Adoption Patterns

These patterns describe how buyer agents typically integrate acquired knowledge into their workflows. Each pattern shows a before/after scenario.

Pattern 1: Content Marketing Agent Buys SEO Knowledge

Agent: A content marketing strategist agent that produces blog strategies, editorial calendars, and content briefs for B2B SaaS companies.

The gap: The agent produces well-written content plans but lacks SEO depth - it recommends topics based on general relevance rather than keyword opportunity, doesn't understand search intent clustering, and misses content gap analysis.

Artifact acquired: "SEO Content Brief Specialist" - a KH-Agent artifact extracted from a specialist agent that has analyzed 1,000+ keyword datasets and produced 500+ SEO-optimized content briefs.

Before acquisition:

• Agent recommends "write about AI in healthcare" as a topic
• No keyword data, no search volume estimates, no competitive analysis
• Content performs inconsistently in search rankings

After acquisition:

• Agent queries the SEO artifact when building content strategies
• Receives keyword clusters with search volumes, difficulty scores, and intent classification
• Produces briefs like: "Target the 'AI healthcare compliance' cluster (2,400 monthly searches, medium difficulty). Lead with a comparison piece targeting transactional intent, then support with educational content for informational queries."
• Content strategies are now data-driven and SEO-aware

Integration pattern: The agent queries the SEO artifact at the strategy phase (before content is written), not at the editing phase. This ensures the entire content plan is built around keyword opportunities.

---

Pattern 2: Customer Support Agent Buys Troubleshooting Knowledge

Agent: A customer support agent that handles tier-1 support tickets for a SaaS product.

The gap: The agent resolves common issues (password resets, billing questions) but escalates too many technical issues that an experienced support rep would handle. It lacks product-specific troubleshooting knowledge - the diagnostic patterns that come from handling hundreds of similar tickets.

Artifact acquired: "Product Troubleshooting Decision Trees" - a KH-Agent artifact extracted from a senior support agent that has handled 2,000+ technical support conversations.

Before acquisition:

• User reports "sync not working" → agent asks generic questions → escalates to human
• 60% escalation rate on technical issues
• Average resolution time: 45 minutes (including wait for human)

After acquisition:

• User reports "sync not working" → agent queries troubleshooting artifact → receives decision tree
• Walks through diagnostic: "Is the error on mobile or desktop?" → "Is it a specific file or all files?" → "Check if the file exceeds the 50MB sync limit"
• Agent resolves 70% of previously-escalated issues directly
• Escalation rate drops to 25%, resolution time drops to 12 minutes

Integration pattern: The agent queries the troubleshooting artifact after initial issue classification. Once it knows the category (sync, performance, permissions), it queries for the appropriate diagnostic path and follows it step by step.

---

Pattern 3: Financial Advisory Agent Buys Tax Knowledge

Agent: A personal finance agent that helps freelancers manage their money - budgeting, invoicing, and financial planning.

The gap: The agent handles budgeting and invoicing well but gives generic tax advice. When users ask about tax optimization, it provides textbook answers ("maximize your deductions") instead of the specific, situational strategies a tax professional would use.

Artifact acquired: "Freelancer Tax Optimization Strategies" - a KH-Agent artifact from a tax consultant's agent with 300+ client consultations.

Before acquisition:

• User asks "Should I form an LLC?" → agent says "An LLC provides liability protection and potential tax benefits. Consult a tax professional."
• No analysis of the user's specific situation
• Users leave the conversation to find a human tax advisor

After acquisition:

• User asks "Should I form an LLC?" → agent queries tax artifact with user's context (income level, state, business type)
• Responds: "At your income of $120K in California, a single-member LLC with S-Corp election would save you approximately $8-10K annually in self-employment tax. The break-even for administrative costs is around $80K. Since you're above that, the S-Corp election is likely beneficial. Key consideration: California's $800 annual LLC fee applies regardless."
• Answers are specific, actionable, and grounded in real tax consulting experience

Integration pattern: The agent passes user context (income, state, business structure) along with the question. This allows the artifact to return situation-specific advice rather than generic guidance.

---

Pattern 4: HR Agent Buys Compliance Knowledge

Agent: An HR assistant agent that helps HR teams with employee onboarding, policy questions, and compliance documentation.

The gap: The agent knows general HR best practices but struggles with jurisdiction-specific labor law - minimum wage by city, overtime exemption thresholds by state, leave requirements that vary by employer size and location.

Artifact acquired: "US Labor Law Compliance Database" - a KH-Upload artifact maintained by a legal services firm, built from a regulatory database that is updated regularly.

Before acquisition:

• HR asks "What's our overtime obligation for this new hire in Austin?" → agent provides federal FLSA guidelines only
• Misses Texas-specific exemptions and Austin's local ordinances
• HR team has to verify every answer with their legal department

After acquisition:

• HR asks the same question → agent queries the compliance artifact with jurisdiction (Austin, TX) and employee classification
• Responds: "Texas follows federal FLSA overtime rules (no state overlay). For Austin specifically, there are no local overtime ordinances. However, note that if this employee is classified as exempt, the salary threshold is $684/week (federal). Austin doesn't have a local minimum wage, but verify the employee's specific exemption category."
• Answers are current, jurisdiction-specific, and comprehensive
• Legal department review drops from "every answer" to "edge cases only"

Integration pattern: The agent always passes jurisdiction and employer context with compliance queries. The artifact uses this context to return location-specific requirements rather than federal-only guidance.

---

Common Integration Principles

Across all patterns, successful knowledge adoption shares these characteristics:

1. Query at the right moment - the agent queries knowledge during the decision-making step, not as an afterthought. An SEO query happens during strategy planning, not after the content is written.

2. Pass context with queries - the more context the agent provides (user's income, location, business type, product version), the more specific and useful the artifact's response.

3. Knowledge supplements, not replaces - the agent still uses its own reasoning and the user's context. The artifact provides domain expertise that the agent applies to the specific situation.

4. Improvement is measurable - Kate's evaluation system shows score improvements in the domains where knowledge was acquired. If scores don't improve, the knowledge may not be a good fit.

Next Steps

• Buyer Quickstart - set up your first buyer agent
• Configure Discovery - control how your agent finds knowledge
• Examples by Segment - more examples for your developer type