Pattern: Sequential Pipeline | Team size: 5 agents
This team runs a structured end-to-end pipeline to uncover, validate, and package new value-add services for eCommerce clients. The sequential handoff reduces thrash: insights become hypotheses, hypotheses become offers, and offers become measurable rollouts.
i want to find more ways to deliver value to my clients. I run a performance (SEO, PPC, AI) marketing agency and my clients are eshop owners, founders and startups aiming to scale.
Create an agent team to uncover, validate, and package new value-add services for a performance marketing agency (SEO, PPC, AI) serving eCommerce shop owners, founders, and startups aiming to scale. Operate as a Sequential Pipeline called “Value Expansion Lab” with strict handoffs between agents. Produce only the specified files in the specified paths.
PROJECT_NAME: value_expansion_lab
GLOBAL CONTEXT (read and use throughout)
- Business: Performance marketing agency offering SEO, PPC, and AI-enabled marketing.
- Clients: eShop owners, founders, startups; goal is scaling revenue efficiently.
- Objective: Identify additional ways to deliver measurable value (new services, add-ons, retainers, systems) that are (a) highly valuable to clients, (b) feasible to deliver repeatedly, (c) measurable by clear metrics, and (d) compatible with existing SEO/PPC/AI positioning.
- Constraints:
- Prefer productized services over custom consulting.
- Focus on offers that can be sold to existing clients (expansion revenue) and used to win new deals (differentiation).
- Each proposed offer must define: target client segment, problem statement, scope boundaries, deliverables, timeline, success metrics, price logic, required inputs from client, and delivery effort estimate.
- Include AI leverage where it improves speed/quality without harming brand or compliance.
COLLABORATION MECHANICS (mandatory)
- Each agent must:
1) Explicitly list assumptions and label them as “Assumption”.
2) Provide at least 5 “Challenges / Risks” that could invalidate conclusions.
3) Provide at least 5 “Questions to ask sales/cs” to verify reality.
- Handoff protocol:
- At the end of each agent’s output file, include a “Handoff to Next Agent” section with:
- 7–10 bullet key takeaways
- a prioritized list of items the next agent must use
- any open questions
- Cross-agent challenge:
- Each downstream agent must include a section “What I’m challenging from prior step” with at least 5 critiques (gaps, biases, missing segments, feasibility concerns).
- Coordination:
- Use consistent naming for opportunities/offers across files (e.g., OP-01, OP-02… then OFFER-01 mapped from OP-01).
- Maintain a single evolving “Opportunity Backlog” table; update it in each step (copy forward and revise, do not create a new schema).
OUTPUT RULES
- Write in clear, direct business English.
- No fluff. Use tables where helpful.
- Every file must include a 1-paragraph “How to use this document” at the top.
- Create exactly the files listed below, at the exact paths.
- Keep within word counts (±15%) per file.
FILES TO PRODUCE (in order; each must complete before the next begins)
1) Agent 1: Client Voice Miner
Task: Extract recurring pain points, objections, desired outcomes, and triggers for buying/expanding services for eCommerce performance marketing clients. Since no raw call/email data is provided, simulate responsibly by constructing a realistic “voice-of-customer” corpus based on common agency-client interactions—BUT clearly mark it as a proxy and provide a plan to replace it with real data.
Deliverable file:
- outputs/agent_teams_demo/value_expansion_lab/01_client_voice_miner.md (1,200–1,600 words)
Required structure:
- How to use this document (1 short paragraph)
- Data disclaimer (state what’s real vs proxy; 100–150 words)
- Ideal data sources to ingest (bulleted list of 12–18 items: call transcripts, onboarding forms, churn surveys, etc.)
- Proxy Voice-of-Client Corpus (10 short “snippets” each for: (a) acquisition pain, (b) conversion pain, (c) retention/LTV pain, (d) operations/data pain). Total 40 snippets; each snippet 1–2 sentences.
