Chatting with Venture Capitalists: Some Questions on Parsnipp & GEO
We are constantly having conversations with all of the stakeholders across the emerging GEO category -- brands and marketers, agencies, technologists, tech media, and investors. Given many people are still learning and catching up on the space, we often get similar questions about Parsnipp and AI Search marketing in general.
We took an email thread with a venture capitalist recently and turned it into a blog post that we hope will be useful for those asking similar questions.
Data Sources
VC Question: As we move increasingly to multi-modal content absorption (e.g. voice, visual content etc.) what are the key data sources Parsnipp looks at to measure brand visibility and share of voice? Or is text the key format for search, and will that continue to be true?
Answer: Our north star isn't LLM visibility — it's helping brands understand and influence interactions between consumers and AI.
That said, today LLMs are the primary consumer application of AI where the interactions most relevant to brands are happening. As AI evolves, consumer behavior will evolve, and so will Parsnipp to make sure we put brands in the best possible position to win in the era of AI. As AI becomes more multi-modal, so will the tools marketers need to compete for visibility and engagement; whether that's vision, voice, spatial, or whatever comes next. If consumers are using it, brands need strategies to market across it.
There's a lot of noise around AI as a technology, but we're taking a marketing practitioner's approach: helping brands react and optimize for the changes AI is having on consumer behavior. In many ways this mirrors what brands go through whenever a new channel gains traction — TikTok, mobile, email, SMS, whatever — except the tactics and tools all need to be reinvented to fit the discipline of marketing through AI.
What are the hardest integrations — technically and organizationally?
The integration load isn't the high hurdle. We've built into the relevant AI search marketing platforms and can solve the technical challenges marketers need us to solve. The bigger lift is subject matter expertise: simplifying the problem, making it digestible for marketing buyers, and focusing on clear prioritization and value — not another shiny exploratory AI tool.
Personas
VC Question: How do you build personas that mirror real consumers (e.g. embedding realistic context)? How prescriptive can a user be on demographic cuts, geos, etc.?
Three core ways:
- Preset personas based on common consumer buyer groups for those who want to set up fast and start testing.
- Toggle across demographic and psychographic attributes to build a persona the same way you'd build a digital ad audience.
- Integrate first-party data from your CRM, eCommerce platform, etc. to build custom personas based on the brand's own data the most accurate replication of their actual shopper.
The more accurate the persona, the higher the efficacy of the resulting AI search marketing data. Consumer context has a huge impact on LLM outputs.
Brands can also run many personas in parallel to segment AI search marketing data in meaningful ways.
How do you benchmark and think about accuracy?
Great question — and a hard problem in a new discipline without deep domain expertise across the market yet.
We have our own models and are constantly testing, but the questions we benchmark against with customers are:
- Are we simulating the most realistic interactions possible between people and AI to give brands high-efficacy AI search results?
- Are we scoring and benchmarking overall GEO Score accurately? We've built our own aggregated scoring system that accounts for hundreds of attributes — constantly being tweaked and optimized.
- Are we preparing for new metrics that will emerge as AI search marketing matures? With agentic commerce, LLM ads, and more engagement and targeting analytics coming directly from the labs, this space will evolve fast. Parsnipp needs to stay at the bleeding edge to be brands' trusted partner for AI search marketing.
Beyond our own metrics, we also make it easy for brands to track the other relevant KPIs of AI search marketing. Not all are super actionable yet, but they're worth watching:
- Traffic via LLMs
- Leads and revenue via LLMs
What are examples (beyond Shopify) of how users can integrate first-party data directly? E.g. proprietary survey data?
Survey data works if it's readily available, but we'll primarily rely on existing marketing stack integrations so it's easy for marketers to implement — CRM, marketing automation, ad tech, eCommerce and shopper data, ratings and reviews, social media, etc.
What are the costs of building personas, and how do you expect these to scale as users derive more value from Parsnipp and increase usage?
Aside from integration work to pull in more datasets over time for more accurate personas, more personas really just means more simulated interactions between consumers and AI. The main cost variable is tokens with OpenAI, Gemini, Anthropic, etc. Hasn't been an issue so far and we are doing out best to make Parsnipp and introductory GEO tool that allows brands to test and start optimizing without hitting budgetary constraints.
