IBM Granite: reaching developers and the C-suite through B2B influencer marketing

Influencer Marketing

Seven developer and AI specialist creators introduced IBM's Granite LLM through two-part Preview and Reveal videos across four platforms, reaching 43M impressions.

9Mengagements
43Mimpressions
68Klink clicks
59Kpaid link clicks
9Korganic link clicks

About the campaign

IBM needed its purpose-built LLM Granite to stand out to developers and C-suite decision-makers in a crowded AI market clouded by confusion about practical applications and scepticism toward vendor claims. The Goat Agency's insight was that a technically sophisticated audience trusts reputable developer and AI voices over brand messaging, and consumes tech content on consumer platforms as well as LinkedIn. Because Granite targets enterprise workloads, both the developers who would use it and the executives who would approve it had to be convinced, so IBM had to earn credibility rather than assert it.

The idea had seven developer and AI specialist creators, with a combined following of over 5 million, each build a real project with Granite and tell it in two parts, a Preview that sets up the project and a Reveal that shows the finished result. Selection ran through the agency's IBEX platform. Short-form video ran on LinkedIn, TikTok, Instagram and YouTube Shorts, briefed on key messages but never scripted, amplified with paid media, with links to IBM.com and to platforms such as Hugging Face where developers already evaluate models.

Matching creator credibility to the audience carried an enterprise message generalist influencers could not, and the Preview then Reveal structure created a narrative gap, with Reveal posts out-performing Previews on engagement. The Goat Agency reports 43M impressions, 9M engagements and 68K link clicks, and was entered in the B2B category at the 17th Annual Shorty Awards. All figures come from the agency's own page and are not independently verified. No launch date, creator names, brand-perception measurement or commercial outcome is published, and the Shorty entry states no result.

Case in 60 seconds

IBM needed its purpose-built LLM Granite to stand out to developers and C-suite decision-makers in a crowded AI market where confusion about practical, real-world AI applications was widespread.

A technically sophisticated audience trusts reputable developer and AI voices more than brand messaging, and consumes technology content on consumer platforms as well as LinkedIn.

Have seven developer and tech/AI specialist creators each build a real project with Granite and tell that story in two parts, a Preview that sets up the project and a Reveal that shows the finished result.

Short-form video on LinkedIn, TikTok, Instagram, and YouTube Shorts, briefed on key messages but never scripted, amplified with paid media, with links to IBM.com and to third-party platforms such as Hugging Face.

The agency reports 43M impressions, 9M engagements, and 68K link clicks, with the Reveal posts driving more engagement than the Previews. The campaign was entered at the Shorty Awards.

Strategy Breakdown

The primary strategy-learning section — interpretation, not a press-release summary.

Two distinct professional audiences: the software developers who would use Granite, and the C-suite decision-makers who would approve it. Granite is built for enterprise workloads, so both groups had to be convinced.
AI is everywhere in professional conversation, but the sources note widespread confusion about what it can practically do in the real world, which breeds scepticism toward vendor claims.
The AI category is increasingly crowded and every vendor claims leadership, so IBM needed to earn credibility for Granite rather than assert it.
IBM stepped back and let credible practitioners demonstrate the product. The brand supplied key messages while creators kept their own voice, positioning Granite through evidence, working projects, rather than advertising claims.
Curiosity carries attention across posts: previewing a project and returning with the finished result gave audiences a reason to come back, and the Reveal posts out-performed the Previews on engagement.

Creative & Execution System

Assets & channels

  • Short-form creator video on LinkedIn
  • Short-form creator video on TikTok
  • Short-form creator video on Instagram
  • YouTube Shorts
  • Paid media amplification

Partners

  • The Goat Agency (strategy, creator selection via its IBEX platform, campaign delivery)
  • IBM (client)
  • Seven developer and tech/AI specialist creators

Participation mechanic

  • Creators briefed on key messages, never scripted
  • Two batches of creative per creator: a Preview post and a Reveal post around a project built with Granite
  • Links drove to IBM.com and to third-party platforms such as Hugging Face

Creator role

  • Developers and tech/AI specialists building and demonstrating real projects with Granite in their own voice

Results & Evidence

MetricValueResult typeConfidence
Impressions43MreachAgency reported
Engagements9MengagementAgency reported
Paid link clicks59KtrafficAgency reported
Organic link clicks9KtrafficAgency reported
Total link clicks68KtrafficAgency reported

Awards & Recognition

Award showYearCategoryResultSource
Shorty AwardsNot stated in sourceBusiness to BusinessEntrant

Why The Goat Agency's strategy worked

  1. 1

    Match creator credibility to the audience: developers and tech/AI specialists carried an enterprise AI message that generalist influencers could not.

  2. 2

    A two-part Preview then Reveal structure created a narrative gap, and the Reveal posts out-performing the Previews suggests the mechanic worked as designed.

  3. 3

    Briefing key messages without scripting kept each creator's voice intact, which is central to why their audiences trusted the content.

  4. 4

    B2B content can run on consumer platforms: TikTok, Instagram, and YouTube Shorts carried the campaign alongside LinkedIn.

  5. 5

    Sending developers to third-party platforms such as Hugging Face, not only the brand site, treated product exploration as the goal rather than site traffic.