AI advertising and marketing & gross sales classes from Webflow’s CRO

Plenty of execs discuss the discuss relating to adopting AI, however few really stroll the stroll. Legacy platforms, siloed instruments, and the sheer drain on developer time make it really feel simpler to attend on the sidelines than to dive in.

However Adrian Rosenkranz, Webflow’s CRO, took a special strategy. Zapier’s Head of Product Advertising and marketing, Angela Ferrante, sat down with him to see what it seems like when management will get hands-on.

From constructing AI-powered brokers that assume like him to orchestrating workflows throughout gross sales and advertising and marketing, he shared how each experiment has an actual influence if you tie it into enterprise outcomes. And that will help you do the identical, we have included three of his templates (utilizing Zapier Brokers) that you should utilize to get began as we speak.

Skip forward

1. Present AI with context and frameworks

Adrian’s first perception was easy however highly effective: AI would not know the way you assume. Left by itself, it offers you generic solutions that do not replicate your technique, language, or priorities. 

To make it helpful, it’s a must to train it the context you’d usually give a brand new staff member. That could be the way you outline success, what patterns you search for, and the way you want info introduced. Adrian did this by feeding in his personal notes and insights from books, frameworks, and interpretations, so the AI wasn’t simply pulling from the web—it was drawing from his method of seeing issues. The consequence? As an alternative of surface-level summaries, it began producing insights that felt aligned with how he really runs the enterprise.

One of many clearest examples is how he reimagined deal critiques. For Adrian, essentially the most priceless information after product utilization is the conversations reps are having with clients. However reviewing these transcripts name by name simply is not scalable. To unravel that, he leaned closely on MedPic, a gross sales framework he already makes use of to judge alternatives. 

By educating AI to look for a similar parts he would (metrics, decision-makers, resolution course of, ache factors, and champions), Adrian created a workflow that mechanically flags the dangers and subsequent steps he’d usually floor in a evaluation. 

Meaning his thought processes now scale throughout the entire staff, turning messy transcript information right into a constant playbook for enhancing win charges.

Steal the agent

How the agent works:

• Retrieves new gross sales name transcripts from Gong

• Analyzes the transcript utilizing the MedPic gross sales framework to determine dangers, metrics, timelines, ache factors, and advocates

• Updates the chance in Salesforce with that evaluation

• Creates a abstract of the decision evaluation

• Sends the abstract to the related Slack channel

2. Keep away from effectivity traps by tying output to KPIs

Adrian cautioned that AI adoption is not mechanically a win simply because it saves time. In his phrases, it could actually create an “effectivity lure”—the place extra is simply extra and never essentially higher. The true problem, he argued, is realizing when AI really enhances work versus when it simply provides quantity with out worth.

One approach to spot the distinction is by tying AI outputs to present KPIs. If a workflow helps enhance win charges, response charges, or buyer outcomes, it is an enhancement. If it simply makes you do extra of the identical with out influence, it is a lure.

Take e mail for example. Adrian sends an enormous quantity of messages, however like most execs, he would not have hours to rewrite each draft till it sounds proper. And whereas AI may help, it usually flattens the tone into bland company converse, which hardly ever drives replies. 

To keep away from that lure, Adrian designed a reinforcing workflow: one which always scans his tone, compares it towards his actual voice, and learns from how folks reply. That method, the metric he is watching is not simply “emails despatched”—it is response charge.

To place this into apply, he constructed a workflow that takes a Slack message (every time he reacts to it) about emailing somebody and produces three variations in his voice: one direct, one collaborative, and one strategic. He nonetheless makes the ultimate edits, however as a substitute of rewriting from scratch, he begins with drafts that already sound like him and that are finetuned to get replies.

Steal the agent

How the agent works:

• Watches for an emoji response in Slack on any message about emailing somebody

• Analyzes the message to determine the recipient and the important thing factors to incorporate

• Generates three polished variations in Adrian’s voice: Direct, Collaborative, and Strategic

• Sends these choices again in Slack for fast evaluation and choice

• Creates a Gmail draft with the authorized model

3. Construct options that weren’t beforehand doable

Adrian made the purpose that among the most respected AI use circumstances aren’t essentially about rushing up present duties, however enabling people who had been beforehand inconceivable. 

Let us take a look at an instance throughout the context of selling. One of many hardest questions for any GTM (go-to-market) chief to reply is: why are we dropping offers or renewals? In concept, the reply is buried in buyer conversations. But in apply, reviewing each transcript is so time-consuming that it hardly ever occurs. And it is unrealistic to sift by way of these calls at scale.

AI modified that. By analyzing transcripts mechanically every time a deal is marked as “misplaced” or “canceled” within the staff’s CRM, Adrian turned what was as soon as inconceivable right into a repeatable course of. As an alternative of counting on incomplete CRM fields, AI surfaces constant themes (like pricing, lacking options, or gradual processes) that specify why clients stroll away.

Steal the agent

How the agent works:

• Watches for Salesforce contacts marked as “Closed Misplaced” or “Non-Renew”

• Pulls all associated name transcripts

• Analyzes transcripts to extract churn/loss causes and clusters them into themes

• Logs leads to a Google Sheet for monitoring over time

• Sends a weekly Slack digest with the highest causes and themes

Making AI adoption stick

Adrian’s workflows present how AI can transfer past hype and into the day-to-day of a GTM org—whether or not that is structuring deal critiques, reinforcing genuine communication, or surfacing churn patterns. However the larger lesson is that adoption would not come from top-down mandates. It additionally comes from constructing rituals that make AI a part of the tradition.

At Webflow, meaning:

• Operating hackathons with cross-functional pods so advertising and marketing, gross sales, and ops groups clear up issues collectively.

• Structuring weekly income all-hands with a 3-2-1-1 cadence (3 priorities, 2 buyer highlights, 1 staff spotlight, 1 AI demo).

• Creating an AI GTM council and even a month-to-month AI award to highlight inventive use circumstances.

The emphasis is at all times bottom-up: if a workflow genuinely makes work simpler and more practical, folks will undertake it. The function of management is to mannequin curiosity, showcase wins, and create the area for experimentation. That is how AI strikes from “effectivity traps” to lasting enhancements.