← Playbook hub
algomarketing
Run Book · Part 4 of 7
Part 4 · Governance & Risk · Delivery Run Book

Governance & Risk

Stand up the control stack that lets a team move fast with AI without flinching: the obligations they must meet, the oversight modes matched to risk, machine-readable rules so agents stay on-brand by default, and the approval, provenance and audit plumbing that proves it. Follow this run book to build control that earns its keep — the steps to take, who needs to be in the room, the effort each takes, and the registers and checklists that keep the framework honest after you go.

Duration
4–6 weekswithin Phase 2 · Build
Effort
~10–14 dayslead consultant + martech
Client team
Legal + brand+ content, ops, martech
Output
Framework livegoverning real work, not shelfware
Overview

What this part delivers, and why

AI didn't remove the need to govern content — it raised the stakes. Once a model can write in the brand's voice at scale, an off-brand, inaccurate or non-compliant piece spreads as fast as a good one. This part builds the layered control stack that turns that risk into a system: the obligations the team must meet, oversight modes matched to risk, rules an agent can actually read, and the approval, disclosure and audit machinery that proves what happened. The goal is a framework that governs real work — not a policy nobody opens.

The six steps at a glance
  • 1 · Map the obligations — the binding floor, the standards buyers ask for, disclosure & provenance, sector rules, into one register.
  • 2 · Risk tiers & oversight modes — four oversight modes; map each content type to a tier so effort goes where the risk is.
  • 3 · Machine-readable brand & editorial rules — guidelines structured so agents conform by default, not by luck.
  • 4 · Approval workflows + disclosure/provenance — review wired into the CMS; C2PA / metadata to prove what was AI-assisted and who signed it.
  • 5 · Prompt library + agent guardrails — version-controlled prompts, least-privilege access, audit logs against the agentic risks.
  • 6 · Governance body & cadence — a content/AI council, how it meets, and who decides what.
The control stack, top to floor
  • Binding floor — EU AI Act Article 50 (machine-readable marking; a human-editorial carve-out). Article 50 transparency duties apply from 2 August 2026.
  • Voluntary standards — ISO/IEC 42001 & the NIST AI RMF GenAI Profile: the credentials procurement now asks for.
  • Industry disclosure — the IAB AI Transparency & Disclosure Framework, with C2PA provenance to make the claim trustworthy.
  • Agentic guardrails — OWASP's Top 10 for Agentic Applications: the newest, least-settled layer.
Why the stakes moved · the numbers
Aug 2026
Article 50 applies from
EU AI Act transparency duties — machine-readable marking — go live 2 Aug 2026.
0%
RAG-grounded hallucination
Even grounded assistants still erred in 17–34% of cases · Stanford, 2025.
0
Oversight modes
From agent-assisted to human-out-of-the-loop — effort matched to risk tier.
Keeping a named human accountable for AI-assisted content isn't red tape — under Article 50's editorial carve-out, it's a route to compliance.
the line that reframes oversight for a nervous legal team
1

Map the obligations

Week 1
Timebox · ~2 daysLead: Consultant + LegalFormat: working session + desk research
Objective

Pin down what this organisation actually has to do — legally binding rules, the standards its buyers expect, industry disclosure norms, and any sector regulation — so governance is built against real obligations, not a generic template. The output is one obligations register everyone can see.

Working session agenda (2 hrs)
  • The binding floor — EU AI Act Article 50: AI-generated content must be machine-readable-marked; content that could mislead (e.g. deepfakes) clearly labelled. Note the human-editorial-responsibility carve-out and the 2 Aug 2026 application date (10m)
  • Standards buyers ask for — ISO/IEC 42001 & the NIST AI RMF GenAI Profile: voluntary, but increasingly the answer that clears a security review (20m)
  • Disclosure & provenance — the IAB AI Transparency & Disclosure Framework and C2PA content credentials (15m)
  • Sector & jurisdiction sweep — finance, health, public sector, advertising standards, regional privacy rules (40m)
  • Score each obligation: applies / partial / out of scope — with owner and evidence (25m)
  • Confirm what feeds the obligations register and who maintains it (10m)
Inputs → outputs
Inputs
  • Markets, channels & jurisdictions served
  • Existing legal / compliance policies
  • Procurement & security questionnaires received
Outputs
  • Obligations register (applies / owner / evidence)
  • Article 50 readiness note
  • Standards-to-pursue shortlist
◆ From the field

