/
PH1 core capbility
YEARS EXPERIENCE
3 to 5 years
TYPICAL CLIENT
CTO, VP Product, Head of AI/ML, Founder/CEO
NECESSARY TIMELINE
2 to 3 months
BUDGET NECESSARY
Quoted individually
Our POV
Every organization is asking the same question: "Where do LLMs fit in our business?" Most are answering it badly — chasing vendor demos, building chatbots nobody asked for, or experimenting with foundation models without understanding what their customers actually need from an AI-powered experience.
PH1's LLM Product Strategy & Specification practice starts with the business opportunity, not the technology. The output is not a recommendation to "use Claude" or "deploy GPT." It is a product strategy and proven workflows that can be monetized. Our focus is on defining a successful AI product from use case, inputs/outputs/outcomes, orchestration, evals, and product delivery optimization. Your engineering team knows exactly what to build, what data it needs, what the user experience must deliver, and what success looks like — grounded in behavioral evidence from your specific market.
What We Do
Opportunity identification. We research your customers, your workflows, and your competitive landscape to identify where LLMs can create a new business opportunity — a new product, a new feature, a new revenue stream, or a fundamentally better experience that your competitors have not yet figured out how to deliver.
Customer needs research. We conduct behavioral research with your target users to understand what they actually need from an AI-powered product — what tasks they want accelerated, what interactions they would trust an LLM to handle, and what the trust conditions are for adoption.
Data and capability alignment. We map the gap between the LLM opportunity and what your organization can actually deliver — assessing your internal data assets, your organizational capabilities, and your technical infrastructure. We define what needs to change to close the gap.
Product specification. We produce a development-ready specification for the LLM-powered product or feature: what the model must do, how well it must do it, what the user experience must deliver, and how the product connects to your business model. The specification is structured around the AI Product Calibration framework.
What We'll Deliver
LLM opportunity analysis: where language models create genuine new business value for your specific organization
Customer needs research: what your target users actually need from an AI-powered product and what the trust conditions for adoption are
Data and capability gap assessment: what your organization needs to deliver the opportunity
Development-ready product specification structured around Power, Speed, Impact, and Joy
Build vs. buy vs. partner analysis for the specific LLM approach
Business case with adoption projections grounded in behavioral evidence
When This is Essential
Before writing product requirements for an LLM-powered product
When the organization has a vague "we should use LLMs" mandate and no clear direction
When you are deciding how to align your data, capabilities, and customer needs toward a specific LLM opportunity
Before a significant investment in data infrastructure or model development
When different teams are proposing conflicting LLM directions and leadership needs one specification
Frequently Asked Questions
Do you help us pick between GPT, Claude, and Gemini?
That is not the primary focus. PH1 identifies WHERE LLMs create business value and what the product must deliver — the vendor decision typically follows from the specification, and your engineering team is better positioned to make the final technical choice based on our requirements.
What data do we need to have ready before you start?
Less than you might think. PH1 starts with users and workflows, not data. We assess your existing data assets during the engagement and identify what needs to change to deliver the opportunity — often revealing that the data gap is smaller than the team assumed.
How do you handle proprietary or sensitive data?
PH1 signs client-specific NDAs and operates under your data governance rules. For sensitive data, we work with anonymized or synthetic datasets during the research phase.
Can you work with our engineering team directly?
Yes. The specification is designed to be handed to your engineering team, and PH1 typically runs working sessions with engineering leads to translate the specification into implementation plans.
Can this inform a fundraising conversation?
Yes. The LLM opportunity analysis and business case are frequently used by founders in investor conversations to articulate the venture's AI strategy with behavioral evidence.
Combine With These Services
AI Concept Validation — Validate the LLM use case behaviorally before defining the product specification
New AI Product Market Research — Extend LLM strategy with ICP trust analysis and monetization evidence
AI Product Evaluation & UX Evals — After build, evaluate the LLM product against the specification criteria
/
Submissions