Ryzeo
The retained public case supports a Tcules UX revamp for marketing-automation software. It is the strongest direct domain proof on this page. It does not by itself establish current AI-agent capability or a quantified revenue outcome.
MarTech and SalesTech
Revenue software must separate what it observed, inferred and changed
MarTech and SalesTech products connect customer data, audiences, content, campaigns, accounts and revenue work, but their interfaces become hard to trust when observation, inference and automation are not separated.
- MarTech
- SalesTech
- Revenue software
Revenue products have different operating models
Revenue products share a condition: data becomes a claim about a customer or an instruction to act. That condition matters more than the presence of large amounts of data.
| Product subdomain | Primary object | Consequential product decision |
|---|---|---|
Customer data and audience platforms | identity, profile, event and segment | source, identity resolution, consent and activation boundary |
Marketing automation and lifecycle engagement | audience, message, journey and trigger | eligibility, suppression, timing, channel and recovery |
Advertising and campaign operations | campaign, creative, placement and spend | approval, pacing, attribution, policy and optimisation authority |
Sales intelligence and engagement | account, contact, signal, sequence and owner | signal quality, prioritisation, outreach permission and hand-off |
CRM and revenue operations | account, opportunity, activity and forecast | ownership, stage criteria, exception and forecast confidence |
Analytics, attribution and optimisation | event, model, comparison and recommendation | data coverage, causal limits, confidence and decision relevance |
The reviewable signal-to-action chain
Source -> event -> signal -> rule or model -> recommendation -> authorised action -> observed result
Each link answers a different question: where the data came from, what happened, what the system inferred, which policy or model interpreted it, what is being recommended, who or what may act, and which later evidence can be connected to that action.
Collapsing the chain makes products look simpler while making them harder to challenge. Tcules uses it to decide where a user needs explanation, comparison, correction, approval or recovery.
- 01
Objective
What result the action is intended to influence.
- 02
Scope
Which people, accounts, campaigns or records are affected.
- 03
Basis
Which signals, rules and exclusions produced it.
- 04
Authority
Who can approve, edit, execute, pause or reverse it.
- 05
Measurement
Which later observation will be treated as evidence, with what limits.
Review a signal-to-action path
Use one signal-to-action path to locate where observation becomes inference and who is allowed to change product state.
Attribution is a model of incomplete history
Revenue interfaces should distinguish recorded events from inferred contribution. A useful attribution view reveals coverage, lookback window, model choice and material missing data before presenting a result as guidance. The same discipline applies to lead scores and next-best-action recommendations.
Compliance affects the experience as well as policy. The FTC's CAN-SPAM guidance (opens in a new tab) distinguishes commercial from transactional or relationship messages and requires truthful routing, opt-out and other controls for covered email. The IAB Tech Lab's Transparency and Consent Framework (opens in a new tab) provides technical specifications for communicating consent choices in relevant advertising ecosystems. Applicability and legal interpretation remain client-owned.
Where are consent, suppression and channel rules enforced?
Consent, suppression and channel rules must be checked before a message, audience activation or automated step executes; advisory text in the interface is not enough.
The authoritative policy or eligibility service, approved workflow and channel integration should return a visible allowed, blocked or review-required state. The product records which rule and customer state were used. Client legal and policy owners define applicability, while Tcules can design and implement the agreed enforcement path.
The marketing-automation product guide traces audiences, signals, rules and actions through campaign state.
AI moves from content generation toward delegated revenue work
Generation can draft. Retrieval can assemble evidence. Recommendation can prioritise. Agents can prepare or execute steps. The larger product change is that these capabilities can now be combined across account context and multi-step revenue operations.
That increases the importance of data permission, source visibility, action scope and human control. A salesperson reviewing one drafted message needs a different safeguard from an agent changing 10,000 campaign records. The interface should adapt control to consequence rather than add the same approval button everywhere.
Where product, experience and implementation meet
Tcules can map customer and revenue objects, clarify roles and states, design operational workflows, prototype AI behaviour, build shared interface systems and implement agreed web or AI-enabled product scope.
In Tcules' current method, the product model is intended to remain connected to components, APIs, evaluation scenarios and release acceptance so automation does not acquire accidental authority during implementation.
Bounded proof
Auxentios
The retained case supports CPQ domain research, stakeholder work, eight qualitative interviews across countries and specialist roles, synthesis, MVP design and a component-led visual system. It demonstrates the revenue-execution side of this category without claiming a production application or measured commercial result.
Continue from here
Start a project
Tcules designs the path from signal to decision and action across product modelling, workflow and interface design, AI behaviour, design systems and agreed software delivery.