What Morgan the AI SDR does in 24 hours
Morgan is Ektie's AI SDR. He runs on a 60-minute heartbeat cycle — 24 times a day, every day. Here's exactly what happens inside one full day: how many prospects he finds, what he researches, what he writes, and where the work goes.
Morgan is Ektie's AI SDR. He runs on a 60-minute heartbeat cycle — 24 cycles a day, 7 days a week. Most founders are surprised by what that actually looks like in practice. Here's a real walkthrough of one full day.
What does Morgan do at the start of every cycle?
The scheduler fires and Morgan assembles his Brain before making any decision. That means: the current ICP definition from the Team Brain, his Personal Brain memory blocks (coaching notes, patterns from prior cycles, account context for active prospects), and any coaching Orion (Sales Director) wrote since the last cycle. One read operation, full context loaded, then work begins.
Prospecting: 5–10 new accounts per cycle
Morgan queries Apollo, Clutch, YC's database, and general web research using the current ICP parameters. He scores each result for fit. Companies that score above the ICP threshold go into his research queue. In a typical cycle, he finds and scores 15–25 companies, of which 5–10 make the research cut.
Account research: 3–5 minutes per account
For each qualified company, Morgan reads the website, checks recent LinkedIn activity, looks at job postings for intent signals, and notes any recent news or announcements. He's looking for two things: what's the most relevant pain for this specific company right now, and what's the one detail he can put in the first line of outreach that shows he actually looked.
How does Morgan write outreach that isn't templated?
Morgan drafts the first-touch email for each researched account. Not a template with variables swapped — a message built around what he just read about the company and the specific pain most relevant to their situation. Coaching notes from his Brain shape every draft. If Orion wrote "lead with compliance angle for healthcare CTOs", that instruction applies to every relevant email Morgan sends until the note is updated.
What happens to prospects already in sequences?
While Morgan works the new-prospect queue, the sequence runtime is executing scheduled steps for 50–300 prospects from prior cycles. Follow-up emails fire on schedule. LinkedIn connection requests go out. Break-up emails send at the end of sequences. None of this requires Morgan's active attention — it runs in parallel.
Reply handling: warm replies route to Astra
When a prospect replies, the reply is classified. Positive replies — "yes, interested", "let's chat", "tell me more" — trigger a handoff to Astra (the AI AE) with full context: every prior touch, the research Morgan ran, the company data, the sequence history. Negative replies remove the prospect and log the reason. Neutral replies get a follow-up question from Morgan.
How does Morgan get better over time?
At the end of every cycle, Morgan writes memory blocks: which segments replied, which subject line formats worked, what patterns he observed. High-importance observations get scored accordingly. On the next cycle, these blocks load alongside the coaching notes and ICP. Each cycle starts with more signal than the last.
What a full day looks like in numbers
24 cycles. 120–240 new accounts researched and scored. Sequences executing continuously for up to 500 active prospects. Memory written and reloaded every hour. Orion's coaching applied on the next cycle after it's written. It runs whether or not anyone logs in.