If you only check one highlight this week:

• B2B CMO? Influencer marketing tips

• Everyone else: what ChatGPT models to pick when

-François

Influencer marketing in B2B = maybe more powerful than AI?

Jessica Gilmartin’s preso on Influencer Marketing opened many eyes among the tech CMOs that she and I brought together.

Including mine.

Here’s what she shared (with some of my commentary weaved in):

Context

Jess advises a B2B company that drives ~ 60% of their revenue through influencer-driven campaigns on TikTok and Instagram (yes, B2B, not B2C, that is not a typo). Now, their product is demo-friendly.

Key finding

Influencer Marketing is still way underutilized in B2B Marketing.

It can yield high impact and high ROI if you ID the right creators, set the right terms, and accept to let go of control.

Here are some of their top strategies:

For Instagram and TikTok: Go for broad awareness, not narrow “Ideal Customer Profile” (ICP) targeting

If you ask influencers to create content that’s narrowly focused on your ICP, it may not be relevant to their large audience. The name of the game is virality and shareability so let influencers make content that has wide appeal to their audiences, in their style (the reason why their followers like them) so it hits more feeds. Trust the algos to then find your ICP.

Finding the right B2B influencers is still manual work, but worth it

It takes dozens of hours to find good influencers - there are no quick shortcuts here.

1. Create a fresh account, search relevant keywords, and train the algorithm by liking relevant creators so you keep finding similar ones.

2. Review the content of all potential partners to assess authenticity, quality content, and strong engagement.

3. Skip the “creator platforms” — they rarely deliver for B2B.

4. Note: I have used AI deep research a few times to ID influencers based on topics they discuss, and came up with lists that were only so-so. At least that helped a little. Pls share your prompts and tips if you had success with that.

Let influencers do their thing. Give some direction

Encourage influencers to:

1. Start with a relatable scenario (the first 10–20 seconds), ideally something fun and attention-grabbing. Usually around the pain point or where other solutions fail.

2. Then introduce your product, ideally with some screenshots or demos (visuals are always better on these mobile first platforms).

3. Keep it authentic and entertaining; those perform better than polished corporate videos.

By the time viewers realize it’s an ad for your product, they should be hooked on the story. Production quality matters less than authenticity.

Contracting best practices

Think of influencers as strategic partners:

1. Sign long-term contracts to avoid post-viral pricing spikes (i.e. an influencer’s subscriber base takes off, and then they charge you more because of their higher follower count!), and to block competitors from using them.

2. Try to get usage rights in your contracts so you can re-use strong content across your website and in paid ads.

Some strategies for LinkedIn

Use creators who genuinely use your product and build long-term relationships with a few targeted influencers.

Many CMOs in the room confirmed that sponsoring third-party posts on LinkedIn often performed better than their own ads by a long shot (more authentic).

Directional attribution rather than perfect tracking

Reliable attribution = very difficult for influencer marketing. Find and be OK with directional signals: self-declared in sign-up flows, annual brand studies. Look for correlation between video views and user spikes.

Conclusion: Try and do more and let them have fun! We're in a creator economy. The best ones have large audiences that deeply trust them. Even in B2B and for sales-led GTM motions, there is great potential.

My selection of tips, news and workflows

🔍 How to rank on LLMs & do AI Optimization

Detailed article on how to do GEO (GenAI Engine Optimization, aka AI SEO).

This should matter more and more as AI usage increases and Google AI Mode is now available to all U.S. searchers without having to opt into it within Search Labs.

🛠️ Which model to pick for what in ChatGPT?

OpenAI confuses us with some many models to pick from. And the differences are not obvious if you don’t keep up. So they clarified things in a recent article:

The bottom line:

• 4o is great for everyday low stakes tasks like summarizing, drafting emails, proof-reading reports, image generation

• 4.5 is supposed to be better at writing, but…

• o3 is the best. It's a reasoning model, it structures things better and IMO it even writes better. Use it if you can afford it, both cost and time wise (it’s slower). Use it for for “complex, or multi step tasks”, such as deep research, analyzing data, as “strategy buddy” and for writing prompts

• you can ignore other OpenAI models for the most part, unless you are coding or doing STEM (then consider 4.1 and/or o4-mini)

Note: This is just for ChatGPT, knowing that Gemini 2.5 keeps getting better, and Anthropic just released Claude 4 which seems to reclaim the coding throne (from Gemini 2.5 Pro) and be great for copy editing (haven’t fully tested it yet).

🛠️ Another CMO’s checklist to build an AI-forward team

Carilu shares her 13-point guide. Consistent with what I shared before, but I especially like four of her categories:

#3. Pick the Right Use Cases (aka: Don’t Boil the Ocean)

• Focus on highest-impact use cases versus trying everything

• Marketing seeing wins in SEO, SDR acceleration, data enrichment

• Demand gen and ops teams driving biggest automation wins

My take: also make sure they have access to foundational docs from Product Marketing.

#5. Create an AI Governance Board

1. Include IT, legal, and business teams.

2. Set protective and informative policies.

3. Consider on-premises solutions for security.

My take: Yes, too many employees are unsure about what they can use and how. Providing clear policies and guardrails will reassure them.

#9. Run Cross-Functional Hackathons

1. Two-day quarterly hackathons build cross-functional AI relationships

2. Marketing, engineering, support, accounting teams collaborate on innovations

3. Relationships formed help troubleshoot and expand AI processes

My take: It’s critical to make time and space away from daily work. We did something similar at Twilio, mixing Devs with Comms, with Events , with Product Marketing folks. Each brought different skills and viewpoints, and learned a ton from each other. It fostered significant innovation and new bonds too.

#10. Host an AI Leaders Retreat

1. Multi-day retreat for most AI-hungry and accomplished employees

2. Creates company value signal: AI innovators are supported/rewarded

3. Sparks relationships, ideas, and implementations across the organization

My take: yes! The version of “Club” for AI champions that will benefit the entire company.

💡This advice from Box’ CEO applies to us all

Final Words

Thanks for sharing these highlights with busy marketing execs around you.🙏 

Someone forwarded you this email? You can subscribe here.

François | LinkedIn 

I'm a CMO, advisor, and "CMO Wingman". Yes, that's a thing :-). Ask my clients: in this AI era, CMOs need a strategic proactive advisor more than ever. I’m former CMO at Twilio, Augment Code, Apollo GraphQL, Decibel, Udacity and Head of Marketing for LinkedIn Talent Solutions.