We like to think we’ve mastered AI in communications.
We prompt, refine, and generate. We use ChatGPT for summaries, Copilot for drafts, and maybe whisper, half-jokingly, that the machine writes better than we do on a Friday afternoon.
But here’s the truth: we’re only using about ten percent of what these tools can do.
Why we only use a fraction of AI’s potential
Large language models are extraordinary at pattern recognition, yet they lack the one thing communicators excel at: context. Without it, even the most sophisticated AI will sound generic, forget nuance, and miss strategic intent.
Most communicators treat ChatGPT or Copilot as a universal Swiss Army knife: good enough for quick rewrites, idea generation, or first drafts. But that’s like buying a Stradivarius and never tuning it.
Custom versions — Custom GPTs in OpenAI’s ecosystem or Copilot Studio agents in Microsoft’s — are where the real transformation begins. They allow us to encode what makes our communications practice unique: our tone, our processes, our brand voice, and our ethical compass.
How to build your custom version
Think of it as training a new team member: one who learns by example.
You start by identifying the most repetitive, time-consuming tasks: writing media pitches, summarizing reports, refining tone for press releases, or proofreading multilingual content. Then you ask: what would excellence look like if AI could do this with us, not for us?
The process unfolds in layers:
- Define the purpose — one clear task per agent: proofreading, drafting press releases, creating social media posts, or maintaining tone consistency.
- Curate the inputs — feed it your best work: successful press releases, past campaigns, tone-of-voice guidelines, brand manuals. These become the training ground.
- Write the “constitution” — precise instructions and ethical boundaries. Include editorial rules, disclaimers, and when the agent should stop or seek human validation.
- Iterate and refine — test, fail, adjust. As in any creative process, clarity emerges through repetition.
The result? An assistant that remembers what good looks like, mirrors your standards, and works at a pace no human team could sustain, without losing quality or integrity.
When customization becomes strategy: BCG case study
At the Boston Consulting Group, this approach has already taken root.
BCG began by analyzing over 300 client organizations to identify where AI could create the most value. Communications ranked among the top three functions with customer service and HR; a finding that led them to pilot AI tools within their own corporate affairs team.
Their communications team began experimenting with tailored Custom GPTs to support the development of press releases and media pitches.
They started by pinpointing where most time was lost: the five to six hours spent turning complex insights from new BCG reports into polished, media-ready content. Then came the groundwork: collecting their best-performing releases and outreach emails, distilling editorial and strategic guidelines, and feeding these into their CustomGPT’s framework.
After several months of testing and iteration, the result was remarkable: the time to craft a press release dropped from six hours to one, with consistency, clarity, and tone aligned across global offices.
It’s not just about efficiency. It’s about embedding institutional memory into tools that scale.
The core of future AI integration for comms teams
Context engineering — as Anthropic beautifully illustrates in their piece on effective context engineering for AI agents — is becoming a discipline in itself. It’s about shaping the conditions under which an AI interprets information and makes decisions. For communications leaders, this means codifying tone, ethics, and strategy directly into the systems that help create content.
Imagine:
- A press release creator aligned with corporate tone and approval workflows.
- A brand voice guardian ensuring every LinkedIn post echoes the same voice of integrity and confidence.
- A proofreading assistant fluent in your internal style guide and sensitive to inclusive language.
- A social content strategist who repurposes key messages across channels without losing meaning.
These are not futuristic scenarios: they are achievable now, with tools like Custom GPTs and Copilot Studio.
When designed intentionally, these systems become more than writing aids. They become knowledge engines that preserve your team’s collective intelligence, accelerate production, and uphold quality in a landscape that rewards speed but punishes superficiality.
The ethics of amplification
Customizing AI is not merely a technical exercise: it’s an editorial act. The datasets you choose, the examples you feed, the tone you enforce — all shape how your organization communicates truth, empathy, and credibility.
This is where communications leaders must step in: to set guardrails, define what “responsible voice” means, and ensure that every AI-generated line aligns with organizational values.
The opportunity before us
In BCG’s example lies a lesson for every communications department: AI will not replace your teams — but teams that know how to train their AI will outpace those that don’t.
The future of communications belongs to those who design their own assistants, rather than merely using generic ones.
At AI Comms Factory, we see this as the next frontier: making these models truly ours — ethical, strategic, and unmistakably human in intent.
So, the next time you open ChatGPT or Copilot, ask yourself:
what if this wasn’t just a tool, but a colleague trained on the very best of your craft?