Menu
A lot of people can draft now.
That used to be the bottleneck. It isn’t anymore. The bottleneck moved. Ten people open the same tool, write about the same brand, and produce ten slightly different versions of “us.”
Tone drifts. Claims wander. One person sounds warm. Another sounds legal. Another sounds like a tool demo. Reviewers spend their time reconciling voices instead of improving the work.
That isn’t mostly a model problem. It’s the brand-voice lottery: AI didn’t create the inconsistency. It inherited it, then scaled it across more hands.
It looks busy. It looks productive. It looks uneven under a second read.
From the hallway, output volume can look like progress. From the review queue, it looks like ten brands sharing one logo.
People rarely resist the tool. They resist what happens when the tool makes every draft someone’s personal version of the company.
We keep seeing the same picture without needing a named account to make the point.
A marketing or content-led team finally has AI in enough hands to matter. There is no single owner of finished copy. Tone guidelines are light. Someone describes the week as ten emails from ten different people, each one almost right and none of them the same. Leadership wants a brand-voice repository or library that survives the lottery: a shared source the team can keep, not another tip deck that disappears after the workshop.
That is one enablement problem with a content surface. You don’t need a louder writing tool first. You need a shared voice and source path people actually use while they draft.
It presents as a writing-tool problem. It usually isn’t.
A writing tool can generate drafts faster. It cannot invent a shared definition of voice, claims, sources, and “done” if the organization never owned those things in one place. When knowledge is scattered and ownership is unclear, more drafters make the scatter visible.
Enablement here means:
If the stuck work is content and knowledge-base consistency, fix that path. Do not buy another blank box and hope tone appears.
Skip the “everyone invents their own prompt” phase.
A useful shared library is not a brand book nobody opens. It is operating material people can use while drafting:
That is content nested inside enablement: same Foundations logic as any other everyday workflow, applied to voice and knowledge.
Consistency and review will still matter. Review is where quality gets tested. It is not where quality starts. Quality starts when the team shares a way of working before the draft hits the queue.
If many people can draft and the brand still sounds like a lottery, do not start with another writing-tool rollout.
Name the stuck work. If it is content, voice, or knowledge-base consistency, treat that as one everyday path: shared library, shared practice, clear owner, light support after the first session. Measure whether drafts arrive closer to “us,” not whether more people clicked generate.
If content and knowledge-base are not the stuck work, don’t force a content module. Go back to the everyday workflow that actually isn’t moving.
What does inconsistent AI voice look like across a team?
It looks like volume without sameness: same offer phrased ten ways, guidelines ignored at draft time, private prompt tricks for power users, and reviewers reconciling brand instead of improving work. Output can look busy from the hallway and uneven under a second read. AI scaled the inconsistency that was already there.
What belongs in a shared voice and source library?
Operating material people use while drafting: concrete voice examples, approved sources and claims, prompt patterns that survived review, a clear review path, and an owner who keeps the library alive. A brand PDF nobody opens is not a library. A keepable path the team still has when the first expert is out is.
Is this a writing-tool problem or an enablement problem?
Usually enablement. Writing tools accelerate drafts. They do not create shared voice, ownership, or source truth if those were never owned. When many people draft without a shared path, the lottery gets louder. Fix the way of working on content and knowledge first; choose tools to support that path, not the other way around.
If a lot of people can draft and the brand still sounds like a lottery, start with the free AI Operations Reality Check. One everyday content or knowledge path. No pitch deck.
If you want a bounded diagnostic and a written next-step memo, the AI Operations Review is the paid step after that ($997, not a retainer, not another writing-tool rollout).
When the stuck work is clearly content, voice, or knowledge-base consistency, the Pragmatic Content Engine is the module for that path after Review names it. It is not the cold front door for this problem.
Susan Westwater is the CEO and Co-Founder of Pragmatic Digital. She helps mid-market and PE-backed teams move from scattered AI pilots to governed, measurable workflows that actually deliver operating leverage. With 25+ years in CX and brand leadership at Leo Burnett and Ricoh USA, Susan specializes in turning AI ambition into repeatable systems that protect brand voice and reduce revision cycles. She is co-author of Voice Strategy and Voice Marketing.
Weekly insights on AI adoption, workflow, and what's actually working in mid-market organizations.