GTM Engineering · All-hands · September 2026

AI in GTM Engineering.

The ten-minute section with Naveen: the problem, the engine, six and a half minutes of it working, and the numbers. Present opens the slides.

0contacts sourced
0Ksegmented, zero credits
0accounts scored
0agents on a schedule

Run of show

Ten to fifteen minutes, Jason Dowden moderating.

MinSegmentWhoWhat happens
0:00IntroDallasName, role, department. One line on what GTM engineering is.
1:00The problemNaveen + DallasThe mandate, and why outbound doesn't scale by hand.
3:00Demo, recordedDallas narratesSix and a half minutes of the engine working: Claude Code, Clay, then lemlist.
9:30ResultsDallasThe numbers, and the hours of manual work they replace.
11:30QuestionsJason moderatesAnswers prepped for the likely ten.

Demo, six beats

One recording, 6:30. Claude Code is the brain, Clay carries the data and the AI research, lemlist carries the send.

TimeBeatOn screen
0:00The brainClaude Code: agents that ran overnight, and a strike plan built live from one ask.
1:10The databaseClay: every contact and account we know, synced from Salesforce. A live segment built in seconds, at zero cost.
2:10AI research + draftsRows enriching live: person found, work email verified, research written, a personalized five-touch sequence drafted per contact.
3:40AutomationA published workflow runs end to end: registrant in, enriched, classified, routed. No one at the desk.
4:30The send layerA loaded lemlist campaign: email, LinkedIn, and call steps per lead, every variable rendered, AI voice note on the sequence.
5:40The loop, beyond outboundReplies route back to the owner, every motion measured on cost per meeting. Closer: the twelve back-office account maps built with the AM team, live as a page.

Open items

  • RecordDemo recording, script in the session folder, one dry pass first.
  • ShapeNaveen's pass on the split and the problem framing.
  • FreezeSlides lock after that pass.
All-hands · AI at work

Dallas Andrews
GTM Engineer

Product Strategy, with Naveen Thilagan. One job: build qualified pipeline, engineered as software.

SourceResearchWriteSendMeasure
The problem

Outbound is manual work.

Find the right person. Learn enough to say something worth reading. Keep customers out of cold campaigns. Every contact, every touch.

0
contacts worth reaching
0
touches per contact
0
hours of manual work.
by hand, it never happens
The answer

One engine.

Signals in Salesforceknown people, customers, owners Market signalsratings, cost mandates, leaders Account mapsbuilt with the AMs Intentsite visits, replies, ad clicks AI engine scores, gates, researches writes every first draft Claude agents, on Clay cleared customers Sequencesemail, LinkedIn, call · on lemlist AM clearancecustomers never go cold cleared names LinkedIn ads, same listads run before the first email Meetingsat the decision maker Salesforcecounted like every channel Read and adjust, every two weeks next wave
Where the AI works

AI at every step.

  1. 1Reads the signalsClaude Code agents, nightlymarket scans, account triggers, briefs by 8am
  2. 2Scores and gatesthe engineaccounts ranked, customers excluded, waves of 30 to 50
  3. 3Researches every contactClayperson found, email verified, role confirmed
  4. 4Writes every first draftClaude, in the rep's voicefive personalized touches per contact
  5. 5Runs the sendlemlistemail, LinkedIn, call steps, AI voice notes
  6. 6Judgment checkpointstuned per motionwidening as each motion earns trust
Demo · 6:30

Watch it work.

  1. 1The brain0:00 · Claude Codeovernight agents, a strike plan from one ask
  2. 2The database1:10 · Clay140,000 people, a segment in seconds
  3. 3AI research + drafts2:10 · Clayrows enriching live, sequences writing themselves
  4. 4Automation3:40 · Claya workflow runs with no one at the desk
  5. 5The send layer4:30 · lemlistloaded campaigns, every variable rendered
  6. 6The loop, beyond outbound5:40replies measured, and the AM team's account maps
Results

Eight weeks in.

0
contacts sourced
outside Salesforce
0
contacts segmented
at zero credits
0
accounts scored,
gated, or mapped
0
hours of sourcing and research
given back to sellers
  1. ·1,016 webinar registrantsone afternoonenriched, classified, routed to owners
  2. ·7 sequences loadedNA and UKpersonalized, gated, ready to send
  3. ·12 account mapswith the AM teamback-office buying committees, delivered as a live page
  4. ·19 agents on a scheduleovernightscans, audits, and briefs before the day starts
Questions

What could this run for you?

If your team has a repetitive workflow, it can run this way. Find me on Slack.

GTM Engineering · All-hands
1 / 7