Our position
We don’t just use AI. We govern it. Modern Impact runs a proprietary AI governance system — our own standards, source controls, and validation rules — that directs how AI is used on every engagement. It works under human direction, and we’re open about when and how. Every deliverable we hand you is owned, reviewed, and accountable to a named person at Modern Impact. The system accelerates our work. It does not replace the judgment you hire us for.
Where we use AI
We use AI in two ways: as an assistant that helps our team work faster, and increasingly as agents that carry out multi-step tasks under our direction.
- Agentic workflows: AI that carries out multi-step work across our tools and systems — retrieving, organizing, producing, and routing — so our people spend their time on judgment, not busywork.
- Research and discovery: synthesizing market, audience, and competitive inputs faster than manual review allows.
- Drafting and iteration: first-pass copy, design exploration, and structure that our team then shapes.
- Quality and analysis: usability review, data pattern-finding, and QA passes on our own work.
A person directs the work, sets the limits on what AI can access and do, and stays accountable for the outcome. No agent ships work to you, or acts on it, without human review.
Where we do not use AI
- We do not ship AI output as finished work without human review and revision.
- We do not put your confidential information into tools that fall outside our data agreements.
- We do not use AI to fabricate research, testimonials, data, or results.
How we govern our AI use
Governance is the part clients rarely see and the part that matters most. We built ours as a working discipline, not a policy PDF. The specifics are proprietary, but here is what they guarantee you:
- Written standards, not habits. Documented SOPs and protocols define how tools are selected, how they may be used, and how output is reviewed. The rules hold the same across every account and every person on the team.
- Controlled source authority. AI on your work draws on approved, MI-maintained reference material — not open-web guesswork. What the model is allowed to treat as fact is something we curate and control.
- Source attribution on every input. We track where each input came from: what you gave us, what our own reference material supplied, and what the model generated. That trace is what lets a named lead stand behind the result.
- Validation before hand-off. Structured validation rules and review gates have to pass before AI-assisted work moves toward you. Missing or unverified information gets flagged, never filled with a confident guess.
- A named lead on everything. One accountable person at Modern Impact owns each deliverable, AI-assisted or not. Our AI Governance Pilot maintains these standards across accounts.
On AI detection and watermarking
As of August 2, 2026, some AI providers embed invisible watermarks in the text their models generate. One point matters more than the rest: a watermark signals that content passed through an AI tool, not that AI authored it. When our team uses an AI tool to proofread, translate, or tighten copy a person wrote, that human work can still carry the mark.
We raise this so a detection result never catches you off guard. Our commitment is simpler than any detector: we disclose how AI was used on request, for any engagement, and we stand behind the human judgment in everything we deliver.
If you want to know how AI was used on your work, ask. We will answer plainly.