Video summary

How to Service Your First AI Automation Agency Client ($3000 EACH)

Main summary

Key takeaways

Business

Overview (case-study driven “first client” delivery)

  • Presenter Liam Otley (Morningside AI) walks through the end-to-end process to deliver a client’s first AI automation agency project: planning → building → reviewing/testing → deployment → handover using a real chatbot case study.

  • The chatbot is an outward-facing customer assistant designed to drive:

    • lead nurturing
    • lead capture
    • conversion to booking/trial
    • customer support via knowledge base + AI fallback

Sales / offer setup & concrete commercial details

  • Client pricing (case study discount):
    • Typical project: $2,000–$3,000 (“two to three thousand”)
    • Case study rate paid: $1,000
    • Proof: Stripe dashboard screenshot shows 3,000 Dirhams ≈ $1,000
  • Positioning: Liam frames the chatbot as a consulting deliverable to:
    • reduce expenses and increase ROI and/or
    • increase revenue through automation

Actionable recommendation

  • For a first project, target low-hanging fruit and “stick to” chatbot-style automation to build a repeatable playbook.

Target client problem: what the chatbot must do

The client wanted an outward-facing chatbot covering four business functions:

  1. Lead nurturing Help website visitors find the right products/services via guided conversation.

  2. Lead capture Collect information (email/phone mentioned).

  3. Conversion event routing Route users to a trial page, book a call, or an external landing page.

  4. Customer support Answer questions using a custom knowledge base.

Delivery framework / “playbook” (end-to-end phases)

The video presents a step-by-step delivery process with explicit stages.

1) Planning (requirements + conversational design)

  • Identify needs & value fit for a chatbot in the client’s niche.
  • Request requirements from the client:
    • knowledge base needs (Q&A content)
    • conversation objectives (buttons/branches)
    • what information to collect
    • which conversion event to route to
  • Conversation flow mapping in Figma
    • Liam builds a conversation diagram including:
      • initial questions
      • user intent prompts
      • where knowledge base queries occur
      • what data is captured

2) Building (system design + tech stack)

  • Tech stack:
    • Botpress for chatbot logic/orchestration
    • StackAI via APIs to improve knowledge-base querying / LLM answering
  • Botpress architecture approach (tactical):
    • Use nodes (messages, info capture, routing transitions)
    • Reduce visual complexity by moving logic into “big JavaScript if/else blocks”
  • Knowledge answering approach:
    • First try: query the client Q&A knowledge base
    • If it fails: use StackAI / GPT-4+ style model
    • Include fallback + loop so users can:
      • receive an answer
      • then re-offer the conversion event (book trial/call)

3) Reviewing / Testing (batch iteration to avoid chaos)

  • Deploy a testing link for the client (Botpress supports URL-based testing).
  • The client stress-tests and returns feedback.
  • Process recommendation: avoid endless ping-pong:
    • use batch iterations (one round of changes, then another) instead of constant back-and-forth.

4) Deployment & handover (client ownership + integration)

  • Client creates their Botpress account and adds Liam as a collaborator.
  • Transfer the bot into the client’s account so:
    • the client is billed (not the agency/dev account)
    • ongoing tweaks are owned by the client afterward
  • Provide an integration script:
    • Botpress Webchat is pre-configured
    • client copies the script URL
    • client installs it on their website to render the chatbot UI

Productized deliverable / scope example (what’s “included”)

The case study chatbot includes the four core functions above, with a near-term upgrade mentioned:

  • Current scope: outward-facing chatbot with:
    • guided lead nurturing
    • knowledge base support with AI fallback
    • conversion routing prompts
  • V2 roadmap: add lead capture (email/phone collection + enriched leads)
    • Result: leads can feed email marketing and SMS marketing

Metrics / KPIs mentioned (and implied)

  • Explicit metrics/targets:
    • Revenue per project:
      • discounted $1,000 for the case study
      • suggested market pricing: $2k–$3k per project
    • “5x revenue per customer” claim (used as the promise for a bonus module)
  • Implied KPIs by functionality (not numerically quantified):
    • lead generation metrics:
      • number of captured leads (email/phone)
      • conversion events (book trial / calls)
    • support quality/deflection:
      • fewer unanswered questions via knowledge base + fallback AI
    • marketing pipeline efficiency:
      • enriched leads for email/SMS campaigns

Bonus module: “How to 5x revenue per customer” (upsell playbook)

Liam’s upsell strategy uses the chatbot as a wedge into deeper automation revenue.

Upsell strategy (stepwise logic)

  1. Trojan Horse foundation: chatbots build trust first.
  2. Upsell based on what can be analyzed and integrated next:
    1. Add integrations to route chatbot-generated leads into systems (lead routing, automations, etc.)
    2. Set up analytics/analysis for knowledge-base questions: - categorize common queries - identify content gaps (what customers can’t find) - drive content improvements and automated responses
    3. Internal automation upsell: - repurpose unused content assets to help the client’s team generate more content
    4. Full business audits + workshops: - “head-to-toe” AI augmentation plan - deliver multiple automations once there are 10–20 deliverables
    5. Convert to monthly recurring revenue: - package into a monthly retainer - target described: $3,000–$8,000 per month per client

Frameworks / playbooks explicitly or implicitly used

  • Delivery lifecycle playbook: Planning → Building → Reviewing/Testing → Deploying → Handover

  • Consultative requirement mapping: value opportunity → define chatbot functions → capture requirements → design conversation flow

  • Iteration process: batch iterations to prevent review bottlenecks

  • Revenue expansion framework: Wedge product (chatbot) → integration upsells → analytics → internal automation → audit/workshop → retainer

Concrete actionable recommendations (pulled from the video)

  • Choose a first client that benefits from a chatbot (especially outward-facing).
  • Build a conversation flow diagram in Figma and validate early with the client.
  • Implement knowledge answering using a two-stage strategy:
    • knowledge base first
    • LLM fallback second
    • with looping Q&A
  • Deploy with a test link, then request feedback via one batch of iterations at a time.
  • Do a proper handover by transferring the bot into the client’s account (collaborator model) so you’re not still the billing owner.
  • For upsells, show results and expand scope through:
    • integrations
    • question analytics
    • full audits/workshops leading to monthly retainers

Presenters / sources

  • Presenter: Liam Otley
  • Company referenced: Morningside AI (also “Morningside Ventures” shown in Stripe screenshot)

Original video