The lobbyist's briefcase is about to become a museum piece.

Photo by Social Estate on Unsplash

There’s a particular ritual in government relations that I’ve watched play out hundreds of times. A well-dressed person walks into a parliamentary office, opens a leather folio, and delivers a set of talking points that were drafted by committee, approved by legal, softened by comms, and stripped of anything resembling a genuine human perspective. The staffer nods politely. The meeting ends. Everyone files it under “stakeholder engagement” and moves on.

That ritual is about to be eaten alive. And the industry performing it has almost no idea.

The government relations profession in Australia — and frankly, most of the Western world — is sleepwalking into an era where the core mechanics of what it does can be replicated, accelerated, and in many cases outperformed by artificial intelligence. Not in some distant, hand-wavy future. Now. Today. With tools that already exist.

And the profession’s response? Largely silence. Or worse, the kind of performative curiosity that involves attending a single AI panel at a conference, nodding sagely, and then going back to doing things exactly as they were done in 2019.

The uncomfortable truth nobody wants to say out loud.

Here it is, plainly. Most of what passes for government relations work is information synthesis, stakeholder mapping, and templated persuasion. Draft a submission. Identify the relevant decision-makers. Map the political dynamics. Monitor legislative developments. Prepare briefing notes.

Every single one of those tasks is now within the capability of a well-configured AI system.

Not perfectly. Not without human judgment. But well enough that the value proposition of a mid-tier GR consultancy — the one that charges a retainer to “monitor and advise” — is eroding underneath their feet while they’re busy updating their CRM.

The industry has spent decades building its moat on access and relationships. Knowing who to call. Having the minister’s mobile. Getting the meeting that others can’t. And look, access still matters. It always will, to a point. But access without insight is just an expensive coffee. And insight is exactly the thing AI is getting terrifyingly good at generating.

An industry allergic to self-examination.

Government relations has always had a peculiar relationship with introspection. It’s an industry that advises others on change while being remarkably resistant to it internally. The tools haven’t changed meaningfully in decades. The same submissions process. The same stakeholder matrices drawn on whiteboards (now Miro boards, which apparently counts as innovation). The same post-budget analysis that reads like it was written by a committee — because it was.

Part of this is structural. GR firms sell certainty to anxious clients who want to believe that navigating government is an arcane art best left to initiated priests. Admitting that a significant portion of the craft is systematisable would undermine the mystique. And the mystique is what justifies the retainer.

But here’s the thing. The clients are going to figure it out. Some already have.

I’ve watched in-house government affairs teams at major corporates start experimenting with AI for policy analysis, submission drafting, and even political sentiment mapping. They’re not waiting for their external advisors to catch up. They’re routing around them. Quietly, without fanfare, and without telling their GR firm that the fortnightly briefing note they’re paying for is now something an internal analyst can generate in twenty minutes.

Where the real danger lives

The danger isn’t that AI will replace lobbyists wholesale. It won’t. The handshake in the corridor, the ability to read a room, the instinct for when a minister is genuinely persuadable versus performing openness — those things remain stubbornly human. For now.

The danger is more surgical than that. It’s the hollowing out of the middle.

Think of it as a barbell. On one end, you have the genuine strategists — people who understand power, who can read political dynamics three moves ahead, who know that the real game is often being played two levels beneath the surface everyone can see. AI can’t do that yet. It lacks the contextual intuition, the relational memory, the ability to sense when something has shifted in a caucus room before anyone says a word.

On the other end, you have the purely administrative functions that AI can automate tomorrow. Hansard monitoring. Regulatory tracking. Stakeholder database management. These are already being done better and faster by machines.

The middle — the vast, comfortable middle where most GR professionals live — is the kill zone. The people who do a bit of analysis, a bit of relationship management, a bit of writing, but none of it at a level that’s irreplaceable. The ones who’ve coasted on being “across the brief” without ever being the person who actually shaped the strategy.

That middle is about to compress violently. And the industry isn’t talking about it.

The skills gap nobody is addressing

Ask a government relations professional what their technology stack looks like and you’ll get a blank stare, a mention of Outlook, and possibly a reference to some monitoring platform they never fully learned to use. Ask them about prompt engineering, structured data analysis, or how to validate AI-generated policy research, and the conversation tends to end quickly.

This isn’t a criticism of individual capability. Most GR professionals are smart, politically literate people who chose this career because they care about policy outcomes. But the industry has done almost nothing to prepare its workforce for a world where their core deliverables can be produced faster, cheaper, and at scale by tools their clients can access directly.

There is no serious, industry-wide conversation about reskilling. No professional development framework that treats AI literacy as a baseline competency. No honest reckoning with the fact that “I know someone in the minister’s office” is a depreciating asset in a world where transparency registers are expanding, revolving door scrutiny is increasing, and the access-for-access-sake model is losing its social licence.

What the industry should actually be doing

The firms that will thrive in this environment won’t be the ones who bolt a ChatGPT subscription onto their existing workflow and call it transformation. They’ll be the ones who fundamentally rethink what value they’re providing and why a client would pay for it.

That means shifting from information delivery to genuine strategic counsel. From monitoring to anticipation. From drafting submissions to shaping the architecture of how decisions get made. From “we’ll get you a meeting” to “we’ll help you understand what actually moves the needle and build the case that makes it move.”

It means being honest with clients about what AI can do — including the parts that make the traditional retainer model uncomfortable. Because the firm that tells its client “you don’t need us for this anymore, but here’s where you absolutely do” is the firm that earns trust. The one that pretends nothing has changed is the one that gets quietly replaced by an internal hire with a Claude subscription and good political instincts.

And it means investing, seriously and immediately, in the capabilities of the people doing the work. Not a lunch-and-learn. Not a webinar. Genuine, ongoing development in AI fluency, data analysis, and the critical thinking required to know when the machine is right, when it’s confidently wrong, and when the answer requires something no model can provide.

The clock is already ticking

I run an advisory firm that operates across politics, strategy, and AI. I see both sides of this equation every day. The technology is not slowing down to wait for an industry that’s still debating whether it needs to pay attention.

The government relations profession has a window — a narrow one — to redefine its value proposition before the market does it for them. The firms that move now, that invest in genuine capability rather than cosmetic adoption, that are willing to cannibalise their own legacy models in service of something better — they’ll be fine. More than fine. They’ll be indispensable.

The rest will join the leather folio in the museum.