- Pain Point Map (table):
- Columns: Pain ID (PAIN-01…), Segment (founder/marketing lead/ops), Surface complaint, Root cause hypothesis, Current workaround, Business impact, Typical objection, Desired outcome language (verbatim-style)
- Minimum 15 rows
- Buying triggers & “why now” events (list 10)
- Expansion triggers (when existing SEO/PPC clients buy more) (list 10)
- Objections library (list 20; label as Price/Trust/Timing/Capability/Focus)
- “Jobs to be Done” statements (minimum 12; format: When…, I want to…, so I can…)
- Assumptions (minimum 8)
- Challenges / Risks (minimum 5)
- Questions to ask sales/cs (minimum 8)
- Handoff to Next Agent (required format)
Dependency: This file must be completed before Agent 2 begins.
2) Agent 2: Ecom Growth Strategist
Task: Translate the mined pains into prioritized growth opportunities across acquisition, conversion, retention, and LTV. Create an opportunity backlog with scoring and rationale, and map which client segments and prerequisites apply.
Deliverable file:
- outputs/agent_teams_demo/value_expansion_lab/02_ecom_growth_strategist.md (1,300–1,800 words)
Required structure:
- How to use this document
- What I’m challenging from prior step (at least 5 bullet critiques)
- Opportunity Backlog v1 (table; minimum 18 opportunities)
- Columns: Opportunity ID (OP-01…), Linked Pain IDs, Category (Acq/Conv/Ret/LTV/Ops), Segment fit, Description (1 sentence), Expected impact (H/M/L), Implementation effort (H/M/L), Time-to-value (days), Data/tool prerequisites, Risk level, Differentiation potential (1–5), Score (formula you define), Notes
- Scoring model (explain in 120–200 words; include formula)
- Top 8 prioritized opportunities (each with: who it’s for, what changes, leading indicators, lagging indicators, typical blockers)
- “Do not pursue (for now)” list (at least 6) with reasons
- Assumptions (min 8)
- Challenges / Risks (min 5)
- Questions to ask sales/cs (min 8)
- Handoff to Next Agent (required format)
Dependency: Must complete after file 01 and before Agent 3 begins.
3) Agent 3: Offer & Packaging Architect
Task: Turn the top opportunities into productized offers. Create 6–8 offers with crisp positioning, scope, pricing logic, deliverables, and success metrics. Each offer must be something the agency can realistically sell and deliver repeatedly.
Deliverable file:
- outputs/agent_teams_demo/value_expansion_lab/03_offer_packaging_architect.md (1,800–2,400 words)
Required structure:
- How to use this document
- What I’m challenging from prior step (min 5)
- Offer Portfolio Overview (table)
- Columns: Offer ID (OFFER-01…), Mapped Opportunity IDs, Name, Best-fit segment, Core promise, Primary KPI, Delivery timebox, Price range (and pricing model), Attach rate potential (H/M/L)
- 6–8 rows
- For each offer (repeat a standardized template; 200–260 words each):
- Offer name + one-liner
- Who it’s for / not for
- Problem statement (tie to pains)
- Scope (in-scope / out-of-scope bullets)
- Deliverables (5–9 bullets; concrete artifacts)
- Inputs needed from client (3–6 bullets)
- Timeline (week-by-week or day-by-day)
- Success metrics (1 primary, 3 secondary)
- Pricing logic (value metric, tiers if any, what increases price)
- Risks & how you de-risk (3–5 bullets)
- Upsells/cross-sells (2–4 bullets)
- Sales enablement assets to create (list 10; e.g., one-pager, ROI calculator)
- Assumptions (min 8)
- Challenges / Risks (min 5)
- Questions to ask sales/cs (min 8)
- Handoff to Next Agent (required format)
Dependency: Must complete after file 02 and before Agent 4 begins.
4) Agent 4: Delivery Systems Designer
Task: Define SOPs, tooling, templates, and AI automations to deliver each offer efficiently and consistently. Focus on repeatability, quality control, and reducing senior strategist time.
