Full Steam 01: Taste, judgment, and the new superpower
Full Steam is an ongoing conversation about the craft of B2B marketing. In each episode, experts from Iron Horse explore a single topic in depth, giving you practical takeaways and a fresh perspective on the challenges you're navigating every day.
The funnel has flattened into a fishbowl, most of the buyer's journey now happens off your domain, and most marketing leaders are trying to figure out exactly what their team actually needs to be good at now.
In the latest episode of Full Steam, Iron Horse senior content strategist Alex Jonathan Brown sits down with Iron Horse CEO Uzair Dada to dig into how AI is changing B2B marketing teams: what's stopped working, what's emerging in its place, and why "taste and judgment" have quietly become the most valuable skills on any modern marketing roster.
In this conversation, Uzair Dada explains why he thinks the marketing industry "effed it up" with over-complication, how the BC/AC ("Before Claude" / "After Claude") shift has reset the floor on what's expected from every marketer, and where the real competitive advantage lives now.
Featuring
- AI has raised the floor for marketing teams. Every contributor is now expected to arrive with informed thinking, a draft, and a point of view.
- First-party data is the moat. Sales calls, CS transcripts, and customer review data are the inputs that make your AI workflows differentiated. Mine them, or stay generic.
- Off-domain discovery is the place to win. LLMs, communities, analyst content, and third-party authority now drive a disproportionate share of high-converting pipeline and the deals they source close at meaningfully higher rates.
How is AI changing B2B marketing teams?
Uzair frames the past two and a half years as a watershed. He notes the biggest shift in December 2025, when Claude and the first wave of true coding and co-working agents arrived. That, he argues, is when the "mechanic of possible" actually changed for marketing teams.
Before that shift, marketing's complexity problem was a tooling problem: 15,000-plus martech logos, overlapping point solutions, and stitched-together stacks where most teams used 20–40% of what they'd paid for. After it, the bottleneck moved. Repeatable work is increasingly automated. Slow work is faster. The question shifted from "do you have the right tool?" to "do you have the right judgment about what to do with it?"
That reframing changes every conversation about hiring, org structure, and what "good" looks like.
What skills do B2B marketers need in the AI era?
Two things, in Uzair's framing: first-party data fluency and taste.
Tooling has been democratized. Every team has access to roughly the same models, the same automations, the same drafting assistants. What can't be commoditized is your understanding of your specific audience. Competitive advantage lies in the friction points buried in sales call recordings, the delight moments your CS team hears every week, the questions your prospects are actually asking in Reddit threads and G2 reviews.
"That's the biggest treasure trove of first-party data — a goldmine that is exclusively yours and no one else's." – Uzair Dada
Taste shows up in the second half of the workflow. The old hand-off model was that a strategist or writer got 20% of the way and passed it to a creative for the remaining 80%. The new shape is the opposite. Marketers are expected to arrive at 60–80% (something like research done, draft sketched, and point of view formed) so the conversation can start at a much later stage. That raises the floor on what every contributor brings to the table.
What should marketing leaders stop doing in the age of AI?
A lot of the post-Claude reset is subtractive. Uzair's pointed list:
- Stop adding tools before you delete tools. One CIO he respects has a flat rule: nothing new gets purchased until something existing gets removed.
- Stop pretending lead gen is mid-funnel. Paid media that drives form fills is still top-of-funnel work — calling it anything else makes the metrics lie to you.
- Stop fighting the brand-vs-demand war. Authoritative third-party content, community presence, and analyst relations are now the highest-credibility signals LLMs use when deciding who to cite. Brand work is discovery work.
- Stop using attribution to prove marketing. Uzair borrows a line from Marketo co-founder John Miller: use attribution to improve marketing, not to defend its existence.
Uzair also points to a HubSpot data point worth sitting with: deals that buyers say were sourced through an LLM close at roughly 4x the rate of other inbound. Vanity metrics on traffic won't show it — but conversion rate will.
How should B2B marketing teams be structured for AI?
The org chart of the near future, Uzair argues, looks heavier on four roles: strategy, taste, judgment, and builders. It will be lighter on the rinse-and-repeat execution layer that AI is steadily absorbing.
But there's a real gap he flags candidly: most AI tooling today is still an individual sport. Sharing a workflow, versioning a custom skill, governing what gets used and how —that team-level harness barely exists yet. The marketing orgs that figure out how to "bench press for the whole team" instead of one person at a time will pull ahead fast.