🏋 Coaching

Seneca Adoption

8 sources Updated Wed May 20 2026 19:00:00 GMT-0500 (Central Daylight Time) Internal
senecaaicoachingnegotiationbotolympusknowledge-basepushbackself-improvement

Summary

Seneca is the Artemis team's AI negotiation coaching bot. Fred repeatedly pushes agents to adopt it for real-time deal support. Trained on Zillow advisor content, negotiation frameworks, Fair Housing compliance, and 960K+ MARIS MLS records synced 5x daily.

Fred's Position

"If you haven't used Seneca you've got a nuke in your pocket and you're fighting with a slingshot."

Adoption is not optional in Fred's view. Seneca is a force multiplier that most agents underuse or ignore entirely.

Live Proof of Concept

Fred used Seneca at a live auction to produce a tax appeal valuation for a stranger on the spot. He pulled the data, generated the analysis, and handed it to the prospect — who became a lead. This is Seneca's value proposition in action: instant credibility through instant competence.

Training Data

  • Zillow advisor content and scripts
  • Negotiation frameworks (objection handling, anchoring, framing)
  • Fair Housing compliance rules
  • 960K+ MARIS MLS records, synced 5x daily

Use Cases

  • Objection handling scripts — real-time responses to buyer/seller pushback
  • Offer strategy — competitive positioning in multiple-offer situations
  • Market data lookups — instant comps and trend data mid-conversation
  • Commission conversations — language for navigating post-NAR-settlement fee discussions
  • Compliance checks — Fair Housing language review before sending

Call-to-Coaching System

Fred records coaching calls and converts them into 5-minute micro-resources distributed through Seneca. This creates a feedback loop: live coaching sessions become searchable, on-demand training content. Agents who miss a call can access the same insights asynchronously.

Adoption Gap

Despite its capabilities, Seneca remains underutilized across the team. Fred identifies this as a coaching priority — the tool exists, the data is live, and the agents who use it consistently outperform those who do not.

Corpus Expansion (2026-04-28) — 700+ Scripts and Playbooks

Fred ran overnight sessions extracting BAMx, Listing Leads, and Curaytor knowledge bases into Seneca's corpus. Net result: 700+ scripts, frameworks, and playbooks now live under the evergreen Voss/Jones/Fred negotiation layer.

What was added:

  • Byron Lazine 2026 Script Black Book
  • 53 Byron Lazine weekly Tuesday role-play scripts
  • 138 Listing Leads scripts
  • 6 LL multi-touch playbooks (Expired, FSBO, Silver Tsunami, Open House, etc.)
  • Tom Toole objection-handling decision tree

Layer architecture (now two layers):

  1. Evergreen master layer — Voss + Jones + Fred. Tactical communication and negotiation. Always primary.
  2. Tactical execution layer (new) — 700+ scripts with WHY (when to use), AUDIENCE (exact list segment), HOW TO EXECUTE (step-by-step), the actual copy, plus a source footer (Byron / Toole / Mackin / etc.) and decay window.

Seasonal scripts surface seasonally. Decay-aged scripts are flagged for refresh. This solves the "stale script" problem that plagues most paid-content libraries.

New use cases unlocked:

  • "Buyer says they want their attorney friend to write the offer — script please"
  • "I have an expired listing — give me the 45-day playbook"
  • "Got an FSBO appointment tomorrow at 6pm — what do I bring?"
  • "Senior couple wants to downsize — playbook?"
  • "Buyer has cold feet pre-inspection — save the deal"
  • "What should I send sellers this week?" (picks something seasonal + current)
  • "Rate locked higher than my buyer expected — they want out"

Hard boundary: Don't ask Seneca to write a full email blast. That's the job of Mercury (in build). Seneca coaches the script + delivers text/phone/in-person language.

Strategic effect: This corpus expansion is the precondition for the listing-leads-sunset-plan — Seneca has officially surpassed LL's strategy library for the agents who need scripts, so LL's Phone/Text + Blueprints tier can be cancelled when convenient.

