Oct 13, 2025
Brand, AI, and the new growth math: Lessons from Webflow’s Josh Grant
Pranav Piyush
,
Co-founder, CEO
Connect with Pranav
The bottom line
Treat LLMs as real distribution, teach them with credible answers, and wire your org so product, marketing, and finance run one plan. Prove lift where it matters, move with judgment where data is thin, and let brand set the ceiling on everything else.
AI is a distribution channel you can influence
Large language models are already reshaping discovery. When ChatGPT added direct links, Webflow saw classic direct traffic give way to LLM referrals. The response was to teach the models, not chase scrapes. Josh’s team publishes content models can reason with, especially deep FAQ and Q&A pages with precise definitions and clear steps. The goal is simple, make Webflow the answer models trust.
Inside the org, AI compresses cycle time. Lightweight automation scrapes community threads, summarizes themes, and routes insights to growth and product within minutes. Josh’s mental model is threefold, use AI for internal leverage, build with and for AI so the product appears in model outputs, and track how AI shifts market dynamics so strategy stays current.
What changed fast
LLMs now send measurable referral traffic, so AEO sits beside SEO
Content that explains and instructs beats content that only ranks
Internal AI workflows turn weeklong research into same day inputs
Tracking LLM referrals shows how answers convert, not only pages
Scale the insight, not the channel
“Scale what works” is a trap if you do not know why it worked. Josh asks teams to isolate message, moment, and intent, then export that insight across channels. He pairs a centralized experiment roadmap with readouts that document the human reason a winner won. He also keeps a ritual that sharpens creative judgment. He screenshots ads he almost clicked but did not, then reverse engineers what failed to earn trust.
At Affirm, this mindset produced a next best action engine that predicted where a user would spend next and which touchpoint would move them. Two people in the same campaign saw different messages and channels because the model optimized both destination and influence. Universal holdouts proved lift, which meant personalization came with causal evidence.
Incrementality that keeps speed
Affirm trained a habit of payback, LTV, and incremental lift on every dollar. Josh brought that rigor to B2B and PLG while accepting that sample sizes are thinner. His rule is to earn budget with causal proof where it matters, then protect velocity elsewhere. Holdouts, geo or time splits, and honest conversion windows build confidence without slowing the team. Growth is math, and it is also the story that explains why a decision is right today.
One growth system across PLG and sales
Webflow kept its efficient agency and freelancer motion, then built an upmarket path by mining self serve behavior for propensity signals and graduating the right accounts to sales assist and enterprise. The company runs this as one system with a small self serve crew and a single roadmap.
Who sits in the room
Growth marketing, accountable for acquisition and spend
Product growth, accountable for activation and expansion
Strategic finance, accountable for pacing and ROI
Josh, accountable for incremental self serve ARR
Marketing brings data that exposes product friction, product ships the fix that lifts marketing efficiency, and finance keeps the score. The setup gives Josh GM style ownership while preserving speed.
Brand sets the ceiling
Josh’s hot take has become operating truth. Experiments can buy efficiency, brand multiplies the machine. When the story lands before the click, every performance metric improves. In a world where AI levels tactics and playbooks spread quickly, distinctiveness and clarity of value become the moat.
That is the new growth math, teach the models with credible answers, prove lift where it counts, move quickly when judgment is enough, connect PLG and sales through one crew, and let brand raise the ceiling on everything else.
