If you only check one highlight this week:

• CMO? How to get better attribution insights.

• Want to create better content with AI? See sections 2 & 3.

-François

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

Someone forwarded you this email? You can subscribe here.

Good AI content. At scale

We all know AI can create content for us. A lot of it. Fast.

But we’ve also mostly seen a lot of mediocre content coming from AI.

It's starting to change.

Recently, many have started to figure out how to create good content, at scale, using AI. You will see some techniques to do that in the sections below.

That didn't come overnight. They worked on their workflows, refined their prompts, iterated, and they, critically, decided to always keep a human in the loop as reviewer and final editor. Check out this article - it's tough to tell it was 90% AI-generated.

That's yielding promising results for their brands and products’ discoverability in search engine and AI search results.

This AI content generation game is unfortunately not a game you can ignore if you're targeting an audience that discovers, learns, and makes decisions based on LLM search results.

Many companies are already using these processes to boost rankings, so the new race has started in earnest.

We used to talk about paying the Google tax. Now it seems we're paying the ChatGPT/AI Overviews/Perplexity tax.

Yes, we are working harder - and paying more - to feed the very models that are disrupting our playbooks! The more content we give them, the more all-knowing they become, the more our content gets bypassed and consumed in the form of summaries or citations via chatbots.

But will that “90% AI, 10% human” content be the one that will help shape and create entire categories? New movements? Establish new brands?

I really doubt it.

Note that I wrote good content above. I did not write great content, and surely did not write exceptional content.

We still seek emotions, human stories, personalities, new insights, connecting new dots. Those are the components needed to truly, deeply shape our perceptions and thinking and spur us to action.

Those are more easily delivered via experiences rather than just text.

That's why marketers are creating more videos and further investing in community and events (i.e. the human side) but also in app creation (for direct interaction, personalization and ease of consumption) - in addition to all that AI content.

And it's only the beginning…

My selection of tips, news and workflows

🎯 Want to create content with AI? Start by creating a Content ICP and add it to your AI’s memory

The best content creation workflows I've seen with AI (whether that's from Marcel at GrowthX or in this case Kieran at HubSpot) start with building a detailed profile of your target audience - including its challenges and preferences - in the context of your product and associated perceptions.

Kieran describes (YouTube - 17 min) how he brings together:

• First-party data (emails, sales calls recordings, customer interviews, survey responses, win/loss notes)

• Third-party data collected via deep research (often using Perplexity)

… to create a detailed Content Ideal Customer Profile (ICP) for each one of his priority personas.

One of the tricks is to extract your audience’s specific words, especially those loaded with emotion.

He then feeds that into Claude (still has the writing edge) and his own GPTs that craft content following his prompts and guidelines. For instance, a GPT (It can be a Gemini gem or a Claude project) that creates compelling hooks based on different devices (e.g. contrast, curiosity gap, metaphor, number, mini-story, paradox, taboo).

👉 It's now really easy now to create various flavors of content for different personas.

👉 Work with your product marketing team on deep and detailed profile cards. Most of the Persona cards that I've seen lately don't go deep enough.

It's worth watching the video if you don't already master that.

📝 An AI workflow to create technical content for AIO that even CEOs end up trusting

How do you significantly scale and accelerate content production for a technical audience and avoid creating AI slop that can damage your brand?

It's a journey. But it's one worth taking.

In his lightning talk (Slides, Recording), Ed Farraye describes the progress he made with his client (who wanted to remain anonymous) to create technical content faster.

The results after 6 weeks:

• 60+ pages published and ranking

• 6 of top 10 organic traffic pages are AI-generated content

• Significant visibility improvement vs. competitor

• A trusting CEO

Their phase 1 was mostly manual. They produced articles in 2.5 hours, but only rated them 3/10 in technical quality. They used Perplexity for deep research, Claude for drafting outlines, and ChatGPT to critique and improve them, followed by human reviews and editing.

For phase 2, they automated their workflow with AirOps:

• Use Perplexity for deep research:

  • Include audit of pages currently ranking for a keyword

  • Keyword density map from first-page results

• Generate 3–4 potential content ideas

• Use Claude for prose

• Use GPT for feedback and to add anchor links, prioritizing higher-quality content

• Run a separate workflow to improve hyperlinking and keyword density of existing content

Critically, they put their human experts in the loop:

• to pick the direction for each article based on 3-4 content ideas

• to lead a technical content review of the final draft

💻 Doing attribution manually. With all that AI, really? Yes!

Show me a marketer who has not experienced the painfully obvious limits of their attribution tool. 

So how can marketers best direct their investments and report to the BOD with confidence?

➡️ Enter "manual attribution" for open opps and post-mortems for deals won.

They complete - and correct - the picture in ways no attribution tool can (just think of all the interactions that take place offline and no system tracks). 

These two techniques help Marketing, Demand Gen, and Sales really understand what's truly influencing and winning deals. 

🤝 Manual Attribution

Gonto gave this lightning talk on the process that his teams (at Auth0 and Vercel) followed for 2-4 hours weekly to analyze which campaigns and channels helped create opportunities (workflow summarized below).

If someone as digitally and technically savvy as Gonto prioritizes such an exercise, it says a lot about the insights a team derives from the process.

The workflow:

- Every Monday, demand gen and growth marketing teams spend 2-4 hours analyzing every sales opportunity

- Goal: identify the “catalyst” - the specific moment that made prospects decide to seriously try the product

- Not necessarily first or last touch, but the thing that triggered real consideration

Analysis includes:

- Call transcripts from all meetings (now searchable with AI)

- All connected emails from Salesforce/HubSpot

- Website analytics and activity data

- Example: Gartner booth didn’t generate leads directly, but validated enterprise credibility for Fortune 500 prospects who had tried Auth0 previously

- Catalyst field added to Salesforce, set by IC doing research

- Cascading validation: manager reviews and validates each catalyst determination

- Weekly review at leadership level for pattern recognition and objectivity

🤝 Group post-mortems on deals-won

Another best practice = bringing together a deal team once you've won it. From Sales, to BDRs/SDRs, to Sales Engineers to Demand Gen, to Events, to Marketing Ops folks.

Take a look at what truly influenced the deal across personas and channels. Bring CRM and campaign data to that conversation as well as sales call recordings and emails.

You can/should even dump it all in an AI project and get insights from that.

Upstream of that, ensure that someone in sales asks prospects:

- during qualification or disco: "How did you hear about us?

- at contract signature, ask the champion: "What was valuable in convincing you and your colleagues?"  

👉 Doing this on a regular basis will reveal patterns of what works and doesn't and help you direct your investments and better report to both the CFO and the board. 

Final Words

François | LinkedIn 

I'm a CMO, advisor, and "CMO Wingman". Yes, that's a thing :-). Ask the CMOs I support: 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.