomously. Human sets the goal and reviews. The agent picks steps, calls tools, retries, produces a deliverable. This is BOD's entry point. L3 is where workflow economics show up, not user productivity. A function gets faster and cheaper.
The industry-standard "climb the stack" story tells portfolio companies to start at L1, build to L2, hope to reach L3, and dream of L4. That story keeps the systems integrator employed. It doesn't deliver EBITDA.
BOD inverts it. We start at L3 because L3 is where workflow economics show up: cost-per-task replaces cost-per-seat as the metric.
From that anchor we move in three directions:
The 80% chat-and-copilot adoption isn't the destination. It's the surface area that exists after the workflow is automated.
BOD starts where workflow economics compound, then expands down, up, and re-incorporates L1 at the edges.
| Function | In practice | Signal |
|---|---|---|
| Finance | Variance, narrative and board-deck agents, chained, evaluated, observable | FP&A 10d → 2d |
| Sales | The agent enriches, scores and routes leads; humans close | Conversion uplift |
| Support | Agent drafts the reply; human reviews; ticket closes faster with audit trail | AI-first triage |
| Legal | Agent feeds the deal desk a risk-scored summary | Not a 60-page redline |
| Engineering | Claude Code / Cursor Agent in CI opens PRs from failing tests & lint debt | PR throughput 2× vs L1 |
This is the shift that makes a function cheaper, not just faster: the steps a person used to run are now executed by the agent, with a human reviewing the one decision that needs judgment.
Same workflow as L1, but cost-per-task replaces cost-per-seat the moment the agent runs the steps.
| Category | BOD Pick | Why | Pricing signal |
|---|---|---|---|
| Durable workflow execution | TPTemporal | BOD default. Production-grade durable execution for AI workflows. | OSS + Cloud tiered |
| Coding agent in CI | CCClaude Code | 78.4% SWE-bench Verified; 5.5× lower token cost vs Cursor Agent; MCP-native | ~$125/seat Premium |
| Coding agent (IDE) | CUCursor Agent | Fortune 500 default; multi-provider; agent + editor | ~$40/seat/mo Teams |
| Eng-grade orchestration | LGLangChain / LangGraph | BOD L3+ default for engineering teams | OSS + LangSmith |
| Data-centric RAG agents | LILlamaIndex | BOD L3 default for data-heavy retrieval | OSS + cloud |
| OSS visual workflow | N8n8n | Self-hostable, sandboxed code, sovereignty-friendly | EUR 20-667/mo Cloud |
| Hyperscaler-native |  Dbrx Workflows · Snow Tasks + Cortex · Bedrock + Step Functions | When the data platform is the lock-in | Bundled |
| Pre-built (rev ops) | LDLindy · Relevance AI | Fast value for SDR / exec admin / research agents | $49-200+/mo |
| CX outcome-priced | SISierra | Paid on resolution. Outcome aligns incentives. | Per-resolution |
| Inside Salesforce | AFAgentforce | Locked to SFDC; three pricing models | $125/seat OR $0.10/action |
| Evals | BTBraintrust · Langfuse | BOD primary; OSS alternative | Tiered |
L3 is where consulting engagements compound. Edge Deploy drops a production-grade agent into your stack in 90 days: finance variance, sales SDR, support tier-1, AP, recruiting, product research.
Each is built to run on Edge Scale, our Agent Ops control plane (product-in-build): governance, audit, cost attribution, RBAC, SSO, and a managed tier targeting 99.9% SLA.
Each is a repeatable 90-day engagement that drops a production-grade agent into the client's stack, wired to the Edge Scale control plane as it comes online. Same architecture, same evals discipline, same control plane.
Close-to-commentary cycle compressed 10d → 2-3d. Variance agent reads the close package, drafts commentary, flags exceptions, routes to the CFO for sign-off.
Research → personalization → send → reply classification → meeting. Pipeline doesn't grow because reps work harder; it grows because the agent works while reps sleep.
Agent drafts the reply; human reviews. Deflection moves from 30% (L2) toward 70%+, without losing the long tail that needed a human.
Invoice intake → matching → approval routing → exception escalation. Pairs with Ramp / AppZen where appropriate.
Inbound screening → scheduling → first-round scorecard. The recruiter takes the shortlisted candidates into final rounds, not the inbox triage.
The agent watches usage signals, drafts feature briefs, summarizes user research, builds the PM the agenda for the weekly review. PM ships, doesn't curate.
"L3 is where the economics compound. L4 is where they inflect. L5 is where exit multiple moves."
BOD Maturity Memo, rev 3Edge Deploy ships production-grade agent blueprints: finance variance, sales SDR, support tier-1, AP, recruiting, product research. Drop one in. Measure it. Then we expand outward.
Talk to BOD