Forrester gave the problem a name. The next step is to give it a number.
The "trust tax" is the drag that ungoverned, unexplainable AI puts on adoption. It is real, it is expensive, and so far nobody measures it:
- Boards approve AI spend they cannot defend.
- Risk committees block deployments they cannot quantify.
- Procurement waves through agents whose controls nobody scored.
A tax that has a name but no price is a tax everyone overpays. What fixed the same problem in lending was not better intentions. It was a score.
Before the credit score, lending ran on relationships and judgment, which could not be compared between two borrowers or carried from one lender to the next. The score turned borrower risk into a number that anyone could compare, set a cutoff against and trade on. A lender could decline below a threshold, price the loan above it, and sell the exposure to someone who read the same number the same way.
Agentic automation is at the moment before the score. Every company is building up ungoverned automation it cannot compare, price or gate. A Trust Debt Index does for agents what the credit score did for borrowers: it turns an agent's controls into a number that production, procurement and deal terms can all read.
Trust debt is a balance you can measure
Trust debt is the gap between what an agent claims and what it can prove. It compounds like financial debt:
- every shortcut taken to ship faster is a liability carried into production, and
- the interest is paid in incidents, audit findings and deployments risk will not sign off.
The index makes the balance visible by scoring the signals that predict failure:
| Signal | What it measures |
|---|---|
| Unsupported claims | Statements the agent makes with no source behind them |
| Citation / evidence gaps | Outputs that cite weakly, partially or not at all |
| Missing controls | No segregation of duties, no approval gate on material actions |
| Absent human-in-the-loop | High-impact actions taken with no person accountable |
| No evidence trail | Actions that leave no record of what happened and why |
| No eval pack / regression gate | No test suite, so quality slips into production unnoticed |
| Stale grounding | Decisions made on information that is out of date |
Each signal can be observed on its own, and each matches a failure a controller already understands:
- an agent that makes claims without sources fails an audit,
- one with no approval gate fails segregation of duties, and
- one with no tests gets worse until a customer notices.
Scored together, they give one number per agent, and rolled up, one number per portfolio.
The standard the index enforces
A score is only useful against a standard. The index measures every agent against one bar. An agent is ready for the enterprise when it is:
- Always sourced: every claim traces to evidence, not to the model's confidence.
- Always owned: a named person is accountable for what the agent does.
- Always tested: a test suite checks every change before it reaches production.
- Always permissioned: the agent acts only within narrow, minimum-necessary authority.
- Always observable: quality, cost, tool use and exceptions are visible in real time.
- Always value-linked: the agent carries a baseline, a target and a lever it moves.
- Accountable to a person: material actions clear an approval gate before they run.
These are not aspirations. They are the seven conditions a low score certifies and a high score flags. An agent that meets all seven carries almost no trust debt. One that meets two carries a liability the index now states as a number, instead of a reviewer's gut feeling.
A number that gates, prices and travels
A comparable score moves the decision out of the meeting and into the workflow. It does three things a named tax never could:
- It gates production. Set a threshold: agents on the right side of it deploy, the rest go back to the harness. The CFO stops funding on faith and the risk committee stops blocking on instinct; both read the same number.
- It prices procurement. The score becomes a buying term, the way a SOC 2 report already is. A vendor shipping high-trust-debt agents finds the tax is no longer abstract: it is a worse price or a lost deal.
- It travels into deal terms. As agents touch revenue and cash, the score becomes part of due diligence. An acquirer, an insurer or a partner will want the number before they sign.
This is not hypothetical. Koko Knows already runs a Trust Debt Index on public companies, turning the control signals disclosed in SEC filings into a comparable score across a universe of companies. Moving that engine from company disclosures to individual agents changes the input, not the method. The scoring discipline exists; agents are the next thing it reads.
The harness is what the score reads from
A Trust Debt Index is not a survey you fill out. It is a meter that reads instruments, and the harness installs them. Each signal maps directly onto a part the earlier essays in this series built:
- The agent card is where sourcing, ownership and the value link are declared. No card, no record of where anything came from.
- The control pack is where segregation of duties, approval gates and permissions live. No control pack, and the missing-controls and human-in-the-loop signals fire automatically.
- The eval pack is the regression gate. No tests, and the test signal fails.
- The evidence ledger is the trail. No ledger, nothing to reconstruct, and the no-evidence-trail signal fires automatically.
That makes the index a diagnosis, not just a grade.
- A high score is the direct result of skipping the harness.
- An agent stood up as a clever script, with no card, no controls, no tests and no ledger, does not score badly because someone judged it harshly. It scores badly because there is nothing for the meter to read, and that absence is the liability the score exists to show.
- Build the harness and the score takes care of itself. Skip it, and the index does not punish you; it shows you the bill you were already running up, signal by signal. That list doubles as your fix list.
The trust tax was always going to get priced. The only question is whether the company prices its own agents first, with a number it controls and can act on, or waits for a procurement team, an auditor or an acquirer to price them instead, on terms it does not set.
This is the seventh essay in the eight-part series The Harness Era. The full-length analysis, with the scoring model for each signal, the full credit-score comparison and how the score reads from the harness, is linked below. The finale brings the series home: Agentic Finance Is the First Killer App, on why the office of the CFO is where governed agents prove themselves first.