Regulate the routing. Restrict the risk. Rule the results.
The Attribution Engine and Shadow Metrics Hub turn what each module tracks, analyzes, and maps into results a leader can stand behind. Oversight stays intentional and unmistakably yours.
The operator who stays in command.
Drive it by hand
FlowTabulator is a human driven platform. The operator runs every module themselves: they capture the spend, they attribute it to the team, they decide whether it proceeds. Oversight stays an intentional human decision at every step, so the record points back to the person who made the call. This is the main way to use the platform, and it is where the modules below live.
Or bring your own agent
Bring Your Own Agent is the automated option for operators who want to hand off the repetitive steps. The operator connects an agent they already trust, lets it run the manual load, and keeps the decisions. The agent automates work without creating homework; no extra dashboards to babysit, no new tickets to triage. The same accountable operator stays in command behind it.
What this pillar delivers
- A human decision recorded at every step of every AI workflow.
- Bring Your Own Agent automation that removes manual work, not authority.
- Spend attributed to the team and module that produced it, in real time.
- Oversight that never costs you the data sovereignty you promised.
The modules that carry the loop.
Three reviewers feed this pillar. Each one captures a different surface of AI work and keeps it tied to the human who owns it.
Agentic Diagnosis
Map every step an autonomous agent runs and keep its token cost tied to the human who launched it.
Oversight preserved
Each agent run is reconstructed step by step, so the operator can see exactly what was spent and approve it before it scales.
Bulk Analyzer
Review large batches of prompts or CSV inputs and surface the patterns traditional FinOps misses.
Oversight preserved
Every batch is kept visible to the team that produced it, so leaders command the spend at the source rather than after the invoice.
Prompt Detector
Sandbox prompt hygiene and catch leakage risk before a single token is spent in production.
Oversight preserved
Detection runs locally in the browser. Prompts never leave the device, so oversight never costs you the privacy you promised.
How oversight stays yours.
Three movements turn raw AI activity into a decision a human can stand behind.
Capture
Every prompt, agent step, and batch is recorded with its cost, tokens, and attribution. Nothing is reconstructed after the fact.
Attribute
Spend is mapped to the team, module, and workflow that produced it. The human who launched the work stays visible in every record.
Command
Leaders act on what the attribution shows: approve, redirect, or stop. Oversight stays an intentional human decision, never a default.
Oversight is the first pillar of ESG attribution.
Human in the loop advocacy is what makes the rest of the platform reportable. The ESG Alignment Summary maps the activity these modules capture onto the governance and sustainability pillars your board already tracks. See how the two connect.