The AI Paradox in Public Affairs: Disruption Without ROI
- Paul Shotton
- Oct 10
- 4 min read
By Paul Shotton, Co-founder Advocacy Strategy
Two stories caught my attention this week.
The first, from Le Monde (7 October 2025), described how artificial intelligence is already reshaping the U.S. labour market. PwC has cut entry-level recruitment (1,300 hires in 2025 versus 1,500 in 2024), citing AI. Accenture is training more than half a million consultants but still letting go those who cannot adapt. Walmart’s CEO declared that AI will “change literally every job.” Some leaders go further: Jim Farley of Ford and Dario Amodei of Anthropic predict that half of all white-collar entry-level jobs could disappear in the next five years, with unemployment rising to 10–20%.
The second, from MIT (2025), presents almost the opposite picture: 95% of organisations report no return from their investments in generative AI. A recent McKinsey survey confirms the same pattern: 8 out of 10 companies are using AI, but 8 out of 10 also say it has had no significant impact on their bottom line. The problem is not the models but how they are used — broad, low-value applications (like meeting minutes) instead of targeted, business-critical workflows (like identifying risky suppliers or generating new product ideas).
At first glance, this seems contradictory:
AI is already destroying and reshaping jobs.
Yet most companies see little or no value from it.
In reality, these are two sides of the same coin. The disruption is real, but the difference between winners and laggards is workflows.
Firms that bolt AI onto existing processes generate little value. But when AI is embedded into the core practices of public affairs, the impact is tangible.
For junior staff — consultants, analysts, researchers, campaigners — this means the day-to-day work of drafting policy briefs, producing parliamentary monitoring reports, scanning news, and preparing first drafts of position papers. With AI, these tasks can be done faster, with higher quality, and with more room to add strategic analysis rather than just description.
For more senior or mid-level staff, the stakes are equally high. Defining priorities and objectives, planning campaigns, and developing messages and message houses all depend on good information and structured processes. AI can support these by surfacing insights, connecting patterns across issues, and stress-testing strategies.
The point is that both juniors and seniors need to map their workflows, acquire the tools, and develop the ability to use them effectively. And when the workflow is properly structured, the line between “junior” and “senior” tasks shifts: what was once considered advanced (such as integrating strategic angles into a policy brief) can increasingly be done at entry level, with AI accelerating learning and pushing younger professionals toward strategic contributions earlier in their careers.
Beyond Workflows: SOPs in Action
Here’s where organisations often stop: they experiment with AI in a few workflows, run some pilots, and leave it at that. But this ad hoc approach is why so many companies see no real ROI.
To create real value, you need to move from workflow mapping to SOPs (Standard Operating Procedures).
A workflow is the map — it shows the sequence of tasks.
An SOP is the turn-by-turn instruction manual — it explains how to perform each step consistently, with the right prompts, tools, checks, and outputs.
Take a concrete example: producing a monitoring report from an EP committee meeting.
The workflow (see downloadable slide) is simple and recognisable:
Select the topic and event.
Gather source materials.
Transcribe the video.
Clean and correct the transcript.
Draft the monitoring report.
Add analysis and recommendations.
Quality control and distribute
But the workflow alone doesn’t guarantee success. That’s where the SOP comes in.
The SOP spells out each step in detail
Which transcription tool to use (e.g. Word’s Transcribe).
How to correct names against the official MEP list.
How to prompt ChatGPT to clean technical errors and align with legislative documents.
How to structure the report into clear sections: summary of key developments, participant statements, action points, and upcoming legislative timelines.
Who reviews the draft before distribution.
The result: instead of an inconsistent transcript or a generic summary, the output is a high-quality monitoring report that’s reliable, accurate, and actionable.
Why This Matters for Public Affairs Teams
Public affairs teams depend on monitoring and reporting — but they are also the very roles most threatened by automation. AI can either hollow them out, or elevate them into higher-value contributions.
By combining workflows (the map) and SOPs (the detailed instructions), organisations can:
Ensure juniors learn by producing higher-quality outputs.
Free up seniors to focus on strategy while still having reliable intelligence inputs.
Make AI adoption consistent and repeatable across the team.
Closing Reflection
AI adoption is not about technology alone. It is about people and processes.
Workflows provide the map.
SOPs provide the step-by-step instructions.
Together, they allow organisations to integrate AI consistently, train staff at all levels, and turn experimentation into measurable impact.
At Advocacy Strategy, this is exactly the work we do with clients: helping them identify workflows, codify them into SOPs, and train their teams with practical exercises so they don’t just discuss AI — they use it in real contexts like monitoring, intelligence gathering, and campaign planning.
Because in the end, the future of public affairs won’t be decided by who has the shiniest AI tool. It will be decided by who can structure their people and processes to work smarter, deliver value, and build the next generation of public affairs professionals.
Comments