[ SERVICES 03 / 08 ] M E S Six playbooks
000%Scroll
00:00:00 UTCDistributed
[ 03 — Services ] Six playbooks

Agents I build.

Six of the agents from my catalog, in depth. Each is a playbook designed to run again and again — the agent looks different at every business, but the architecture, the eval harness, and the operating model stay the same. See all ten on the Agents page.

Playbooks06
To production21 days
Your cloud100%
Service catalog6 playbooks
01Speed-to-lead AIwk 1–3
02WhatsApp receptionistwk 1–3
03Document intelligencewk 1–3
04Content repurposingwk 1–3
05Lead qualifierwk 1–3
06AI cold outreachwk 1–3
same rails · different agent
Continue
Index

Jump to a playbook.

01 · Speed-to-lead 02 · WhatsApp receptionist 03 · Document intelligence 04 · Content repurposing 05 · Lead qualifier 06 · Cold outreach
Anatomy

How an agent actually works.

Every agent I ship looks roughly like this — a signal, a brain, the right tools, an action, and an audit log. The pieces change. The shape doesn't.

Signal
EMAIL · TICKET · WEBHOOK
Reasoning
CLAUDE · GPT
Tools
CRM · ERP · EHR
Action
SEND · POST · SUBMIT
↓ BEFORE

A person, doing it slowly.

→ Read ticket / lead / invoice
→ Switch to CRM / ERP / EHR
→ Look up context (5 tabs)
→ Write reply / fill form
→ Double-check, send, log
≈ 12 — 45 minutes per item
↑ AFTER

An agent, doing it constantly.

→ Signal received
→ Context retrieved (parallel)
→ Reasoning + tool call
→ Action executed
→ Logged + observability
≈ 3 — 30 seconds per item
SVC · 01 / SPEED-TO-LEAD

Reply to every lead in seconds.

The agent watches your forms, WhatsApp, and missed calls. The moment a lead lands, it replies in your voice, answers the first question, books a slot, and logs everything to the CRM — before a competitor opens the email. Leads contacted in the first minute convert far more than ones left for an hour.

  • Instant detection from forms, WhatsApp, and call misses
  • Personalised first reply in your brand voice
  • Auto follow-up cadence if there's no response
  • Calendar booking via Cal.com / native calendars
  • Two-way sync with HubSpot, Salesforce, Zoho
[ LIVE FEED ]
● new lead · web form
→ reply sent · WhatsApp
✓ replied · 3 seconds
✓ booked · Tue 2pm
SVC · 02 / CX
[ WHATSAPP · 24/7 ]
"Are you open this Sunday?"
"Yes — 10am to 2pm. Want me to book you a slot?"
"11am please"
"Booked ✓ — see you Sun 11am"
✓ HANDLED · no staff · 18s

A receptionist that never sleeps.

A 24/7 agent on the app your customers already use. It answers FAQs, qualifies enquiries, books slots, and escalates only the genuinely complex questions to you — capturing the after-hours messages that today go to a competitor by morning.

  • FAQ, intent, and qualification handling
  • Appointment booking inside the chat
  • Draft-first or full-auto reply modes
  • Escalation rules + human-in-the-loop checkpoints
  • Native on WhatsApp, Instagram DM, and web chat
SVC · 03 / DOCUMENT INTELLIGENCE

Make the back office disappear.

An OCR + LLM pipeline that ingests any document — invoices, contracts, packing slips, expense receipts — extracts every line item, validates against your master records, flags anomalies, and posts to your accounting system. Built to process documents at scale across PDF, image, email, and EDI.

  • Vendor-agnostic invoice extraction (PDF, image, email, EDI)
  • PO and GL coding with confidence scoring
  • Three-way match and anomaly detection
  • NetSuite, QuickBooks, Xero, SAP B1 integrations
  • Human review queue for low-confidence items
[ INVOICE EXTRACT ]
VendorACME CO
PO match✓ PO-4451
Total₹ 18,420.00
Confidence98.4%
Posted→ NetSuite
SVC · 04 / CONTENT
[ 1 VIDEO → 4 POSTS ]
▶ SOURCE · 18-min YouTube
"How we cut onboarding in half"
✓ Blog post
✓ X thread
✓ LinkedIn
✓ IG caption

One recording, a week of content.

Drop in a YouTube video or podcast. The agent transcribes it, then Claude turns it into a blog post, an X thread, a LinkedIn post, and an Instagram caption — in your voice. You approve from a single queue; it learns from your edits.

  • Auto-transcription from video or audio
  • Blog + thread + post + caption, generated per recording
  • Brand-voice tuning from your existing content
  • One-click approval, then auto-schedule
  • 10× the output for the same effort
SVC · 05 / REAL ESTATE

Call only the buyers who are ready.

Every inbound property enquiry gets scored by budget, timeline, and intent. Hot leads are routed to an agent on WhatsApp within seconds; everyone else drops into an automatic nurture sequence — so no lead is dropped and no agent's time is wasted on tyre-kickers.

  • Scoring by budget, timeline, and buying intent
  • Instant WhatsApp routing of hot leads to agents
  • Automatic nurture drip for cold leads
  • Duplicate detection and enquiry de-duping
  • Two-way sync with your CRM / portal
[ LEAD · SCORED ]
SCORE
92/100
INTENT
High
BUDGET
✓ Fit
ROUTE
→ Agent
SVC · 06 / OUTBOUND
[ CAMPAIGN · Q3 ]
✦
SaaS founders · seed–series A
personalised by Claude
Sent1,240
Opened38%
Replied7% · 87 warm

A week of qualified conversations.

The agent scrapes and enriches leads that match your ICP, writes a hyper-personalised email for each with Claude, sends at scale, and follows up automatically — tracking every open and reply so only warm threads ever reach your inbox.

  • Lead scraping + enrichment from a dozen sources
  • Per-lead personalisation written by Claude
  • Sending via Smartlead / SMTP at scale
  • Open + reply tracking with auto follow-up
  • Warm replies handed straight to you
07 / Stack

Boring tech. Boring on purpose.

Anthropic
OpenAI
LangGraph
Temporal
Postgres
Vercel
AWS
GCP
Pinecone
Braintrust
Datadog
Sentry

Don't see your
workflow?

If you can describe a repeatable, rule-driven workflow in a 30-minute call, I can probably build an agent for it.

Describe your workflow