Mi3 report publication date. The company changes described occurred on different or undisclosed dates.

· Link

Your brand is being assembled outside the marketing department

AI assistants do not see departments. They assemble a view of a brand from product data, policies, documentation, reviews, media coverage, community discussions and owned content. Much of that evidence sits beyond marketing’s direct control.

That makes AI discovery an organisational test, not simply a search task.

Mi3’s The Prompt Economy provides a useful example. Sitecore found that AI tools were citing old web pages it intended to remove. It improved the useful material instead, then brought together people from web, content, documentation, social, community, public relations and customer advocacy. Sitecore says its measured share across monitored AI answers rose from 6.7 to 12 per cent, while appearances across priority prompts increased from 23 per cent to almost half.

The numbers are company-reported, but the operating lesson is useful: a marketing team cannot improve the brand’s representation if the underlying evidence is fragmented, outdated or owned by teams that rarely work together.

Key takeaways

  • Discovery reflects the whole business. Product information, service quality, policies and external commentary can shape an AI answer as much as campaign content.
  • Presence and preference are different. A brand may appear in an answer without being recommended. Track whether it is cited, described accurately, preferred and acted upon.
  • Search specialists become connectors. Their value is increasingly in finding evidence gaps, testing changes and directing work to the people able to fix the source.
  • Independent evidence carries weight. Reviews, communities, journalists and partners can support or undermine the claims a brand makes about itself.
  • A single visibility score is too blunt. Results need to be examined by customer need, platform, competitor and commercial outcome.

What could this mean for your marketing team?

  • Give one senior marketer responsibility for coordinating AI discovery, with authority to involve product, customer, communications, data, technology and legal teams.
  • Begin with two or three customer questions that matter commercially. Map the answer, its sources, the gaps and the owner of each corrective action.
  • Review findings regularly and escalate stalled work. Coordination without time, authority or accountability becomes another meeting.
  • Keep ownership where the truth lives. Marketing can organise the response; it should not rewrite product facts, service failures or legal decisions.
  • Report recommendation and business impact separately from raw visibility.

AI discovery is making the seams in marketing visible. The opportunity is not another dashboard. It is a better way for the organisation to produce, maintain and prove what it wants to be known for.

Mi3 produced the report with Omnicom Oceania. Its cases are useful operating accounts, not evidence that one structure will suit every organisation, and reported performance remains company-supplied.

Read the full study: download The Prompt Economy: AI Search and Discovery.

Mi3 · The Prompt Economy

Event: 17 Nov 2026 Evidence source