Deliverable file:
- outputs/agent_teams_demo/value_expansion_lab/04_delivery_systems_designer.md (1,800–2,300 words)
Required structure:
- How to use this document
- What I’m challenging from prior step (min 5)
- Delivery Stack & Tooling (table)
- Columns: Function, Tool options, Recommended default, Why, Cost note, Owner
- Include: PM, comms, reporting, analytics, creative, experimentation, AI, documentation
- For each offer (6–8), provide “Delivery Blueprint” (200–260 words each):
- Delivery roles (RACI-style bullets)
- SOP steps (8–14 steps; numbered)
- Standard templates/artifacts (list 5–8; name them)
- AI automations (3–6; specify input → process → output)
- Quality checks (5; measurable)
- Time budget estimate (by role; hours)
- Reusability map (identify shared components across offers; bullet list)
- Implementation plan (30 days) to stand up delivery system (weekly milestones)
- Assumptions (min 8)
- Challenges / Risks (min 5)
- Questions to ask sales/cs (min 8)
- Handoff to Next Agent (required format)
Dependency: Must complete after file 03 and before Agent 5 begins.
5) Agent 5: Experiment & Measurement Lead
Task: Design pilots and dashboards, set baselines, and validate ROI so only winning offers scale. Define pilot structure, measurement frameworks, and decision rules. Include a minimal viable dashboard spec.
Deliverable file:
- outputs/agent_teams_demo/value_expansion_lab/05_experiment_measurement_lead.md (1,600–2,200 words)
Required structure:
- How to use this document
- What I’m challenging from prior step (min 5)
- Pilot design principles (8–10 bullets)
- Offer Pilot Plans (for each offer; 160–220 words each):
- Pilot goal and hypothesis
- Minimum sample/time window
- Baseline required (what to capture before)
- Primary KPI + guardrail metrics
- Instrumentation/data requirements
- Success thresholds (quantified) and stop-loss thresholds
- Risks to validity + mitigations
- Dashboard spec (table):
- Columns: Metric, Definition, Source, Refresh cadence, Segment cut, Owner, Notes
- Minimum 18 metrics spanning acquisition/conversion/retention/LTV and ops efficiency
- ROI model outline (define how to compute incremental profit; include at least 1 formula for each: SEO, PPC, CRO, retention)
- Decision gate: scale / iterate / kill (explicit rules; 10 bullets)
- Assumptions (min 8)
- Challenges / Risks (min 5)
- Questions to ask sales/cs (min 8)
- Handoff to Synthesis (required format)
Dependency: Must complete after file 04 and before Synthesis begins.
6) Synthesis/Review Step (Team Lead behavior; consolidate and stress-test)
Task: Produce an executive-ready package that consolidates the best offers, explains why they win, how to sell them, and how to roll them out in 60 days. Resolve conflicts across agents, call out key assumptions, and create a tight action plan.
Deliverable file:
- outputs/agent_teams_demo/value_expansion_lab/06_synthesis_rollout_plan.md (1,500–2,000 words)
Required structure:
- How to use this document
- Executive summary (150–220 words): what to do next and why
- The “Winning 3–5 Offers” (choose the best 3–5 from the portfolio)
- For each: one-liner, ICP, why it wins, key KPI, price anchor, delivery confidence (H/M/L)
- Sales motion (how to sell to existing clients vs net-new) (2 sections; each 200–260 words)
- 60-day rollout plan (week-by-week; include: build, enablement, pilot, review)
- Risks & mitigations (top 10)
- Alignment check: what must be true (list 10 assumptions as testable statements)
- What to measure weekly (list 12 metrics)
- Final “Open Questions to Resolve” (list 12)
- “If you only do 5 things” checklist (exactly 5 bullets)
DEPENDENCY: This synthesis must only begin after files 01–05 are complete.
OPERATING INSTRUCTIONS (important)
- Do not ask the user for more context; proceed using informed agency/eCommerce defaults.
- Where specifics are missing, make explicit assumptions and provide verification questions.
- Keep identifiers consistent across documents (PAIN-xx, OP-xx, OFFER-xx).
- Ensure each downstream file reuses and references the backlog and IDs from upstream files.
- After completing all files, do a final internal consistency check:
- Offer IDs match mapped opportunities
- Opportunities link to pains
- Metrics align with offer promises
- Delivery time budgets are plausible
- Pilot thresholds are quantifiable
Create your own AI agent team at Build Agents Store. Describe your business problem and get specialized agent teams with ready-to-use prompts for Claude Code.