Commercial Gate
What question does this solve?
Section titled “What question does this solve?”Technical progress creates a temptation: the design is accepted, so implementation must be next. The model needs an explicit point where someone with the authority to commit resources decides whether the work continues and on what terms, so that progress in the work is never read as permission to spend.
Definition and role
Section titled “Definition and role”A Commercial Gate is a decision point about economic commitment: to continue, revise, or stop. It makes explicit a decision that can authorize, keep, revise, defer or end a commitment, relating the information available, the alternatives and the authority to assume resources or obligations.
It is not a charge at the end of every stage. A gate appears when continuing depends on an economic decision not yet taken, or on revising the scope in force. An accepted output may still have no decision to invest in the next transformation; conversely, an earlier commitment may authorize several stages without a new decision between them. In internal operations the decision may concern budget, priority or use of capacity; the absence of a commercial transaction does not remove the need to recognize who can authorize consuming resources.
Why it matters for the hypothesis
Section titled “Why it matters for the hypothesis”If AI reduces the effort of clarifying a demand and preparing usable options, it may be worth deciding about continuation in smaller commitments. The gate organizes that choice. Adding authorization points without lowering the cost of reaching them only adds work to buying.
How it works
Section titled “How it works”What justifies a gate
Section titled “What justifies a gate”A gate should exist because of a relevant decision. Moving from one stage to another may be a good occasion, but is not a justification by itself. An analysis can provide information to authorize implementation; a discovery can reveal work outside the approved scope; a requested change can need new terms of price or time. There can also be agreed limits, such as a ceiling on consumption, and the operation must recognize when it is approaching one before exceeding what was authorized. No extra gate is needed for activities already covered by delegated decision rules.
Information for deciding
Section titled “Information for deciding”| Information | Question it supports |
|---|---|
| State of the demand | What has already been produced, accepted and consumed? |
| Proposed work | Which result or transformation would be authorized next? |
| Current estimate | What effort, time and cost are expected, under which assumptions? |
| Commercial terms | Which price, form of charging or resource limit is being discussed? |
| Dependencies and risks | Which conditions could change continuation? |
| Alternatives | Is it possible to reduce, defer, investigate or stop? |
| Effects of the decision | Which obligations, reservations or consequences follow? |
| Authority | Who can decide, and within what scope? |
Not every decision needs new studies. When an essential basis is missing, the operation should acknowledge that clarification is needed rather than present artificial certainty. Past consumption must be kept apart from the proposal for the future: work already done may give context, but it does not show that continuing is the right choice.
Possible outcomes
Section titled “Possible outcomes”Authorizing is one outcome, not the only one. The gate can keep the current commitment, ask for revision, authorize a smaller exploration, defer, or stop. Each needs an understandable scope: approving further analysis does not authorize the full solution, and asking for revised terms does not provisionally accept the proposed work.
Silence is not consent. A missing answer is not an authorization, unless a rule agreed in advance gives it that meaning within clear limits. A conditional decision must state its conditions and how meeting them will be recognized; otherwise execution can start on differing interpretations of what was approved.
Authority and delegation
Section titled “Authority and delegation”Governance establishes who may assume or revise a commitment. Knowing the problem or taking part in its evaluation does not confer economic authority: someone may accept business rules without being able to raise a budget, and someone else may approve resources without the competence to judge whether an implementation meets the agreed behaviour. Delegation can allow decisions within a limit already set; the mechanism can then be simple or automated, as long as it recognizes the authorized scope and routes exceptions. When a decision exceeds that scope, it goes through Escalation. Being able to act is not being authorized: see Mandates.
Not the Quality Gate
Section titled “Not the Quality Gate”The Quality Gate concerns whether an output meets technical or quality conditions. The Commercial Gate concerns the economic commitment. They may use some of the same information, but they answer different questions. An accepted analysis can show that a solution is feasible under the conditions examined, and the buyer can still decide not to take on its cost; the economic result does not invalidate the technical analysis. Conversely, an available budget does not show an output is adequate. Both decisions can happen in the same interaction, by one person or by systems with the relevant delegations; the record must show their separate scopes, especially where a generic “approved” could be read as authorizing different actions.
Price, billing and payment
Section titled “Price, billing and payment”Authorizing a commitment, setting a price, issuing a charge and making a payment are distinct events. The gate may be part of that sequence without coinciding with all of them. An agreement can foresee payment in advance, by milestones or by consumption; the model chooses no universal form. Accepting a stage does not, by definition, trigger a charge. Confusing an operational stage with a financial instalment makes the design of the flow depend on a correspondence nobody needs. See Pricing.
Waiting and the cost of deciding
Section titled “Waiting and the cost of deciding”A decision may need analysis, negotiation or people who are unavailable, and that time affects continuity. While a gate is pending, independent authorized work may continue; keeping resources busy does not justify anticipating the missing commitment. Capacity reserved or environments kept during the wait may cost something, and that should stay visible. Too many gates increase coordination costs; too few leave consumption and changes of scope undecided. The effect of AI should be examined from the preparation of the decision onward: summarizing context, checking dependencies, comparing alternatives and relating them to the mandate. Automating only the final answer may shorten one wait without making the rest of the route sustainable.
In the running example
Section titled “In the running example”The barbershop has completed a discovery priced at 20. The provider offers 80 to implement, evaluate and deploy booking; reminders would add 15; advance payment remains outside the available offers. No amount is a market price.
The customer has delegated the continuation under an additional ceiling of 90, separate from the price of discovery. The mandate requires usable booking, an identified integration, defined acceptance criteria and compatible commercial terms. The authorized system checks those conditions and the validity of the offer. Booking at 80 fits; booking with reminders at 95 does not. Budget availability alone is not enough.
The system accepts the offer at 80 and records scope, price and references. Ten remains as uncommitted authorization: not spending, not revenue for the provider, and not general permission to produce extra features. The reminders remain unauthorized; their proposal can return for revision or a new competent decision. If the essential technical conditions had not been clarified, even the offer at 80 would not go ahead under this mandate. A direct human decision could have produced the same choice; the example shows what the gate does and how it can be automated, which is a hypothesis about reducing latency and coordination effort, to be evaluated together with the quality of decisions and the cost of handling exceptions.
Boundaries and failure modes
Section titled “Boundaries and failure modes”A decision that cannot say no adds nothing. A gate that only records an approval that was inevitable, with no real possibility of revision or stopping, contributes little.
Authorizations can lapse. An authorization may stop applying if its conditions no longer hold. Coordination must recognize that and route a revision, instead of treating the record as permanent, unrestricted permission. The Orchestrator uses the decision to coordinate the next step and Routing mobilizes resources accordingly; neither may widen the economic scope because it can execute more work.
It does not guarantee a good decision. Incomplete information, unclear criteria or unavailable authority can limit its usefulness. The right boundaries, limits of delegation and cost of deciding by context remain open. The model sets no contract clauses, payment schedule or universal financial approval criteria.
Connected concepts
Section titled “Connected concepts”- Economics overview
- Previous: Progressive Commitment
- Next: Pricing
- Quality Gate, the separate technical decision
- Mandates, how the decision can be delegated
- Glossary
Software Delivery as a Service is a proposed model by Renan Deves and Marcelo Scheidt, Independent Researchers. See the paper for citation and license.