STB Creative Solutions — AI Policy

• Human-led outcomes: AI supports our work; it does not replace strategic thinking, empathy, and experience.

• Client safety first: confidentiality, security, and brand governance take priority over speed.

• Accuracy over automation: we verify facts and avoid unverified claims or fabricated sources.

• Originality & value: We aim for distinctive positioning and useful insights, not generic AI text.

• Transparent collaboration: clients may request disclosure and/or restrictions on AI use.

Generative AI: tools that generate or transform text, images, audio, video, or code.

Agentic workflows: AI-enabled systems that can plan and execute multi-step tasks (e.g., research → draft → optimize → report).

Client Confidential Information: any non-public client information (data, strategies, internal documents, credentials, financials, customer lists, etc.).

AI-Assist Standard: AI may be used for research, ideation, drafting, and editing with strict human review and confidentiality controls.

AI-Restricted: AI may be used only for non-confidential tasks (e.g., public research, outline ideation), with client materials excluded.

Human-Only: no generative AI is used in producing client deliverables (except basic non-generative tools such as spellcheck, if permitted by client policy).

Audience, market, and competitor research based on public sources

Topic discovery and question mapping to support AI search visibility (answer-driven content)

Synthesizing complex industry information into clear insights—then verified by humans
 

Brainstorming headlines, angles, hooks, and campaign themes

Drafting content briefs, outlines, editorial calendars, and repurposing plans

Structuring content to improve clarity, scannability, and decision-maker experience

Generating rough drafts for blogs, landing pages, LinkedIn articles, newsletters, scripts, and frameworks for human refinement

Creating alternative versions for A/B testing, segment-based personalization, and platform repurposing
 

Grammar, tone alignment, summarization, readability improvements, and structural editing—followed by human review

Consistency checks for terminology, voice, and formatting

Transcribing interviews, webinars, and recorded calls (with permission where required)

Extracting key themes, quotes, and insights for case studies, white papers, and thought leadership
 

Role-playing: simulating how a prospect might interpret messaging to reduce confusion and improve conversion clarity

Objection handling: testing claims and benefits language against common buyer concerns

Formatting content for direct answers, FAQs, and structured sections suitable for AI-powered discovery

Drafting schema-ready content blocks and supportive microcopy (implementation may require developer approval)
 

Using AI to assist with workflow design, content ops automation, and reporting—under defined governance controls

Prototyping agentic workflows to improve speed and consistency while maintaining human approval points

• Data minimization: prompts include only what is necessary to complete the task

• No sensitive uploads by default: we do not paste/upload confidential documents, credentials, customer data, or internal reports into third-party AI tools unless the client explicitly approves

• Redaction/anonymization: where possible, we remove names, identifiers, and proprietary specifics before AI assistance is applied

• Access control: only authorized personnel may use AI tools for client work; credentials must never be shared in prompts

• Retention awareness: We consider vendor terms and data retention policies when selecting tools and workflows

• Publish AI-generated content without human review and editing

• Share client proprietary or confidential data with AI tools without explicit consent

• Fabricate case studies, testimonials, performance results, customer quotes, references, or citations

• Use AI output as the sole basis for compliance-sensitive decisions or regulated claims

• Bypass security controls, client governance rules, or the tool's terms of service
   

• We can provide an AI usage summary for a project (what tools were used and for which tasks)

• We will highlight any areas where assumptions were made, and what was verified versus inferred

• If a client requires “Human-Only” work, we will confirm that AI was not used (as defined in Section 3)

• Client deliverables are treated as client work product under the applicable engagement agreement.

• We avoid plagiarism and ensure content is sufficiently original and brand-specific

• We cite/confirm sources for data, quotes, and factual claims included in content.
   

• Tool governance: STB maintains an internal list of approved AI tools and acceptable use settings for client work

• Training: team members using AI tools must understand prompting hygiene, confidentiality rules, and fact-check requirements

• Continuous improvement: this policy may be updated as AI capabilities, vendor terms, and client expectations evolve

• Human-in-the-loop review: a qualified STB strategist/editor reviews every deliverable before publishing or client submission

• Fact-checking: we confirm key statistics, references, and claims using reliable sources; AI output is not treated as a primary source

• Hallucination controls: if a claim cannot be verified, it is removed, rewritten as opinion, or clearly labeled as an assumption (with client approval)

• Compliance awareness: we avoid regulated claims and ensure disclaimers/approval steps are followed when required by client sector policies

• Originality: We do not copy competitors or reuse copyrighted material; we prioritize original frameworks and client-specific differentiation