What does each role own?
A Finance Controller FTE reconciles bank, processor, and invoice records daily and proposes every match with evidence. A Customer Support FTE answers first-line categories from the knowledge base with citations and escalates the rest with summaries. A Recruiter FTE screens against your scorecard with evidence per score and schedules interviews. An SDR FTE researches, qualifies with evidence, drafts outreach, and keeps the CRM honest. A Search Visibility FTE samples buyer questions, proposes with verdicts, drafts through the fact gate, and measures citation presence. A Social Growth FTE runs native presence with reply rules. A Performance Ads FTE monitors and proposes inside caps. A Marketing Strategist FTE runs campaigns to briefs. A Content & SEO Engineer FTE builds clusters and publishes as code. Market Entry and Research Analyst FTEs return ranked hypotheses with sources.
The how-it-works pieces in the library cover each role mechanically; the discover pages describe each for hiring.
What does every role never do?
Decide about a person: hire, reject, offer, discipline, or coach. Set or change a price, a discount, or terms. Move money: refund, credit, post an entry without approval, or spend beyond a cap. Contact a customer, candidate, or prospect about money, a complaint, or a decision without an approved draft. Continue a conversation a person owns. Publish or post unattended unless a narrow class has earned it under the Promotion Standard. Widen its own scope or credentials. And anything on its remit's always-escalate list, which reaches a person with a summary at every autonomy level. The never-automated list is the platform's and cannot be changed by anyone.
The Three-List Remit framework describes how each role's Never list is written on top of the platform's.
Which role should we hire first?
The one whose function has the clearest evidence and the most hours, under the First-Function Matrix. For most teams that is a Finance Controller FTE for reconciliation, because the right answer lives in the bank feed and the ledger, or a Customer Support FTE for one high-volume category, because the answer lives in the knowledge base. Search visibility is the highest-leverage and slowest to show; SDR needs a protected list written first; recruiting needs a scorecard and a data protection decision. Judgement-heavy roles, market entry and research, come later, once the approval habit exists.
The industry pieces in the library apply the same choice to each vertical, and the persona pieces to each seat.
Can a role be composed for a function the catalogue does not cover?
Yes. A Digital FTE is a remit with three lists, connections, targets, and a manager, and the catalogue roles are compositions of that shape for common functions. A function the catalogue does not name, admissions enquiries, HR policy questions, employee onboarding, supplier queries, is composed the same way: sort the work, write the remit, connect the tools with scoped accounts, set targets, and run the thirty-day pattern. The library's pieces on admissions enquiries, HR policy questions, and onboarding show composed roles. The never-automated list applies to every composed role exactly as to the catalogue.
The comparison of a Digital FTE with an agent framework describes the line between composing a role and building an agent.
How do roles share knowledge without sharing scope?
Through the Brand Brain, which is organization-wide: the company's voice, facts, policies, products, and competitor facts, each with provenance, pre-filled at hiring and confirmed by managers, read by every role. Nothing captured anywhere on the platform is re-asked by a new role. Scope, by contrast, is per role: each has its own connections, credentials, budget, approval record, autonomy ladder, and targets, so a promotion in one role never touches another and a Finance Controller FTE's credentials never serve a Support FTE. The myth piece on one agent doing every job explains why the sharing stops where it does.
For multi-property organizations, each property has its own memory and scope within the organization's.
How many roles should a team run, and how are they managed together?
As many as the record supports, added one at a time under the Second-Role Test: the previous role settled with a promoted class, the manager's approval time falling, the next function's evidence clear with a baseline, and the block funded from usage. Managed from the workforce view, which combines every role's morning brief into one list ordered by what blocks the most, with routine approvals routed to the people closest to each function and each role reviewed monthly on its own record. A COO running four roles usually spends less time than on the first two, because the routing and the ladder did their jobs.
The COO persona piece and the hybrid team whitepaper describe the shape of a team with several roles.
Do roles talk to each other?
Through the workspace and the shared memory, and through a manager's coordination, not through unsupervised handoffs. A Marketing Strategist FTE running a campaign coordinates the pieces the content and social roles produce, but each role still proposes to its own manager through its own gate. A Support FTE's knowledge gaps become topics a Search Visibility FTE's verdicts consider, through the manager reading both. A connector can expose a Digital FTE to an external agent, or to one your team built, as an MCP server with the approval boundaries intact. Nothing crosses a gate because another role asked.
The launch planning playbook shows several roles working from the same approved facts through their own gates.