AI in Practice: From AI Selection to Go-to-Market Readiness

At EM’s Summer Soirée in Redwood City, Petra Neiger and Lakshmi Peri showcased practical AI systems that marketers can use today. Petra demonstrated how to get your brand recommended by AI platforms, and what to do if it’s not happening yet. Lakshmi presented an agentic AI workflow that turns weeks of cross-team status-chasing into a single, governed dashboard. Here’s a brief recap of what they shared.

Increase your chances of being recommended by AI 

Petra Neiger has spent her career building and transforming marketing functions from the ground up for startups and global enterprises. She opened with a story from her own consumer brand, Red Pantz. She wanted to give her USDA-certified organic skincare products higher priority in her portfolio since this certification is so unique and rare in the market. Her goal was for AI to start recognizing and recommending her MY24/7 by Red Pantz organic skincare line. As an experienced AI marketer, she discovered that three layers have to work together for AI to consider a brand for recommendation. This revelation led her to the creation of the DUV™ framework:

1. Discoverability. Can AI find you at all?
2. Understanding. Does AI know what you actually want to be known for?
3. Validation. Does AI trust third-party signals enough to recommend you?

When she ran the new AI Selection Audit on how ChatGPT, Claude, Gemini, and Perplexity evaluated the offering, two of the four had misclassified it as “female erotica and apparel.”

“That was a total shock to me,” Petra said. Everything looked fine when she typed questions in one at a time. The problem only surfaced once she ran systematic, live probes across all four platforms at once.

Petra says you have to think of this as optimizing systems, not content. “You’re building a spider web of interconnected signals,” she said, all pointing to the same message, and it has to be actively managed.

Here’s what fixing each layer actually has looked like for Red Pantz so far:

• Discoverability: Restructured the site’s schema and built out entries on high-authority knowledge graphs, such as Wikidata, ZoomInfo, and Crunchbase
• Understanding: Reorganized, connected, and built out new content, including a dedicated organic skincare content hub
• Validation: Became active on off-website platforms, review outreach, PR pitching to build other trust signals largely outside her direct control

To track all of it, she vibe-coded an AI Selection Audit™ tool: agents that send live probes to Claude, ChatGPT, Gemini, and Perplexity, benchmark her brand against named competitors, and generate a scored report with 30/60/90-day recommendations. Her score started in the 40s. It’s climbing, but there’s no finish line: “It’s like gardening. You plant the seeds, and then you start watering the garden. And one day, it will happen.”

Since AI systems are built on proprietary technologies, there’s no guarantee that you will be recommended. However, what the tool is designed to do is to help you uncover how AI finds, understands, and validates you to increase your chances of being selected.

Automate your go-to-market readiness reports

Lakshmi Peri has spent 18+ years in enterprise program operations and GTM transformation. For the last two, she’s been building AI agents for go-to-market program operations, starting with workflows where friction is highest, frequency is constant, and slow coordination is most visible. In her demo, she focused on go-to-market feature launch readiness.

The problem is familiar to any marketing or program manager. Before a launch, you’re manually chasing sales, marketing, engineering, and product for status, then compiling it into a deck for leadership. It takes days, sometimes weeks, and by the time it’s presented, it’s already stale.

How her system works:

• Six single-purpose agents, each pulling one signal from Jira, Slack, Zoom, Salesforce, and Google Drive into a shared “operating picture.”
• Scoped access. Each agent reads only the sources it needs, and only within the boundary it’s been granted.
• Human-in-the-loop approval. Nothing updates automatically until an accountable  owner  reviews and approves it first.

The live demo made the payoff concrete:

• One agent tracked sales certification by region, flagging anything above 60% as on track and below as a risk.
• For a hypothetical September 15 launch, the dashboard showed 78% overall readiness, 72% audience quality, and 57% sales-rep readiness.
• Two decisions, including what to do about 340 double-targeted accounts, needed leadership to sign off before the status would update.

Each agent is single-purpose and configuration-driven, so a different program is a config change, not a rebuild.

Lakshmi now advises B2B SaaS and fintech teams on AI portfolio governance and GTM workflow transformation. See her work here.

The Common Thread

Both talks landed on the same idea: AI is only as good as the system you build around it. Petra’s was about building systems that earn trust from AI the same way you’d earn it from a human buyer. Lakshmi’s was about keeping humans firmly in charge of decisions while agents handle the busywork. That’s the kind of hands-on, AI-fluent thinking EM’s consultant bench brings to client work every day. If you want that thinking on your team, reach out to EM.