Use of artificial intelligence
How the collective uses AI, where it is never allowed, and who is accountable for the work. This is the same policy our own team signs.
Under no circumstances may any of the following be entered into any AI tool — approved or otherwise — without explicit written client authorization: client data, client-identifying information, regulated health information, financial records, personally identifiable information (PII), proprietary strategy documents, product specifications or technical information, M&A or transaction-related information, personnel information, or any other confidential information. AI tools may not be connected to or integrated with any client systems. This is a data security and confidentiality requirement, not a preference. Violation is grounds for immediate termination.
If a client contract, SOW, or master services agreement prohibits the use of AI tools, that prohibition is absolute. AI may not be used on that account in any form. Before beginning work on any client account, the Account Lead is responsible for reviewing all agreements for AI restrictions and communicating them to the full team in writing. When in doubt, do not use AI — confirm first.
Our position on artificial intelligence
AI is changing marketing — how work gets done, how buying decisions are made, and what clients expect from their partners. At JXT, we’re not standing on the sideline.
AI runs through our operations end to end. It’s how we share knowledge across the collective, maintain quality standards, and keep the delivery model lean. It gives every team member access to senior-level frameworks. It handles coordination and overhead so experienced people can focus on the thinking and the work that actually matters.
But AI doesn’t do the thinking. That’s yours. The insight, the idea, the strategic direction — those have to be human. Knowing what you’re trying to achieve and how to get there is your job. AI enables you to get there faster and at a higher level of quality. The expectation at JXT is that every person brings the judgment and every tool amplifies it.
Our clients operate in complex, regulated industries where accuracy, compliance, and human judgment are non-negotiable. AI enhances our work. It does not replace our expertise.
What’s expected of every team member
- Develop working fluency with approved AI tools — this is part of competency at JXT.
- Use AI to increase efficiency and quality, not to replace thinking. The insight, the idea, the strategic direction — those are yours. AI helps you get there faster and at a higher level of quality.
- Apply human review, judgment, and editing to all AI-generated content before it is used. AI assists; it does not decide.
- Stay current on how AI is evolving in marketing and in the regulated industries we serve. What’s changing is part of the job.
- Bring AI-informed insights into client recommendations where relevant and appropriate for the client’s regulatory environment.
Approved tools
Use only AI tools approved by JXT management. Current approved tools include Claude (Anthropic) and other tools designated by the Operations team. If you want to use a tool not on the approved list, request approval from your manager before use. The approved tool list is maintained by Operations and updated as the landscape evolves.
What’s allowed
- Drafting, brainstorming, outlining, and summarizing non-confidential content
- Research support and competitive landscape analysis
- Content editing, optimization, and repurposing
- Proofreading and quality review against brand guidelines, style guides, client standards, and approved AI skills
- Formatting content into templates and standard deliverable formats
- Internal process documentation and template development
- Data analysis and reporting — with human validation before use
- Meeting prep and follow-up — agendas, notes, and action item capture
- Summarization of internal meetings or recordings (non-confidential only)
- Administrative and operational tasks: scheduling, formatting, organization
- Learning and professional development — exploring tools, building knowledge, staying current on industry shifts
What’s not allowed
- Uploading confidential client data, personal information, or proprietary business information into any AI tool not approved for that use
- Submitting AI-generated content to clients as final without thorough human review and editing
- Using AI in ways that violate client agreements — if a client prohibits AI use, do not use it
- Using AI to generate content that is misleading, discriminatory, or non-compliant with applicable regulations
- Representing AI-generated work as original human work when disclosure is required
- Using AI to produce content on a subject you have not read and do not understand. AI overuse shows. It produces generic, shallow output that demonstrates unfamiliarity with the topic. You are expected to understand the subject matter you’re writing about. Read the brief. Know the client’s business. Understand the industry. AI can help you write faster and better, but it cannot substitute for knowing what you’re talking about. Submitting AI output you haven’t genuinely reviewed is a quality failure, not a time saver.
Client work requirements
- Review client contracts and SOWs for any AI restrictions before using AI tools on that account
- Document AI tool usage in project notes when required by client agreement
- When in doubt about whether AI use is appropriate for a specific client, ask your Account Lead
Human accountability
You are accountable for everything that goes out under the JXT name, regardless of whether AI assisted in creating it. Factual accuracy, brand alignment, compliance, and quality are your responsibility. AI is a starting point, not a finish line.
Violations
Misuse of AI tools — including unauthorized uploads of client data, bypassing review processes, or violating client restrictions — may result in corrective action up to and including termination.
FROM THE QUESTION DESK
AI at JXT, answered straight
Does JXT Collective use AI?
Yes, openly. AI does the heavy lifting where it is good at it, and a human reviews, corrects, and signs everything before it ships. The policy our own team signs is public on this page, which is the point. AI-powered, human-signed is how the house works, in writing.
Will AI write my content?
AI may draft. It never publishes. Every deliverable is edited and signed by a named person, and nothing leaves the house unsigned. That trail matters in regulated industries, where you have to answer for what went out and who approved it.
How does JXT keep AI from getting things wrong?
Two ways at once. We build our standards into the AI itself, so the house voice, editorial rules, and compliance guardrails are embedded as skills the AI works under and drafts start closer to right. And we treat its output as a fast junior’s, never an authority’s. Facts are verified against sources, claims trace to evidence, and a named reviewer signs the work before it ships.
Why publish your AI policy?
Because clients in regulated industries have to answer for their vendors. Publishing the actual policy, not a summary, lets you read exactly what we do and hold us to it. It is the same document our own team signs, word for word.