How our strategy helped a global energy supermajor see who really speaks for it in AI answers
A global energy supermajor's YouTube channel was absent from AI answers about its own brand. Juan Felipe Arango led the audit and the plan to fix it.
Situation
A global energy supermajor’s main YouTube channel held about 1,500 videos carrying three unrelated businesses: retail and service-station ads, motorsport sponsorship, and a thin strand of corporate film. No one owned its editorial line. The client saw it as a regional workflow problem: too many teams uploading.
Task
An agency partner brought me in to review the channel against four peer energy majors, with one page on metadata and AI search, one on paid opportunities, and a conclusion answering four questions. How do you turn this into a strategic channel? What format and rhythm would make it useful? Where do Shorts and paid fit next to long-form? And how do you keep regional video needs under control? The budget was two days. On the kickoff call, the agency asked for disagreement and thinking specific to this brand, not polish.
Action
I reframed the brief from metadata hygiene to a sharper question: who narrates the brand where AI answers get written? The answer reset the project. A wire service held about 3.5x more YouTube citations about the brand than all five majors’ own channels combined, and the most-cited YouTube video about the brand was theirs. The channel wasn’t underperforming against its peers. It was absent from the conversation about its own brand.
Six research blocks backed that up: an AI-citation audit of 106 prompts across 7 AI answer surfaces, a 90-day engagement pull comparing long-form and Shorts, a metadata scrape of four majors’ video libraries, a capture of 21 paid creative concepts from public ad libraries, a branded-search audit, and a check of every third-party claim against first-party sources. Five of 12 didn’t hold up, so they were cut.
The evidence pointed the same way. In 90 days, a sponsorship clip took over 100x the views of a corporate film posted weeks apart. Both were well made. The corporate one was publishing into a feed built for something else. Across three majors, 61 financial-results and investor videos had zero chapters: open ground. The brand’s median description was about a fifth the length of its closest peer’s, and it linked back to its site from 39% of videos against the peer’s 97%. And only 9 of 124 YouTube citations were Shorts, so long-form earns the citations and Shorts earn reach.
The plan made three asks. Decide what the channel is for and give it a named owner. Retrofit the corporate videos so AI answers can quote them, starting with chapters, with no new production. Connect corporate paid to the corporate channel and measure both on the same audience. For the regional question, the answer was one hub channel with audio per region through multi-language audio and auto-dubbing, not regional channels or a tighter approval queue.
Four KPIs set the baseline, led by citation share: how often the brand’s own videos are the source when an AI answer talks about the brand. It started at 6%. Retrieval readiness (how many corporate videos carry chapters, a full description, and a link back), qualified attention, and downstream referrals completed the set. A 10-point upload checklist and next steps in 30, 60, and 90-day windows followed, with the first 30 days limited to quick wins the central team controls. I delivered it all as an 18-page report, a client-facing executive summary, an internal summary, a video walkthrough, and a call script.
- 6%Citation shareHow often the brand's own videos are the source when an AI answer talks about the brand
- 4.6%Retrieval readinessOf ~109 corporate videos: 4.6% have chapters, 6.4% a full description, 61.5% a link back
- Not yet measuredQualified attentionAverage share watched on corporate long-form, not raw views
- Not yet measuredDownstream referralsVisits the videos send back to the brand's site
Result
I presented the work directly to the client’s corporate communications and social teams, alongside the agency. The client’s corporate comms lead called it “really clear” and “really useful,” said it gave the team “loads of food for thought,” and named the senior colleagues they would take the insight to. The social lead agreed to start with the quick wins, since they depend only on the central team. The client’s own early plan already matched the recommendation: one shoot cut into long-form and Shorts, with pieces left over for paid. This was the second project the agency brought me in for, after an account-based pilot for a global life-sciences supplier.
