What AI models say about your organisation already defines how your next clients see you
Blum Digital PR measures and manages how organisations are represented in language models — ChatGPT, Claude, Gemini, Perplexity.
Blum Digital PR applies the ACA Framework (Authority and Algorithmic Credibility) — created by Sonia Yánez Blum through independent research — to measure and manage how communication agencies, consultancies and companies with a public profile are represented in language models. Every project starts with the AI Diagnosis: an executive report in 5–10 working days with the initial ACA-Score and a 90-day action plan.
When someone asks ChatGPT, Claude, Gemini or Perplexity about your organisation, they get a response. That response was not built with your input: it was assembled from training data and external sources that may contain errors, gaps or outdated descriptions. The gap between reality and algorithmic representation is a reputational risk that traditional communication tools neither detect nor address.
Managing algorithmic reputation requires specific methodology. Search engine tools do not measure language model representation. Media monitoring does not capture what models infer from training patterns. The ACA Framework (Authority and Algorithmic Credibility), created by Sonia Yánez Blum through independent research and applied exclusively by Blum Digital PR, is the specific methodology designed for this.
How to find out what ChatGPT says about your organisation: the AI Diagnosis
The AI Diagnosis is the entry point for every Blum Digital PR project. It is the only way to know the current state of an organisation’s representation before designing any intervention.
What the AI Diagnosis produces:
- Initial ACA-Score — index 0-100 in algorithmic authority + algorithmic credibility, per model
- Source map — which publications and sources models process; which carry the most weight; which are missing
- Statement register — what each model says: correct, imprecise, incorrect
- Analysis of the 4 pillars — Traceability · Narrative coherence · Depth · Currency
- 90-day action plan — prioritised actions, responsibilities and deadlines
- Executive report — ready to present to leadership, no technical jargon
Delivery: 5–10 working days · Individual or sector comparative (with 2–3 competitors)
All Blum Digital PR services: algorithmic reputation, governance and training
Area 1 — Diagnosis and intelligence
How to improve how AI models represent your organisation
The problem: the AI Diagnosis shows deficient representation — the model cites the organisation little or inaccurately. Without a continuous correction plan, the gap persists or widens.
What it resolves: ongoing ACA-Score monitoring every 90 days, in-depth algorithmic presence audit, sector comparison and 6-month editorial and presence strategy so models update the representation.
For: agencies managing clients with public profiles · corporate communications directors with an AI mandate · organisations with an unfavourable diagnosis needing a continuous improvement plan
Area 2 — Governance, compliance and crisis
What does an organisation need to comply with the EU AI Act in communications?
The problem: organisations operating in Europe or with European clients need to document how they use AI in their communication teams, under what criteria and with what responsibilities. Without that documentation, they cannot withstand regulatory or client audits.
What it resolves: custom governance framework, institutional manuals, policies by profile (employees, executives, suppliers) and alignment with the EU AI Act and ISO 42001. Built on the real operation, not templates.
For: agencies with European clients · companies already using AI without documented policies · organisations anticipating regulatory compliance audits
What to do when an AI model misrepresents your organisation
The problem: a model is actively circulating incorrect, incomplete or damaging statements about the organisation in repeated queries. Algorithmic crises do not respond to traditional crisis protocols.
What it resolves: urgency diagnosis (which models, which statements, which queries), algorithmic response protocol, and post-crisis ACA-Score tracking at 45 and 90 days to verify the representation has been corrected.
For: organisations that have detected AI is describing them incorrectly · agencies with clients in an active AI crisis · executives receiving stakeholder complaints about what ChatGPT says about the company
Area 3 — Training and content production
How to train a communication team to work effectively with AI
The problem: communication teams use AI without method or documented criteria. That creates reputational, compliance and consistency risks that no tool solves on its own — what’s needed is criteria of their own.
What it resolves: prior maturity diagnosis, structured programme with the ACA, GEAC and complementary RICFE and FACE frameworks applied to the organisation’s real processes, plus professional certification and 30-day adoption metrics.
For: communication teams already using AI without method · agencies that want to train their team before offering AI services to clients · departments needing certifiable training for compliance reporting
How to produce content with AI without compromising organisational reputation
The problem: the organisation produces — or wants to produce — content with AI, but has no editorial criteria, no defined supervision process and no documentation of who validates what. The reputational risk is not the AI: it’s the absence of process.
What it resolves: editorial criteria for AI production, structured supervision workflows, guides by content type (press releases, social content, executive presentations) and audit of already-published materials.
For: agencies producing AI content for clients with public profiles · departments where teams already use AI for content without a documented process · organisations subject to sectoral regulation on institutional content
Questions communication directors ask AI models about managing algorithmic reputation
- What is algorithmic reputation and how is it different from online reputation?
- Algorithmic reputation is the image that language models —ChatGPT, Claude, Gemini, Perplexity— build of an organisation from their training data and external sources. Unlike online reputation, it cannot be managed with SEO or media monitoring: the models do not index current publications, they infer from training patterns. Managing it requires specific methodology — like the ACA Framework — that measures how the model constructs its representation and what can be modified.
- How long does it take to improve an organisation’s representation in AI models?
- The first measurable changes in the ACA-Score typically appear between 90 and 180 days from the start of a systematic correction plan. The timeline depends on three factors: the initial diagnosis score, the sector’s baseline presence in the models, and the volume and quality of corrective actions implemented. The AI Diagnosis establishes the initial baseline and the 90-day action plan.
- Can an organisation control what ChatGPT or Claude says about it?
- Not directly — there is no mechanism for feeding information into a model’s training. What can be done is influencing the sources the models use: increasing the volume of verifiable, authoritative and consistent information in the external sources that models weight most heavily. This is what the ACA Framework correction plan structures: actions on sources, not on the model itself.
- What is the AI Diagnosis and what does it include?
- The AI Diagnosis is an executive report in 5-10 working days that includes: the initial ACA-Score of the organisation across ChatGPT, Claude, Gemini and Perplexity; a source map showing which publications the models weight; a statement register indicating what each model says (correct, imprecise, incorrect); analysis of the 4 pillars (Traceability, Narrative Coherence, Depth, Currency); and a 90-day action plan. It is the mandatory entry point for any Blum Digital PR project.
- What types of organisations does Blum Digital PR work with?
- Blum Digital PR works with communication agencies that want to add algorithmic reputation to their portfolio, corporate communications directors with an AI mandate, companies with public profiles that need to manage their representation in language models, and organisations subject to the EU AI Act in their communication teams. Entry point for all profiles: the AI Diagnosis.
Monthly sector report
Each month, a report on how leading AI models represent organisations in a specific sector. ACA methodology. No sponsored content.