Half the room will assume the EU AI Act doesn't apply because "we're not in Europe." If the content reaches an EU audience, it usually does. Settle that question in the first hour with legal in the room, not over email three weeks later.

2

Set content risk tiers & oversight modes

Week 1–2
Timebox · ~2 daysLead: ConsultantFormat: 2-hr tiering workshop
Objective

Not every piece needs the same scrutiny. Agree the four oversight modes, then place each content type into a risk tier so human effort lands where the risk actually is — and so no one has to argue it case by case later.

The four oversight modes (least → most autonomous)
  • Agent-assisted — human drives, AI suggests. Lowest autonomy, highest control.
  • Human-in-the-loop — AI drafts, a named human approves before anything ships.
  • Human-on-the-loop — AI acts, a human monitors and can intervene.
  • Human-out-of-the-loop — fully autonomous; reserved for genuinely low-risk content only.
Explore the autonomy ladder
Workshop agenda (2 hrs)
  • Agree the risk dimensions — brand exposure, regulatory weight, factual sensitivity, reach (20m)
  • List the content types the team actually produces (25m)
  • Tier each type (low / medium / high) and bind it to an oversight mode (55m)
  • Agree escalation rules — when a piece jumps a tier (20m)
Inputs → outputs
Inputs
  • Obligations register (Step 1)
  • Content-type list from the pipeline work
Outputs
  • Content risk-tiering matrix (type × risk × mode)
  • Escalation rules
◆ From the field

Teams instinctively want to tier everything "high" to be safe. That's how you end up reviewing social captions with the same rigour as regulated claims — and the whole policy quietly stops being followed within a month. Push them to name what genuinely is low-risk; that's where the speed comes from.

Interactive · content risk heat-map

Likelihood × impact. Cells run green → amber → red by severity. Click a plotted risk (●) — or a chip — to see why it lands there and how to mitigate it.

Impact →
Likelihood →
3

Make brand & editorial rules machine-readable

Week 2–3
Timebox · ~3 daysLead: Consultant + BrandMethod: structure + test
Objective

A PDF brand book is invisible to an agent. Restructure the brand and editorial rules so an AI conforms by default — producing on-brand, compliant content the first time, instead of generating "AI slop" you then have to police.

Activities
  • Extract the rules from the brand book into explicit, testable statements — voice, tone, terminology, banned phrasings, claims that need a disclaimer
  • Encode them where agents read them — structured prompt context, a style config, or a checking agent's rule set
  • Separate hard rules (must never break — regulated claims, prohibited terms) from soft preferences (tone nudges)
  • Test against real prompts and tune until the agent conforms without hand-holding
Inputs → outputs
Inputs
  • Brand & editorial guidelines
  • Approved terminology / glossary
  • Risk-tiering matrix (Step 2)
Outputs
  • Machine-readable rule set (hard vs soft)
  • Test results & conformance notes
◆ From the field

The exercise that makes this real: take the brand book and force every rule into a sentence an agent could pass or fail. "Be confident but not arrogant" becomes a banned-words list and two worked before/after examples. The vague half of any brand book simply doesn't survive contact with a machine — and writing the rules down that plainly tends to improve them for the humans too.

4

Build approval workflows + disclosure / provenance

Week 3–4
Timebox · ~3 daysLead: Consultant + Martech
Objective

Wire the oversight modes into the CMS so review happens by default, not by goodwill — and attach the disclosure and provenance data that lets you prove, on demand, what was AI-assisted and who approved it.

