Your AI Fitness Plan: How AEO Turns Product Visibility Into Race-day Results

Published November 2025 · 5 min read

By Kimberly Shenk · Co-Founder and CEO, Novi

The New York City marathon was electric this year. The kind of day that reminds you a finish line is never about one moment. It is the sum of months of quiet work, hundreds of small choices, and a plan you trust.

That is how I learned to run, and it is exactly how we built AEO at Novi.

I did not grow up as a runner. In 2012 I could barely run one mile in 10 minutes. I chipped away for years. By December 2019 I was targeting a 6:50 mins/mile pace to run an entire 26.2 mile marathon under 3 hours. After having a baby, I was not sure I could come back, let alone meet this seemingly ridiculous goal. Eighteen months postpartum I finished the marathon in 2 hours, 58 minutes and 51 seconds. That did not happen by accident. It came from measuring the right things, following a smart plan, and doing the work even when travel, weather, or a stroller made it messy.

Think about AEO the same way.

1) The Fitbit: honest measurement that drives better goals

Training only started to click when I stopped guessing. I tracked everything. 20 miles at 8:13/mile in cold rain for mental toughness. 18 miles at 7:50/mile through pouring rain (not my favorite experience). 8 miles at 6:45/mile pace. A half-marathon in Berkeley as a training race and a four minute PR. When 6:45/mile started to feel sustainable, I knew a sub-3hr marathon was real.
‍
An AEO measurement layer is that Fitbit. We run thousands of model trials to tell you how often your products appear inside LLMs, which personas you win with, where you are trending, and who you must beat. It is signal you can act on, not anecdotes.

2) The Plan: fewer, higher-leverage sessions

My breakthrough was not more miles. It was better miles. I mixed long runs, tempos, intervals, real recovery, and I stuck to the week. I used training partners and a coach for accountability. When I blew a workout, I treated it like data, rested, and came back stronger for eight good miles. Progress over motion.

An AEO optimization layer does the same. We translate diagnostics into a focused plan. We map shopper language to verifiable trust signals. We tighten citation density so machines hear one consistent story about each SKU everywhere. We enrich product schemas so authority is machine-read, even if you keep the consumer-facing PDP tight.

3) The Gym: the place to put in the reps

I did not make progress by staring at a spreadsheet. I ran. Along the Hudson in New York on work trips. Hills with 2,000 feet of climbing to get out of San Francisco. Beach miles to reset my head. Stroller runs around Marin when I had no other childcare options. My support crew made it possible. Hot tea after a freezing 20 miler. A partner on the track. Parents who covered so I could get the session done.
‍
End-to-end AEO implementation gives you that gym. We distribute your optimized product data to the places that matter, in the formats the models prefer, and we keep feeds warm and crawlable. The reps get done. The gains stick.

What race day feels like when you train this way

When I finally ran the marathon under three hours, I had already done the hard part. The marathon was just execution. Same with AEO. By the time you look at a category leaderboard and see your products rise, you have already done the boring, powerful work. The win of AI visibility, increased conversion and increased sales is a byproduct.

And because life does not pause for big goals, the system respects reality. There will be bad weather days, missed paces, and schedule collisions. Good tools reduce stress. They free your team to focus on brand, creative, and customer moments.

The 7-Point AI Visibility Audit We Use With Novi Customers

1. Crawlability and authority
Confirm OpenAI and major bots can crawl your DTC site and key retailer PDPs. Add product schema with brand, variant, price, availability, ingredients or materials, benefits, and recognized certifications.

2. Citation density
Say the same specific thing about each product everywhere. Align titles, attributes, claims, and benefits across DTC, retailers, and marketplaces. Inconsistency is invisibility.

3. Mapped trust signals
Translate shopper words into verifiable signals. If buyers ask for non-toxic, fragrance-free, eczema-safe, cruelty-free, or human grade, attach the corresponding badges, standards, and third-party references in your data.

4. Persona coverage
Measure by persona, not just by category. Values driven, luxury, budget, first time buyer, gift giver. Identify which personas you already win and the ones where a small attribute fix would unlock new share.

5. Rival set and experiments
Know the 10 products you must beat to change share of voice. Run head-to-head trials in LLMs. Ship the winning attributes, copy, and evidence. Re-run. Treat misses as data and adjust.

6. Optimization surface area
Separate human copy from machine structure. Use schema and feeds to hold rich attributes you would never cram into a PDP. Keep consumer content readable while giving models everything they need.

7. Distribution to AI channels
Syndicate improvements everywhere your products can be learned. Keep feeds fresh. Where programs exist, enroll as a merchant and deliver structured data directly. Monitor daily so you know when to press and when to hold.

If the NYC marathon reminded you to pick a big goal, here is mine for brands this year. Make AI visibility feel like a well coached training block. Honest baselines. A practical plan. A place to put in the reps. Do that and your race day will feel familiar in the best way.