The disappearing direct link: Why AI is rewriting the playbook for investor relations 

For decades, the ‘direct link’ has been the holy grail of Investor Relations. An IR Director speaks; the market listens. Whether through an earnings call, a physical roadshow, or a meticulously crafted RNS, the connection between the equity issuer and the financial analyst was personal, nuanced, and immediate. 

But in some instances that direct link is disappearing. And being replaced by an AI Proxy. 

In a recent conversation with Ross Hawley, Director of Investor Relations at ZIGUP PLC and a member of the IR Society’s AI working group, a new reality emerged: we have entered the era of ‘machine-first consumption’. 

Today, your most influential ‘investor’ might not simply be a human at a fund - it’s also the AI model they’ve programmed to digest your data before they even see it. 

The rise of the ‘AI Proxy’ 

As professional investors and newsrooms increasingly rely on ‘Agentic AI’ to summarise performance, the corporate narrative risks being stripped of its nuance. When an AI ‘scrapes’ a PDF, it doesn't hear the CEO’s conviction or understand the specific context of a non-GAAP reconciliation. It looks for patterns, probabilities, and keywords. 

Ross Hawley describes this as a fundamental shift in how ’sell-side’ and ‘buy-side’ consume information, with a risk that the narrative is diluted or lost. If your company’s story isn't ‘machine-readable’, the AI proxy fills in the gaps itself - often with hallucinations or outdated peer comparisons. 

AI as a mirror, not a ghostwriter 

The most sophisticated IR teams aren't using AI to write their messaging; they are using it to audit it. At ZIGUP, Hawley has utilised Burson’s Decipher tool to track the sentiment progression of results announcements. 

He is clear however that this isn't about letting a machine take the wheel. It’s about using AI as a ‘sentiment mirror’—insight as to how the market’s digital sentinels will ‘tag’ and ‘label’ your performance before you go live. If the AI sees a ‘hedging’ tone where you intended ‘cautionary optimism’, you need to be aware. ZIGUP uses it more subtly – not as a standalone view of the upcoming statement, but to look over time at the progression of messaging and tone – over the past 3 results statements. 

From ‘passive scraping’ to ‘active authority’ 

The challenge for listed companies is how to reclaim the direct link. At Burson Buchanan, we believe the answer lies in moving from being a company that is summarised by AI to one that provides structured data for AI. 

This requires two technical leaps: 

Semantic infrastructure: Moving beyond PDFs. Every disclosure must now include invisible data tagging - machine-readable snippets that tell an AI exactly what your Revenue, EBITDA, and Dividend figures are, leaving zero room for ‘hallucinated’ interpretations. 

The ‘Direct Pipe’: Utilising the Model Context Protocol (MCP) to host an ‘Authorized Data Server’. This allows professional AI agents to ‘subscribe’ directly to your verified, official data stream, bypassing the noise of the open web. 

The Bottom Line 

The future of corporate communications might not be about ‘beating’ the bots; but about owning the data layer they rely on. And whoever structures meaning best, owns narrative visibility. 

I finished our conversation by asking Ross whether he thought our role is primarily becoming one of ‘data guardianship’. In perhaps a note of hope, he rejected this conclusion, stating that he felt the ‘narrative was ever more important’. 

What might be increasingly true, is that in a world where machines are often the gatekeepers, your most important strategic asset isn't just what you say, it’s how well the machines can hear you.