Activities
  • Map each risk tier to a concrete review step in the CMS workflow (who approves, what gate blocks publish)
  • Capture an approval record — named approver, timestamp — so the Article 50 editorial carve-out is evidenced, not asserted
  • Set the disclosure policy — disclose where AI materially changed what someone sees, not a blanket "AI" stamp on everything
  • Stand up C2PA content credentials / metadata so a disclosure claim is tamper-evident and verifiable
Inputs → outputs
Inputs
  • Risk-tiering matrix & oversight modes
  • CMS workflow capabilities
  • Obligations register (disclosure rules)
Outputs
  • Tier-matched approval workflows (live in CMS)
  • Disclosure policy
  • Provenance / C2PA setup
▲ Watch out

Don't over-stamp. Labelling every asset "AI-generated" trains your audience to ignore the label and can flag content that a human genuinely authored. Disclose when AI materially changed what the reader sees — and keep the provenance trail so the claim holds up if challenged.

The approval & escalation flow
tier jumps → escalate approved rejected · rework AI draft from governed prompt Risk-tier the content type × tier matrix Oversight mode? Low-risk monitor / autonomous Named approver signs Send back rework loop Publish C2PA stamped
approved path reject / escalate conditional hover a node for detail
5

Stand up the prompt library + agent guardrails

Week 4–5
Timebox · ~2 daysLead: Consultant + Martech
Objective

Turn prompts into governed assets and put the safety net under any agent before it runs: version control, least-privilege access, and audit logs — sized against the OWASP agentic risks rather than learned the hard way.

Activities
  • Build a version-controlled prompt library — prompts as reusable, reviewed assets, not scattered through people's chat histories
  • Apply least-privilege access — scoped agent identities that can only touch what they need
  • Turn on audit logs for every agent action, so you can reconstruct what happened
  • Set guardrails against the OWASP agentic risks — goal hijack via prompt injection, memory / context poisoning, and cascading failures where one bad output ripples through an automated pipeline; require sign-off on anything high-impact
Inputs → outputs
Inputs
  • Machine-readable rule set (Step 3)
  • Oversight modes (Step 2)
  • OWASP Top 10 for Agentic Applications
Outputs
  • Governed prompt library
  • Access & audit-log configuration
  • Agentic-risk guardrail checklist
◆ From the field

Nobody has decades of practice with agentic risk yet — so don't pretend to. The honest move is sensible limits: least privilege, a human sign-off on anything that touches the public or the regulated, and a log you can actually read. Conservative now, loosened deliberately later, beats clever-and-exposed.

6

Governance body & cadence

Week 5–6
Timebox · ~1 dayLead: Consultant + Sponsor
Objective

A framework with no owner decays. Stand up a content/AI governance council with real decision rights, and a cadence light enough that people actually attend — so the stack stays alive as tools, rules and risks move.

Activities
  • Charter the council — membership (legal/compliance, brand, content, martech), remit, and explicit decision rights
  • Set the cadence — a short monthly review plus an exception path for urgent calls
  • Agree what the council owns — the obligations register, risk tiers, the prompt library, and incident review
  • Define the metrics it watches — including the honest one: AI still gets things wrong, so track factual-accuracy / hallucination rate on owned assistants
Inputs → outputs
Inputs
  • All Step 1–5 artifacts
  • Sponsor mandate
Outputs
  • Governance council charter
  • Meeting cadence & decision-rights map
  • Live framework, owned
▲ Watch out

Governance-as-shelfware is the default failure here: a polished policy PDF that nobody reads and no one owns. The test of success isn't whether the document exists — it's whether a real piece of content was held, changed, or escalated because of it last month. If nothing ever gets stopped, the framework isn't governing, it's decorating.

Good content engineering lowers the AI error rate sharply. It never makes human oversight optional.
RAG-grounded assistants still hallucinated in 17–34% of cases · Stanford, 2025
Roles & effort

RACI & effort summary

Who does what across the part. R Responsible · A Accountable · C Consulted · I Informed.

