Marisa, from I Already Said No, should be able to check what her cancellation changed. The service should supply a dated record from billing, identifying the subscription and its effective end date, with any final amount due. A later statement then lets her test whether collection agrees with that record.
A statement without a charge is evidence for the period it covers. Collection may also be absent because a card was blocked or a payment failed. A final invoice can be legitimate after renewal ends. The customer needs enough detail to tell these states apart without another tour through support.
An institution’s own record can carry evidence of effect. Outside corroboration makes it easier to test. The useful question is what the record establishes and who can challenge it.
Capacity and delivery
A cancellation record and a power grid have little in common, except that both make promises somebody downstream has to test.
Installed capacity measures the generating equipment in place. Delivered service concerns electricity that reaches people when they need it. Transmission constraints and curtailment can keep available generation from becoming usable supply.
A capacity target should be tested against capacity, and meeting it is an achievement on the measure specified. Household reliability requires another measure, including interruptions and the hours when supply is available. Counting the plant does not settle what reaches the socket.
China’s official announcement of its 2021–2025 energy-technology plan describes grid development to accommodate renewables, alongside enhanced storage. Its published objectives treat integration work separately from generation technology. The plan’s window closed at the end of 2025, so its commitments can now be read against delivery records. The announcement, by design, cannot supply those.
A public plan is a commitment whose implementation can be evaluated. Its authority differs from a customer’s contractual cancellation. The accountability questions still travel: who owns the result, what effect is promised, when is it due, and what evidence demonstrates it?
Performance scrutiny also stays separate from rights scrutiny. A successful energy project cannot establish that people displaced by it could contest the decision. One household’s reliable supply and another household’s loss each require their own account. Delivery in one domain buys no credit in the other.
Who chooses the measure
An institution can choose a measure that catches its preferred activity and leaves the affected person’s result outside the frame. A cancellation dashboard might count acknowledgments sent within a target window. That records response speed. Whether valid instructions changed renewal by the effective date is a different check. (We do love a dashboard.)
The metric should distinguish prohibited new renewal from a legitimate final obligation. It should also track unresolved cancellation states directly. Waiting for a successful charge to reveal them means a failed payment can make an unfinished cancellation look like a completed exit.
The customer’s time spent reconciling screens is a cost the service’s dashboard can miss. A household’s spending on backup power can vanish from an installed-capacity measure. Each omission has a scope. Asking affected people what they need to verify exposes costs the preferred metric never captured.
Checkable by whom
A billing-generated cancellation record gives the customer something definite to keep and contest. An independent payment record helps test subsequent collection. Where a bank exposes a mandate, its status is evidence about that payment authorization. The service’s renewal state still needs its own record.
India’s RBI framework assigns the issuer an online mandate-withdrawal facility with authentication, and identifies a pipeline-transaction exception. It is a bounded payment safeguard. Whether the underlying contract has ended is a question it leaves to the contract.
For a public energy commitment, independent measurement can test an official report against supply experienced outside the reporting institution. The method and coverage of each dataset matter. A national aggregate may miss local outages even when its totals are accurate. An outside record deserves scrutiny of its own.
Evidence should reach the person affected in a form they can use. A promise checkable only from inside a billing console or a specialist grid model is checkable by very few of the people who depend on it.
The failed check
A record that contradicts the promised effect should trigger a named route to someone empowered to correct it. Marisa should be able to point to the canceled subscription and the inconsistent renewal attempt without proving that the vendor’s entire platform is defective. The service should reconcile its own records and explain any remaining obligation.
An institution can define a precise promise and still fail it. Checkable evidence makes that failure demonstrable. It also needs a recipient with authority to respond.
The socket is where the promise meets someone else’s life.
Companions
- Reversibility before action: Before the Commit.
- The execution route: The Stop Must Reach the System.
Bibliography
Reserve Bank of India, “Processing of e-mandate on cards for recurring transactions” (21 August 2019), paragraphs 15–17, notification.
Xinhua, “China Unveils 5-Year Plan on Energy Technology Innovation” (2 April 2022), Chinese government website.
These notes come out of Sociable Systems, a practice that reads AI-shaped documents the way a hostile reviewer will, before a lender or a court finds the gap. The argument has an operational form: the Interim Protocol sets out four rules for AI use in environmental and social deliverables, covering disclosure at touch-point grain, evidence custody, the phrases no automated screening may settle, and a hostile read before anything ships. Free, and written to be cited or retired once institutional guidance arrives.