Blog post

Why is tracking government visibility on LLMs different from tracking brands?

The advent of LLMs and AI search means that there has been a colossal shift in how audiences are consuming information today,  and a reciprocal shift in how all types of organisations, from government agencies to brands are responding . But while discussion amongst PR and comms pros tends to fall disproportionately on how brands are impacted, the needs of the former are just as keenly felt, and often quite distinctive.

So how can government agencies respond, especially in a time of global flux, when major policy changes need to be communicated and important stakeholders need to be managed? After all, government organisations do care about reputation, much as brands do, but have quite distinctive goals when it comes to ensuring accurate information reaches the right audiences.

There is a new entry point for stakeholders

The era of “Let me Google that” is rapidly fading. Instead of clicking through to official websites, people are asking chatbots for direct answers.  What’s striking is that government agencies often have no visibility on how they’re being talked about in the LLM space, even as it becomes a central channel for messaging and reputation. 

When AI models become the primary gatekeeper, audiences bypass official portals entirely — driving down site traffic and leaving agencies vulnerable to misinformation, negative sentiment, or worse, being left out of the conversation altogether. 

Therefore, the entry point is different. For commercial brands, this shift is profound but in some ways mediated – an FMCG brand, for instance, often discovered through third-party platforms in any case. But for government entities, the stakes are entirely different. As the sole, authoritative source for public information, they need citizens on their websites to get accurate details. Government agencies, especially if they are the authority or regulator in a particular industry or sector, need to make sure audiences know where to go to get the right information.

Different types of government agencies have different considerations

Not all government agencies are alike, and they all have different parameters that are quite non-negotiable for them, just by the way they function. 

1. Service delivery agencies – these rely heavily on content freshness. They can’t risk outdated sources impacting eligibility or process changes not reaching audiences.
   
2. Regulators – these need to transmit authority and trust. Regulators have to make it a priority that they’re amongst the first place audiences go to for information and that the industry they’re regulating does not put them in the shade on the channels stakeholders are actually using
   
3. Policy departments – did the AI’s account of a policy match what was actually announced? They want to be able to make sure the accuracy of a policy (and ideally, its effectiveness) is translated when audiences search through LLMs.
   
4. Local and state authorities want to make sure the services they carry out on behalf of locals are visible too. Much like service providers, there is a question around access and awareness of programmes and regulations, but also an added consideration: not appearing on LLMs discontent amongst those wishing to see a return on taxation and electoral mandates, and give credence to bad actors.

What do government agencies want to get out of this new LLM-mediated landscape? 

Reputation is important, but that exists downstream from maintaining a flow of accurate information. It’s useful for communications teams in government organisations to self reflect and ask themselves the following questions:

• Where are citizens going to find information about your services if not your website — and do you know what they’re being told?
  
• If there was a significant policy change or incident in the last twelve months, do you know how it’s currently being characterised when someone asks an Al tool about your agency?
  
• When you communicate a major service change or policy update, do you have any way of measuring whether it’s surfacing in searches about you?
  
• How do you currently understand the gap between what your agency publishes and what citizens actually receive when they search for information?
  
• Are there community groups, advocacy organisations, or media outlets shaping perception of your agency – and do you know if that’s feeding into what Al models say?

These are gaps they already realise, but they don’t actually know what to do about it – how to manage or measure them. They need a tool that allows them to know this critical piece of information and make informed decisions. 

Ngaire Crawford, Pulsar Group’s Executive Director for AI Strategy says, “It’s easy to disregard LLM reputation as part of a difficult AI landscape or something that is only really relevant to more product-based communication or marketing. Government communications is about social licence, and ensuring the public have access to up to date and accurate information about things that matter to them, the role that Generative AI plays in how a community understand an issue will only continue to grow, and knowing the impact that you can have in that through small shifts in channel strategies or more consistent messaging is a crucial part of the communications toolkit.”  

Lumina AI View: AI visibility for PR & Comms

Lumina AI view is built for communicators who want to understand how their organisation and their competitors are being talked about by various AI models – including ChatGPT, Gemini, and Claude. AI View users get an insight into which sources are being cited, and how they would need to respond as a way of protecting their reputation or making sure correct information about them is being disseminated. 

The tool provides an AI view score — a composite metric ranging from zero to 100, designed to help track brand performance over time and facilitate comparisons against competitors. It is calculated using five weighted factors — sentiment, visibility, authority, dominance and freshness. Beyond the overall score, the platform provides a summary of a brand’s AI narrative based on four distinct reputation pillars — direction, performance, integrity and innovation. 

These pillars help users identify exactly which dimension of a brand’s reputation is under pressure, offering specific, actionable insights for board presentations or reviews. Ultimately, while the AI view platform provides the necessary intelligence, the strategic decisions regarding how to respond to these insights remain with the organisation.

Spotlight: An Australian Council

This progressive local government council is located in Australia’s leading center for culture and sports.

The council earned an AI view score of 74 reflected by strong reach and authority. Publications like CBD News and its own website are the most cited by LLMs — interestingly, most of them being cited by Claude.

Content freshness scored lower at 48. Their website still carries error pages and annual reports from a few years ago. If a report — one that is seen as an organisation’s most comprehensive and authoritative content, is still being cited even if it’s older, might potentially be in the way of the organisation’s own perception. Which means more work is needed to prevent outdated content from still appearing.

What type of content is showing up?

Most citations for the council originate from government sources, followed by news outlets, reports, and blogs. The domain is evenly split between owned content and content earned from external sources media articles and independent authorities. 

While most are recent, some older articles from major outlets such as The BBC remain visible and may significantly influence how LLMs perceive the council. Owned content typically addresses last year’s budget plans and the council’s latest vision for the city, which LLMs are referencing. Government sources are the major content type, however, external sources have a greater impact on the council’s overall LLM score.

​

The stakes are higher for government agencies

When brands track LLM visibility, they often ask, “Are we shown in a positive light?” or “Are we cited accurately?” For the government, additional questions arise: “Is this information accurate enough for someone to act on?” and “Are we still viewed as more authoritative than what we oversee?” Mistakes can have serious consequences, such as individuals applying for ineligible programs or missing critical deadlines for new initiatives or elections. This can quickly lead to public frustration. It is essential for government communicators to recognize these risks.

If you would like to know more about our Lumina suite, please reach out here and our team will get in touch with for you a quick demo.