Does AI write generic content? Look at your brief
AI writes generic rubbish.
We hear this in almost every organisation we work with. And our answer is always the same: does it? Or did you give it a generic rubbish brief?
Because here’s what we’ve learned across years of content and AI work: when AI output erodes trust, the root cause is almost never the model. It’s the input. And the input is fixable — fast.
The brief most teams actually write
Let’s look at the brief that produces the “generic rubbish”: “Write a LinkedIn post about our new product.”
No audience defined. No tone of voice guidance. No objective stated. No format or length specified. The result is exactly what you’d expect: generic, off-brand, unusable output — and a team that now distrusts AI entirely.
Be honest: if you handed that same brief to a freelance copywriter, would you expect anything better? The technology was never the problem. The brief was
The brief that works
Now compare a structured brief for the same task:
- Audience: senior marketing leaders, B2B
- Objective: drive registrations for a webinar
- Tone: authoritative, direct, no jargon
- Format: three hook options, 150 words maximum
- Constraints: no competitor mentions
Same tool. Same model. Night-and-day output — a usable first draft that needs minimal human editing instead of a rewrite from scratch.
This is why we tell clients to invest in briefing standards before investing in more tools. A briefing standard costs almost nothing to build and improves every single AI output across the organisation. Another tool licence improves nothing if the briefs feeding it stay broken. Briefing is implementation at its most granular.
The question almost nobody asks: who owns the decision?
There’s a second discipline that separates trustworthy AI content systems from risky ones, and it’s this: clarity about who — or what — owns each decision.
Here is the split that works in practice, mapped across all eight stages of a content system.
| Stage | AI role | Human role |
|---|---|---|
| 1. Strategy | Synthesise insights, analyse trends, surface opportunities | Set direction, priorities and point of view |
| 2. Ideation | Generate angles, campaign ideas, content territories | Choose the strongest, most relevant ideas |
| 3. Briefing | Turn strategy into structured briefs and prompts | Approve audience, message, proof points and risk |
| 4. Creation | Draft, adapt, repurpose, create variants | Edit, sharpen, apply judgement |
| 5. Review | Check tone, consistency, claims, SEO and readability | Own quality, compliance, reputation and final decisions |
| 6. Publication | Adapt by channel, format and audience | Approve final version and timing |
| 7. Measurement | Summarise performance patterns and insights | Interpret results and decide what changes |
| 8. Learning loop | Feed insights into future briefs and content planning | Improve the system, not just the asset |
Notice the pattern: at every single stage, AI does the heavy lifting and a human owns the call. Accountability cannot be automated away — and the moment it’s defined this explicitly, teams stop hesitating and start shipping.
The point isn’t to limit AI. It’s to make AI safe to use at speed.
And here’s the blunt part: if you can’t currently say who owns each decision in your content process, AI didn’t create that problem. It just switched the lights on.
Why does this matter more as you scale
A weak brief in the hands of one person produces one bad draft. A weak briefing culture in an organisation running AI across marketing, comms and sales produces hundreds of off-brand assets a month — each one a small bet against your brand, some of which will reach customers.
This is the amplification effect we describe across this blog: AI doesn’t create content maturity, it amplifies whatever maturity — or dysfunction — already exists. Briefing standards and decision rights are two of the cheapest, fastest ways to make sure what gets amplified is the good stuff.
Both belong inside a designed workflow — what we call the Content Operating System. If you haven’t read that piece, start there; it’s the frame everything else hangs off.
Fix the inputs first
Before your next tool evaluation, run this test: take your five most recent AI briefs and score them against the five elements above — audience, objective, tone, format, constraints. Most teams score one or two out of five. That score, not your tool stack, is your real AI maturity.
Want help building briefing standards and decision frameworks your teams will actually use? That’s a core part of what we do at Ubi — practical, unglamorous, and where the ROI hides.
30 minutes. Zero cost. Let’s talk.
❓ FAQ
Q: Why does AI produce generic content?
A: Most often because the brief lacks the five essentials: defined audience, objective, tone of voice, format, and constraints. Structured briefs produce usable first drafts; vague briefs produce generic output regardless of the tool
Q: What should a good AI content brief include?
A: Audience, objective, tone, format (including length), and constraints. Treat AI like a capable freelancer: it performs to the quality of the brief.
Q: Which content decisions should humans always own?
A: At every stage: direction and priorities (strategy), idea selection (ideation), approval of audience, message and risk (briefing), editorial judgement (creation), quality and compliance (review), final sign-off and timing (publication), and interpretation of results (measurement). AI supports; humans decide.