To what extent is it important for Parsnipp to build these capabilities in-house vs. partnering with others (e.g. Simile, Listen Labs, etc.)?
It's a use-case-by-use-case analysis. We absolutely don't believe we need to build all insight capability ourselves and we can and will partner in a range of areas while building in others.
Where Parsnipp is uniquely positioned is our ability to deliver a the platform to manage and deploy campaigns, targeting, and optimization on the channel of AI itself. We're building the marketing system of record for AI, the way HubSpot is for email marketing, Criteo for digital ads, Sprinklr for social, Shopify for eCommerce, etc. (this is an oversimplification of these companies of course).
Recommendations
How opinionated does Parsnipp ultimately need to be on the "what" vs. the "how" e.g. "add more attribution and factual density to improve AI confidence" vs. "improve your social media presence"?
Both — I don't think those are mutually exclusive. We should be expert, trusted partners and thought leaders for brands on how to drive measurable marketing value from AI channels.
Everything Parsnipp does needs to be reinforced by testing and hard ROI analysis. Marketing across LLMs can't be a "pie in the sky" marketing strategy. It needs the same data fundamentals as buying search ads, sponsoring a PR placement, or running an affiliate program with influencers. That said, downstream from really good recommendations and strategy is the execution. We can help there via more agentic solutions within Parsnipp, agency services, etc.
On the roadmap, is creating content important or is it better to qualitatively guide?
Both, with nuance. A few other entrants have decided that generating infinite agentic content to flood LLMs is a long-term GEO strategy. We believe there are real content and communications strategies that stem from good AI search marketing insights, but it's more nuanced than an "AI slop content strategy." There are also so many other parts of a brand's digital footprint that can be optimized to improve AI search marketing readiness, particulary as we move closer to widespread adoption of agentic commerce.
Ideally Parsnipp delivers both: strong recommendations and best practices baked into the platform, plus solutions to help marketers execute not just content, but web optimization, eCommerce and agentic commerce enhancements, etc.
How do you track improvements in AI visibility against individual recommendations to help users double down on what's working?
We close the loop by continuously running our AI search analysis and simulating interactions between AI and consumers. This lets us track brand visibility over time, what GEO strategy changes brands are making, and ultimately build stronger modeling that correlates tactics to ROI. From here we recommend the highest impact actions brands should take, whether that be spending more time and resource on existing activity or solving other gaps.
Do you then track performance metrics? Can a customer calibrate Parsnipp with their internal engagement data, e.g. conversion metrics?
For some metrics, yes. Some are measurable in other systems and can be used to validate Parsnipp's data. For others, the space is still new enough and the data accessible from leading LLMs limited enough that Parsnipp is the only place marketers can go for this type of insight today (aside from running a another GEO toolside-by-side to test and compare).
Agency Model
How do you plan to scale the agency? Is this primarily a GTM flywheel trigger, or is the service model a long-term differentiator?
The agency isn't there to scale in the traditional sense as a standalone business; it's there to make sure we can be as high-touch as needed and solve the GEO challenge fully for marketers, even as the discipline evolves fast and the platform changes with it. It is an extension of customer success, a "forward deployed" function, and runs as a complementary but separate team to our other forms of customer support.
Ultimately we want to work as closely as possible with brands to solve the new marketing challenges of AI and the agency lets us do that. In the early days of GEO, that's an asset since most brands are still building domain expertise in GEO internally.
On the flywheel — I think our PLG motion will be a stronger flywheel than the service-led agency because it's our core business. But we'll also be selling to agencies, which I do think helps with adoption since a lot of marketing mindshare for brands still lives inside agencies. Many marketers trust agencies to bring in the right tools to build a best in class marketing stack.
Product Roadmap and Long-Term Vision
Where is Parsnipp focused today?
- Help brands understand and benchmark how AI is talking to shoppers about their brand, products, competitors, and category.
- Build an easy-to-use set of recommendations to help brands improve their position in AI.
- Track ongoing progress across LLM channels and give marketers the tools to continuously optimize and win.
- Be the trusted partner brands rely on to both implement GEO strategies and upskill their teams on AI marketing — because marketers can't afford to also become AI search experts on top of everything else they manage, at least not right away.
What's coming next?
- New use cases most relevant to marketers: agentic commerce and LLM ads.
- New forms of AI that consumers are adopting and that marketers need to start planning and optimizing for now.