ActivitySponsorLegal / ComplianceBrandContent / OpsMartechLead consultant
Map obligationsIACCIR
Risk tiers & modesCCCCIR
Machine-readable rulesICACCR
Approval & provenanceICICRA
Prompts & guardrailsICICRA
Governance bodyACCCIR
WeekFocusConsultant days
Week 1Map obligations, start risk tiering~3
Week 2–3Finish tiering, machine-readable rules~4
Week 4–5Approval & provenance, prompts & guardrails~4.5
Week 6Governance body, cadence, handoff~2
Templates & worksheets

The artifacts you use and leave behind

Three core templates are spelled out below; the full set produced in this part is indexed at the end.

Template 1 · Obligations register

What must we comply with — and who owns it?

ObligationWhat it requiresApplies?OwnerEvidence
EU AI Act · Article 50Machine-readable marking of AI content; clear labelling where it could mislead. Human-editorial carve-out (applies 2 Aug 2026)__Legal__
ISO/IEC 42001AI management system — the credential buyers increasingly ask for__Compliance__
NIST AI RMFGenAI Profile — risk-management practices for security reviews__Compliance__
IAB disclosureTransparency & disclosure of materially AI-changed content__Brand__
C2PA provenanceTamper-evident content credentials backing disclosure claims__Martech__
Sector / jurisdictionFinance, health, public-sector, advertising or privacy rules as applicable__Legal__

Score each "applies / partial / out of scope" with evidence. The council reviews this register on its cadence — obligations move.

Template 2 · Content risk-tiering matrix

Content type × risk × oversight mode

Content typeRisk tierWhyOversight mode
Regulated / claims contentHighLegal & brand exposure; factual sensitivityHuman-in-the-loop (named approver)
Thought-leadership / long-formHighBrand voice carries reputational weightHuman-in-the-loop
Product / web copyMediumOn-brand matters; lower legal riskAgent-assisted → human review
Social captions / variantsLowLimited reach, low claim riskHuman-on-the-loop (monitored)
Internal drafts / ideationLowNot published; no external exposureHuman-out-of-the-loop (low-risk only)

Rows are illustrative — tier against this client's real content types and obligations. Add escalation rules for when a piece jumps a tier.

Template 3 · AI-content QA / audit checklist

Before this AI-assisted piece ships

Tick each control to tally pre-ship coverage. Every box must clear before the gate opens.

Pre-ship control coverage
0 / 8 · 0%
Full template index for this part
Obligations register — obligation, requirement, applies, owner, evidence (above)
Content risk-tiering matrix — type × risk × oversight mode (above)
AI disclosure policy outline — when to disclose, how, and the C2PA setup
Prompt-library standard — naming, versioning, review & reuse rules
AI-content QA / audit checklist — the pre-ship gate (above)
Oversight-mode definitions — the four modes and when each applies
Agentic-risk guardrail checklist — least-privilege, sign-off, audit logs vs OWASP risks
Approval-workflow map — risk tier → CMS review step → approver
Machine-readable brand rule set — hard rules vs soft preferences
Governance council charter — membership, remit, decision rights, cadence
Done criteria

Entry & exit gates

The quality bar that says this part is genuinely ready to start, and genuinely finished.

Before you start (entry)
  • Content model & pipeline work underway (governance has something to govern)
  • Legal/compliance and brand stakeholders engaged and available
  • CMS workflow and agent tooling access granted
Before you finish (exit)
  • Obligations register complete, owned, and signed off by legal
  • Risk tiers set and oversight modes wired into the CMS
  • Brand rules machine-readable; prompt library & guardrails live
  • Disclosure / provenance in place; governance council chartered and meeting
algomarketing
Run Book · Part 4 · Governance & Risk (v0.1) · one of seven part run books. ← back to the playbook hub