Notes from the sales stack.
Two series written in order. Twenty-five years of outbound email, era by era. Then one deal, from the recording to the proposal. Every claim sourced, every diagram drawn from the system itself.
A new era of outbound email
Twenty-five years of cold email, era by era: every tool that made sending easier handed the inbox a new pattern to defend against. Then the operating model that works now.
- 012001 to 2010: email becomes infrastructure, and the filters learn rules
- 022011 to 2017: sequencers turn follow-up into a system, Gmail starts grading the sender
- 032018 to 2021: enrichment plugs into the sequencer, identity becomes the filter's strongest signal
- 042021 to 2023: mailbox rotation and AI drafting make campaigns cheap, and the filters learn faster
- 05February 2024: authentication stops being best practice and becomes the gate
- 062026 to 2027: research is nearly free, and visibility inside the inbox is the next gate
- 07Sending capacity is no longer the advantage: the hatchet and the scalpel
- 08Manage outreach as a portfolio of campaign cells, not a send queue
- 09A signal earns the right to become a message
- 10Learn which message earns the conversation: openings, value lines and asks, scored separately
- 11Time the conversation: send window, day of month, follow-up rhythm, season
- 12Trace a reply-rate drop from the outside in: the seven-layer runbook
- 13Automate the known. Escalate the consequential.
- 14Turn qualified replies into the next right action
After the meeting
The sales call already holds the answers. This series follows one deal from the recording to the proposal: every field with the sentence behind it, every promise tracked, and the rep still making the call.
- 01The rep still makes the call: the whole system on one page
- 02No call wasted, no thread wasted, no note wasted
- 03It does not move your numbers. It argues with you about them.
- 04The forecast that reads the calls
- 05Deals do not die on price. They die on a promise nobody wrote down.
- 06Every meeting ends with a recap. Every thread gets an answer or an owner.
- 07Press on anything and hear where it came from
- 08The proposal writes itself, and every claim shows its source
- 09Use last quarter's demos to fix next quarter's
- 10Hand the rep the answer while the call is still open
- 11The harness: the same webhook twice does nothing twice
- 12What to measure, and what the published numbers already say
- 13Automate the transcription. Keep the judgment.
Latest
2001 to 2010: email becomes infrastructure, and the filters learn rules
Mailchimp and SendGrid made sending at scale a service. Blocklists, scoring rules, Bayesian filters and the CAN-SPAM Act answered. The pattern that still runs outbound in 2026 started here.
The rep still makes the call: the whole system on one page
Every call, thread and note becomes a drafted CRM record with the sentence behind each field. The rep approves it in about two minutes, keeps stage, dates and amount, and the system only argues with a quote.
2011 to 2017: sequencers turn follow-up into a system, Gmail starts grading the sender
Sequencers made follow-up a system from 2011 to 2017, and Gmail began grading senders by their history. Since then, a domain's past shapes where its mail lands.
No call wasted, no thread wasted, no note wasted
Run every call recording, email thread and rep note through one extraction. Each fact is typed and keeps its sentence, so your CRM becomes queryable.
2018 to 2021: enrichment plugs into the sequencer, identity becomes the filter's strongest signal
From 2018 to 2021 data, enrichment and sequencing became one workflow. Mailbox providers answered by weighing domain, IP, authentication and feedback together.
It does not move your numbers. It argues with you about them.
Stage, close date and amount are the rep's commitments. The system argues with a quote instead of overwriting them, and the rep accepts or dismisses.
2021 to 2023: mailbox rotation and AI drafting make campaigns cheap, and the filters learn faster
Warmup, rotation and AI drafting made cold campaigns nearly free from 2021 to 2023. Filters learned faster, and polished copy stopped meaning much to readers.
The forecast that reads the calls
Adjust the weighted pipeline only where the CRM contradicts what a buyer said on a recorded call, and show the sentence. The gap is your pipeline review agenda.
February 2024: authentication stops being best practice and becomes the gate
Since February 2024 Gmail and Yahoo require bulk senders to use SPF, DKIM, DMARC and one-click unsubscribe, with spam under 0.3%. Outlook.com followed in 2025.
Deals do not die on price. They die on a promise nobody wrote down.
Stalled deals usually wait on a promise, not on price. Track every commitment either side makes as it is spoken, then draft what is owed for a person to send.
