AI Search visibility assessed and improved across eight portfolio companies.

The challenge

The client was a private equity firm with a B2B portfolio of eight companies spanning enterprise software, professional services, and industrial technology. They had identified AI Search visibility as a strategic concern — not from an academic interest in emerging technology, but from a practical observation: their portfolio companies' sales teams were hearing prospects cite AI-generated answers as discovery sources in sales conversations.

The challenge was that nobody in the portfolio had any structured visibility into how their brands were performing in AI search. There was no measurement framework, no baseline, and no way to prioritise which companies needed attention first. The firm wanted a programme that could establish this visibility and begin improving it.

Our approach

A portfolio engagement is structurally different from a single-brand engagement. We needed to build something repeatable and comparable — a measurement framework that could be applied consistently across eight different companies in different sectors, then used to prioritise the active programme.

Phase 1 — Measurement framework design. We defined a consistent methodology for assessing AI Search citation share: query selection criteria, platform coverage (ChatGPT, Perplexity, Gemini, Google AI Overviews), citation classification rules, and a scoring model that allowed cross-company comparison despite different market positions and query volumes.

Phase 2 — Portfolio audit. We applied the framework to all eight companies over six weeks. Each company received a citation share baseline across 20–40 target queries, a competitive benchmarking report against their two or three named competitors in AI responses, and a prioritised action list.

The audit produced a clear priority tier. Three companies had both significant citation gaps relative to competitors and the content and technical foundations to improve quickly. Three were in early-stage positions where entity work needed to happen before content could be effective. Two had strong AI visibility already and needed monitoring rather than intervention.

---

The work

Active programme — three priority companies. For the three highest-priority companies, we ran a structured six-month AI Search programme: entity repair, structured data implementation, content restructuring for AI citation patterns, and co-citation seeding.

Each company had a different gap profile. One was an entity recognition problem — the brand had weak presence in the sources AI systems rely on. One was a content architecture problem — good brand recognition but content that didn't answer the question forms AI systems were being asked. One was a co-citation problem — the brand was known but not consistently associated with the authoritative sources AI systems trusted.

Monitoring infrastructure. We configured our citation monitoring platform for all eight portfolio companies — not just the three in the active programme. This gave the firm a live view of AI Search performance across the entire portfolio, with alerting when significant changes occurred.

---

The outcome

After six months, the three companies in the active programme showed material citation share improvements — from a combined average of 12% to 34% across target queries. The measurement framework has been adopted as a standing KPI across the portfolio, reported quarterly alongside traditional organic traffic metrics.

"Before this programme we had no way to answer a basic question: are our portfolio companies visible in AI search? Now we have a framework that answers that question consistently and can track it over time."

The engagement is ongoing. The three monitoring-only companies have been reprioritised as conditions change, and one has moved into an active programme following a competitive incursion.

---

Related disciplines

This engagement draws on AI Search strategy, measurement framework design, entity visibility, GEO, and AEO. It is unusual in being a portfolio-level programme rather than a single-brand engagement, which required a different approach to scoping and prioritisation.