AMS Whitepaper V5
A practical reader version of Keigen's AMS framework for qualifying Intent, Attention, Trust, Policy, and Governance before commercial value is released.
Most systems count activity. AMS helps organisations decide what should count.
AMS is Keigen's trust architecture for commercial systems where value is triggered by digital participation. It helps teams decide whether activity is meaningful, whether attention is worth acting on, whether signals can be trusted, whether policy conditions have been met, and whether the decision can be governed after the fact.
The framework is designed for a world where humans, bots, buyers, fans, workers, vendors, and AI agents all produce signals. In that world, organisations need more than dashboards. They need a disciplined way to decide:
What deserves attention? What deserves action? What deserves value release?
This reader introduces the five-layer AMS spine and shows how it applies across Keigen's product adapters: BuyerRecon, RealBuyerGrowth, Fidcern, and Time-to-Point.
Most analytics, fraud, monitoring, and attribution systems were built to observe activity. They count visits, clicks, entries, completions, conversions, sessions, time spent, and logged effort. That is useful, but it is no longer enough.
The commercial question is harder: should this activity trigger anything? Should it influence budget? Should it affect account prioritisation? Should it count as sponsor value? Should it release a reward? Should it support billing? Should an AI-assisted action be accepted?
AMS starts from the moment before release. It asks whether activity is sufficiently meaningful, trustworthy, policy-compliant, and governable before money, access, rewards, priority, billing, or commercial trust are released.
| Layer | External-facing question |
|---|---|
| Intent | What is the activity moving toward? |
| Attention | Is the attention meaningful enough to prioritise? |
| Trust | Can the signal be believed? |
| Policy | Is action allowed under the rules? |
| Governance | Can the decision be explained, reviewed, and improved? |
This is the core shift: from activity tracking to value-release discipline.
AMS helps organisations avoid releasing money, access, rewards, budget, priority, or trust on signals that are too weak, too synthetic, too early, too distorted, or too poorly governed.
Intent looks for direction. It asks whether behaviour is moving toward purchase, evaluation, reward collection, participation, work completion, delegated execution, or another commercially meaningful outcome.
Attention tests signal quality. It asks whether activity is repeated, deep, timely, sequenced, durable, or relevant enough to deserve commercial attention.
Trust checks signal integrity. It asks whether the activity is genuine, eligible, attributable, human or authorised, and not clearly distorted by automation, fake engagement, duplicate identities, or synthetic behaviour.
Policy applies the organisation's own rules. It asks whether thresholds, permissions, eligibility conditions, consent boundaries, sponsor terms, billing rules, or escalation requirements have been met.
Governance protects the decision after it has been made. It asks whether the organisation can explain, defend, review, learn from, or reverse a decision if challenged.
The Benevolent Holding Field, or BHF, is not a sixth layer. It is the operating condition within which the five AMS layers work as intended.
In a weak field, the same verification logic becomes defensive, brittle, expensive, and adversarial. Monitoring costs rise. Escalations increase. Participants optimise for appearing compliant rather than being truthful.
In a strong field, truthful participation becomes easier, manipulation becomes more costly, and repair becomes faster. This matters even more in the AI-agent era, because the same field condition that helps human cooperation remain truthful also helps human-AI collaboration remain verifiable.
For the deeper treatment, read the companion paper: AMS Field Theory — The Operating Condition Behind Governed Allocation.
BuyerRecon helps B2B and complex-commerce teams recognise serious pre-form buyer motion earlier. It turns fragmented behaviour into governed commercial evidence so teams can decide whether a visitor, company, or account deserves attention now.
RealBuyerGrowth helps merchants and growth teams distinguish genuine commercial demand from growth inflated by bots, fake engagement, coupon abuse, or low-quality promotional attention. It improves budget direction and protects future decision data.
Fidcern helps operators verify whether participation in draws, activations, incentives, sponsor campaigns, or limited-release events deserves value attribution or reward release. It protects commercial yield without reducing genuine participant experience to a crude fraud screen.
Time-to-Point helps organisations verify whether claimed work, delegated execution, or AI-assisted output is supported by adequate evidence before acceptance, credit, or billing. It is not time tracking. It is evidence portability for distributed and mixed human-AI work.
AMS should not begin as a large transformation programme. It should start at one economically meaningful control point where release quality matters and evidence can be generated quickly.
Once the first control point proves useful, the same trust spine can expand into adjacent release decisions.
Ready to apply the framework to a live commercial control point?