policy-recommendation

When to use

• One recommendation, put to lawmakers. When a brief must land a single policy ask — adopt, amend, oppose, or defer — with the evidence that earns it and the exact legislative move to make. The stance badge is the verdict; the ask is the call to action.
• The stance sets the color. Pick the variant by the posture: adopt (support, green), amend (conditional, amber), oppose (against, red), defer (study first, neutral). The badge, rail, and ask bar all take the stance color, so the reader sees the position before reading a word.
• Reasons carry a citation. Each reason leads with a bold claim and a nested evidence line ending in an inline-code citation chip. Without the chip the reason reads as opinion; with it, it earns its place in a legislative record.

When not to use

• Weighing two options. If the slide compares alternatives before choosing, use split-compare — its right zone is a 2-option grid plus a verdict card. policy-recommendation states one already-chosen position.
• More than four reasons. Past four the panel reads as a memo and the ask loses force. Keep the three strongest reasons here and move the full evidence to list-criteria.
• A recommendation with no ask. Omitting the closing blockquote leaves the reader with a position but no action. Always name the specific legislative move — the bill, the section, the vote.
• A topic-label heading. ## Breach Notification is a topic, not a recommendation. Say the action: ## Adopt a 30-day breach-notification deadline.

Common mistakes

• Writing a reason's evidence line without a trailing inline-code citation chip. Without the trailing citation chip, the evidence line reads as unsupported assertion rather than record-grade evidence — the nested evidence bullet should end in an inline-code citation, not just prose.
• Placing the bill/docket eyebrow after the heading instead of before it, or leaving it unwrapped in backticks. A code-only paragraph immediately before the heading is lifted into the masthead and picked up by the shared mono-caps eyebrow rule — it must be the section's first line, wrapped in backticks, or it either stays a plain paragraph (if unwrapped) or becomes an italic subtitle instead (if placed after the heading).

Authoring

<!-- _class: policy-recommendation adopt -->

`Bill reference · Domain`

## The recommendation as a complete sentence.

One line naming the problem or the stakes.

- First reason
  - The evidence for it, with a `Citation` chip.
- Second reason
  - The evidence for it, with a `Citation` chip.
- Third reason
  - The evidence for it, with a `Citation` chip.

> The specific legislative ask — sponsor, vote, or amend, with the section.

Slots

Variant decision rule

• adopt — The recommendation is in support of the measure — green verdict badge and rail.
• amend — Support is conditional on a specific change — amber badge and rail.
• oppose — The recommendation is against the measure — red badge and rail.
• defer — The evidence base is incomplete and the ask is to study first, not decide yet — neutral badge and rail.

Variants

<!-- _class: policy-recommendation adopt -->

`HB 214 · Consumer Data Protection`

## Adopt a 30-day deadline for breach notification.

Consumers now learn of a breach 78 days late on average.

- Faster notice cuts identity-theft losses
  - 22% lower per-victim costs where the rule exists `FTC 2025`.
- Already industry practice
  - 14 of the top 20 processors notify within 30 days `IAPP`.

> Vote YES on HB 214 § 3.

<!-- _class: policy-recommendation amend -->

`SB 88 · Algorithmic Hiring`

## Amend SB 88 to require bias audits before deployment.

Post-hoc audits catch discrimination only after applicants are screened out.

- Pre-deployment testing is preventive
  - Later audits cannot un-reject filtered candidates `EEOC 2025`.
- The change is one clause
  - Move the trigger from § 4(b) to § 3 `Leg. counsel`.

> Adopt the § 4(b) to § 3 move in the committee substitute.

<!-- _class: policy-recommendation oppose -->

`AB 402 · Encryption Backdoors`

## Oppose AB 402's mandated key-escrow provision.

A state-held decryption key is a single point of failure for every resident's data.

- It weakens security for everyone
  - A mandated backdoor is exploitable by any attacker who finds it `NIST`.
- It won't stop determined actors
  - Open-source encryption stays available outside the state's reach `EFF`.

> Vote NO on AB 402, or strike the § 5 escrow mandate.

<!-- _class: policy-recommendation defer -->

`HR 19 · Municipal Facial Recognition`

## Defer HR 19 pending an independent accuracy study.

Error rates for the proposed system are unpublished for the populations it would scan.

- The evidence base is missing
  - No demographic accuracy audit exists for this vendor `GAO request`.
- A study is already funded
  - The oversight office can report within one session `fiscal note`.

> Refer HR 19 to interim study; revisit with the audit in hand.

Related

• split-comparetwo options weighed before a verdict card
• decisionnaming a chosen path among options already presented
• list-criteriaa flat enumeration of requirements without a stance or an ask
• regulatory-updatea period digest of what changed, not a recommendation on one measure