2026 to 2027: research is nearly free, and visibility inside the inbox is the next gate
In 2026 agents made research and drafting nearly free. Gmail's AI Inbox ranks mail by relationship and message signals, so delivered no longer means seen.
Every meeting ends with a recap. Every thread gets an answer or an owner.
Draft the recap minutes after the call, every claim traced. Answer questions from approved material or route them, and revive quiet deals on their own rhythm.
Sending capacity is no longer the advantage: the hatchet and the scalpel
Mass cold email still sends, but no longer wins: every team buys the same data, sequencer and drafting model. The advantage is deciding who, why now and what.
Press on anything and hear where it came from
Accurate enough to draft, never to trust blind. Every field should show its sentence, so reps check it in seconds and their corrections become training data.
Manage outreach as a portfolio of campaign cells, not a send queue
Split outbound into campaign cells by company size, industry and persona, count each eligible pool honestly, then decide where to build, fix or stop sending.
The proposal writes itself, and every claim shows its source
Yes, if every claim shows its source. The draft opens on the buyer's words, prices only the scope they stated, keeps cut lines out, and waits for the rep.
A signal earns the right to become a message
Score each signal for fit, freshness, source quality and relevance before writing. Then wait, reach out now or disqualify, and pass the research to the rep.
Use last quarter's demos to fix next quarter's
Average a quarter of recorded demos of one deck and the same moments go quiet. Score the deck, not the rep, and turn the drops into an edit list with owners.
Learn which message earns the conversation: openings, value lines and asks, scored separately
Treat each cold email as an opening, a value line and an ask. Tag every send, score each part in comparable cells, and promote only past a preset evidence bar.
Hand the rep the answer while the call is still open
AI can help during a live sales call in one narrow way: one silent line from a rep who won, shown only to the rep. Five guardrails, and why it ships last.
Time the conversation: send window, day of month, follow-up rhythm, season
Generic studies point to midweek mornings. Learn the send window, day of month, follow-up gaps and season for each campaign cell from your own replies instead.
The harness: the same webhook twice does nothing twice
A reliable sales agent is a compiled graph: typed steps, fixed tools, one approval interrupt, idempotent writes, checkpoints, and a CRM login with hard limits.
Trace a reply-rate drop from the outside in: the seven-layer runbook
When cold email replies fall, test the delivery path before the message: sender health, lead source, placement, then timing, construction, strategy and market.
What to measure, and what the published numbers already say
Track accept and dismiss rates per field, fields that survive to close unedited, and adjusted forecast vs actual. Plus four published sales figures, caveated.
Automate the known. Escalate the consequential.
Known events with approved actions run on their own: renewals, auth drift, bounces, opt-outs, scheduling. Ambiguous ones arrive as a brief a person decides.
Automate the transcription. Keep the judgment.
Automate known events: extraction, records, the promise ledger, recaps, quotes. Keep consequences with a person: stage, dates, amount, send. Where to start.
Turn qualified replies into the next right action
Classify each reply by intent, confidence and policy. In scope, an agent answers from approved material and books the rep. Everything else goes to a person.
Earlier writing
Posts from before the two series, on agents, automation and implementation.
- Cold Outreach That Learns: Inside an Agent That Researches Before It Writes
- The Signal Timeline: How Meeting Transcripts and Email Become Deal Memory
- Your Website Traffic Is Full of Buyers You Can't See
- SEO Got You Into Google. AEO Gets You Into the Answer.
- Pitch Decks That Answer Objections Before They Come Up
- Churn Warnings Live in Your Support Tickets
- Why 70% of AI projects fail (and how to be in the 30%)
- AI implementation strategy that actually works
- Lead generation with AI: quality over quantity
- Marketing automation ROI: beyond the hype
- Campaign funnels that actually show ROI
- AI workflow automation: 10 tasks to automate first
- Building AI agents for small business
- AI agents vs traditional software: cost of ownership
- AI agent security and compliance
- Scaling AI agents: proof of concept to production
- How voice & email agents cut support workload
- From PDFs to data: automation that saves hours
- AI agents transform customer service
- Business process automation: implementation guide
- AI workflows for remote teams
- Industry-specific AI workflows
- Voice AI agents vs chatbots
- AI agents ROI: measuring business impact
Fifteen minutes to see if it fits.
A free call with Jared about your team, your tools and the first thing you would fix. Or walk through a sample sales team first.