Does your organization's AI still sound authentic?
Social impact and sustainability teams have never had more to say or less time to say it. Impact reports, employee campaigns, partner updates, social posts, and investor narratives all compete for the same small team. AI promises to close that gap, and in many ways it can.
But there is a catch. Nearly every company is drawing on the same few models. According to Menlo Ventures, Anthropic, OpenAI, and Google together accounted for about 77% of enterprise LLM usage in mid-2025. When everyone uses the same tools the same way, the content starts to sound the same.
For purpose-led organizations, that sameness is a real risk. Your purpose is what makes you distinct. If your communications lose that voice, stakeholders notice (even if your dashboards don't).
Where AI can amplify purpose
Used with intent, AI can help purpose teams do more of what they already do well:
Turn impact data into stories. Program metrics, grantee reports, and volunteer hours become narratives people want to read.
Reach each stakeholder in their language. One impact story can be tailored for employees, investors, community partners, and customers without starting from scratch each time.
Speed up reporting. Drafting, summarizing, and formatting take less time, which frees people for the judgment calls that matter.
Pedigree's Adoptable campaign is a great example, Created by Colenso BBDO and Nexus Studios, machine learning turned amateur shelter photos into ad-ready images, so real local shelter dogs could appear in Pedigree's own media placements. Ads updated as dogs were adopted. The results in test markets: 50% of featured dogs were adopted within two weeks, and shelters saw a sixfold increase in visitors. The campaign won the Outdoor Grand Prix at Cannes Lions in 2024.
Note what made it work. The technology did not invent Pedigree's purpose. It carried a long-standing commitment to shelter dogs into every media dollar. Purpose came first; the AI simply amplified it.
Where it goes wrong: purpose drift
Most AI failures in communications won't look like a scandal. They will look like drift: a slow slide away from what an organization stands for. Each AI-assisted draft seems fine on its own. Over months, the voice gets flatter, the stories get more generic, and the claims get a little further from the evidence.
Part of the reason is human. In a Microsoft Research and Carnegie Mellon study of 319 knowledge workers and 936 real examples, higher confidence in generative AI was linked to less critical thinking. The more people trust these tools, the less they push back.
The financial stakes are real. EY's Responsible AI Pulse survey of 975 C-suite leaders found that 99% of companies had taken financial losses from AI-related incidents, averaging an estimated $4.4 million. Companies with real-time monitoring and oversight committees were more likely to report revenue growth and cost savings from AI. Communications is where drift becomes visible first, because it is the part of the business outsiders actually see.
Talking about your own AI use
The second job for purpose communicators is explaining how the organization uses AI at all. Right now, few companies are doing it.
JUST Capital's Spring 2026 research found that only 37% of 110 companies analyzed disclose responsible AI principles or guidelines. Meanwhile, the public is warming to AI but with conditions: 66% expect AI to benefit society within five years, and 47% name safety and security as their top concern.
That gap provides an opportunity. Companies that explain what they use AI for, why, and where humans still make the call can earn trust before it is demanded. The leaders are already moving: PwC's 2026 AI Performance Study of 1,217 executives found that 20% of companies capture about 74% of AI's economic value, and those leaders are 1.7 times more likely to have a responsible AI framework.
The lesson is familiar. In our B2B Purpose Paradox study with ANA and The Harris Poll, 86% of B2B leaders said purpose mattered to growth, but only 24% said it was embedded deeply enough to shape operations. AI is the same test in a new form. Saying you use it responsibly is easy; showing how is what builds trust.
Four checks for purpose communicators
You don't need a new framework to keep AI-assisted communications on course. Start with four questions:
The resonance test. Read it out loud. Does it still sound like you, or could any company in your sector have published it?
The override count. When did someone on your team last look at an AI draft, say "that's not us," and change it? If no one can remember, humans aren't really in the loop.
The evidence test. Does every impact claim trace back to a real number, a real program, or a real person? Polished visuals without data to back them up are drift in disguise.
The disclosure test. Could a stakeholder find out, in plain language, how you use AI and where people stay in charge?
None of these checks slow the work down much (especially when you factor in the efficiency gained by using AI tools in the first place). They add a small amount of friction on purpose, and that friction is what keeps your voice your own.
Comments