Product Frameworks for the AI-Era Enterprise

When to Apply

• Entering a new market or segment where problem framing is unclear
• Evaluating whether an AI capability should be assistive or autonomous
• Reframing a stagnating product around unmet outcomes

AI-Era Notes

Separate the human's job-to-be-done from the agent's job-to-be-delegated. Document escalation triggers where the agent must return control. Capture the trust threshold required before customers delegate irreversible actions.

Common Antipatterns

• Using JTBD as a marketing exercise without evidence
• Assuming AI can perform emotional or social jobs without user consent

Pairs With

When to Apply

• Product teams shipping regularly but uncertain about impact
• Enterprise contexts where customer access is gated and precious
• Any environment where AI synthesis of research is being introduced

AI-Era Notes

AI can accelerate transcription, clustering, and pattern surfacing — it cannot replace direct customer contact. Enforce source citations on every AI-derived insight and require human review before an insight influences prioritization.

Common Antipatterns

• Replacing interviews with AI persona simulation
• Losing verbatim quotes and provenance in summarization
• Running discovery without a decision it will inform

Pairs With

When to Apply

• Translating a strategic outcome into a portfolio of bets
• Deciding whether a problem is best solved with UI, workflow, or agent
• Communicating trade-offs to executives without over-committing to solutions

AI-Era Notes

Tag each solution branch by autonomy tier and reversibility. AI-native branches inherit additional preconditions: eval set, guardrails, monitoring, and rollback authority.

Common Antipatterns

• Trees that stop at solutions and skip experiments
• Uniform treatment of AI and non-AI branches

Pairs With

When to Apply

• Quarterly planning where AI features materially affect the outcome
• Aligning cross-functional teams on both business and model quality
• Any AI feature where user harm is possible from silent regression

AI-Era Notes

Every AI-affecting objective should carry at least one leading model metric alongside the lagging business metric. Regression in model metrics is a leading indicator of business regression.

Common Antipatterns

• Treating eval scores as vanity metrics divorced from outcomes
• OKRs that measure only shipping activity, not customer or model health

Pairs With

When to Apply

• Organizations with fragmented metrics and unclear priorities
• Products where AI outputs may inflate activity without value
• Aligning multiple product lines around a shared measure of success

AI-Era Notes

Define whether AI-assisted actions count toward the metric and how. Undifferentiated counting inflates metrics without reflecting real customer value.

Common Antipatterns

• Choosing a metric that cannot be moved by the team
• Ignoring model-driven noise in the input metrics

Pairs With

When to Apply

• Prioritizing across a mixed portfolio of AI and non-AI initiatives
• Regulated environments where reversibility materially changes cost
• Roadmap trade-offs requiring executive-visible justification

AI-Era Notes

Add a risk multiplier that penalizes low-reversibility, high-blast-radius, or model-dependent work. Shipping an autonomous capability is not equivalent to shipping a UI change.

Common Antipatterns

• False precision in confidence scores
• Prioritizing agent capabilities purely on reach without risk weighting

Pairs With

When to Apply

• Any capability powered by an LLM, model, or autonomous agent
• Regulated products requiring documented acceptance criteria
• Cross-functional work spanning product, engineering, and AI/ML

AI-Era Notes

Acceptance criteria for AI features must be expressed as eval sets, not prose. The eval spec becomes the contract between product intent and model behavior.

Common Antipatterns

• AI PRDs with prose acceptance criteria and no measurable evals
• Eval specs written after the model is already in production

Pairs With

When to Apply

• Decisions with long-term architectural or product consequence
• Model, vendor, or dataset selection for AI capabilities
• Any decision that establishes organizational precedent

AI-Era Notes

Model selection is a decision that expires. Every model ADR should include the eval basis for selection and a review trigger for re-evaluation as the model landscape changes.

Common Antipatterns

• ADRs written after the fact to justify a decision
• Missing model version and eval basis in AI-related decisions

Pairs With

When to Apply

• Teams shipping continuously while validating new opportunities
• Organizations introducing AI features into existing products
• Environments requiring both learning velocity and shipping discipline

AI-Era Notes

Discovery-track AI experiments must never leak into production without crossing delivery-track gates. The evaluation set is the artifact that transitions between tracks.

Common Antipatterns

• Discovery-track prototypes shipped to real customers without evals
• Delivery track blocked waiting for perfect discovery

Pairs With

When to Apply

• Before production release of any model-driven capability
• Continuously across model, prompt, or context changes
• In response to observed customer harm or regulator inquiry

AI-Era Notes

Evals are not a one-time test; they are a first-class product artifact with version control, ownership, and review cadence equivalent to code.

Common Antipatterns

• Evals maintained by a single individual without redundancy
• No adversarial or safety cases in the eval set

Pairs With

When to Apply

• Any agent capable of taking action beyond retrieval
• Customer-facing or regulated workflows involving automation
• Internal agents with access to sensitive data or systems

AI-Era Notes

Autonomy tier is a product decision, not a technical one. It carries governance, liability, and support implications that must be reviewed by legal and executive stakeholders.

Common Antipatterns

• Undocumented tool access granted to agents
• No defined threshold for human escalation
• Missing kill-switch authority or runbook

Pairs With

When to Apply

• Products making consequential decisions affecting customers
• Regulated environments including healthcare, financial services, employment
• Any deployment requiring model cards or algorithmic accountability disclosures

AI-Era Notes

Model cards and system cards are becoming procurement requirements, not academic artifacts. Treat them as first-class deliverables owned by product.

Common Antipatterns

• Responsible AI relegated to a legal review at the end of the cycle
• Transparency artifacts that are inaccurate or out of date

Pairs With

When to Apply

• Before wiring an agent to any internal system, SaaS connector, or MCP server
• When computer-use or browser agents can click, send, file, or pay
• When a tool catalog has grown faster than product ownership

AI-Era Notes

A tool grant is a product launch. Adding Slack send, Jira write, Salesforce update, or file-system access is a change of blast radius, not a config toggle. Re-attest the contract on any new tool, new scope, or new model that can call it.

Common Antipatterns

• Granting tools because the demo looked better
• Shared admin credentials used as agent identity
• Tool lists maintained only in engineering config with no product owner

Pairs With

When to Apply

• Before shipping any retrieval-augmented or multi-agent workflow to production
• When agents can read tickets, email, EMR/EHR notes, CRM records, or private docs
• When hallucination, stale policy, or cross-tenant leakage is a material risk

AI-Era Notes

Prompt engineering is a subset of context engineering. Product owns the context architecture: which sources are in-scope, how they are selected and compressed, how memory persists across turns, and what must be cited. Re-attest the contract when sources, embeddings, chunking, or memory policies change — those are product launches, not infra tweaks.

Common Antipatterns

• Dumping entire corpora into the window and calling it RAG
• Shared memory across tenants, cases, or customers without isolation
• No freshness SLA (agents citing expired policy as current)
• Citation optional for consequential answers

Pairs With