Olympus Integration (2026-04-17)

Seneca now auto-reads the olympus-operations-wiki. Every SOP, sales tactic, coaching theme, and market intel note filed into Olympus becomes immediately available to Seneca without manual re-training. This closes the old gap where Seneca could be weeks behind on team-level context changes.

Practical effect: an agent asking Seneca about a pricing strategy or a ZHL objection gets an answer grounded in the current team playbook — not a generic negotiation framework.

Fred's framing: "This is 10x more capability than agents are going to have in the field." Combining real-time negotiation coaching with live team operational knowledge is what makes Seneca different from a stock AI chatbot.

Seneca v2 — Self-Improvement Loop + 4 New Rules (2026-05-11)

Fred shipped a major Seneca update on 2026-05-11 documenting both how the bot evolves and four new rules now live in the prompt.

The improvement loop (how Seneca actually levels up)

Every agent conversation gets vacuumed into the corpus. Each week, Seneca runs a self-correction pass over the week's chats and flags every moment where the answer was off, the agent pushed back, or the response didn't land. Most of those get auto-corrected. The bigger ones get DM'd to Fred as a "you should probably look at this" list — Fred writes the rule, ships it back, bot gets sharper.

Fred's framing of the maturity arc: "Right now Seneca is a really smart high school kid. In a couple months, PhD. In a year? Mr. Spock would look dumb next to Seneca."

The agent's job is pushback. When Seneca tells you to do something that feels off — "actually, no" / "that's wrong" / "let me clarify" — that's the signal that closes the loop. The agents pushing back the hardest get the best version of Seneca. Agents who treat it as a generator and never push back get the dumbest version.

Rule 1 — Never quote a prospect's Zillow activity back to them

Telling a lead "I saw you saved that $1.6M home" reads as surveillance and burns trust. Three-move replacement (credential the source → generalize the specifics → invite restatement). Tactical script lives in zillow-pro-sphere-lock.

Rule 2 — One-word criteria are capability-demo triggers, not appointment cues

When a lead says "Schools!" / "Big yard!" / "Investment property!" — do NOT pitch a 15-min call. Run the six-move capability-demo framework. The appointment ask comes AFTER engagement, not instead. Tactical script lives in vip-buyer-program.

Rule 3 — Cancelled showings + short replies = supply-side problem

Cold leads aren't a demand-side issue — they don't yet see what makes you different. Two scripts: the "moving parts" walk-through for cancelled showings, and the bench-vs-field metaphor for short replies. Tactical scripts live in vip-buyer-program.

Rule 4 — Never send a seller a message with [X] for the answer they asked for

If a seller asks "what do I net at $375K?" — actual numbers in the answer slot, never placeholders or "I'll send the breakdown in 5 minutes." Reference delphi-closing-calculator by name and sell the calibration (900+ real ALTA statements vs. generic national templates). Tactical doctrine lives in delphi-closing-calculator.

Why these four together

All four rules attack the same pattern: agents shortcutting to credentials-and-appointment instead of demonstrating live capability. Seneca v2 enforces capability-first responses across the four highest-friction touchpoints (saved-search outreach, criteria probes, recovery from cold leads, net-proceeds questions).

Success Stories — May 2026 (proof Seneca is working)

Fred's 2026-05-11 callout for success stories surfaced three concrete wins inside 20 minutes:

  • Jackie Poponi — Used Seneca at her son's baseball game for a challenging price conversation. Fed seller replies, got tailored responses with O'Fallon-specific market data (homes advertising new roofs, 30-day price drops, DOM by criteria) at her fingertips. Texted Fred mid-game.
  • Chad Klein — Used Seneca to pull his Aviston 4th St listing ($550K) with a performance-based commission structure (got the exact email script). Used to negotiate a $710K counter on his J Rock listing. Daily use across his full sales cycle — solar company seller correspondence, showings, agent conversations. Taught Jack Daltagiacomo to use it the same way. Quote: "Guys, ask it questions about helping you and provide as much details as possible about the situation or scenario or goal you are looking to accomplish. It's freaking spectacular!"
  • Alicia Howard — Walking into her son's athletic banquet, client wanted her uncle's opinion of a house (after mom and dad already weighed in). First Seneca response coddled the buyer. Alicia pushed back — told it to position her as the knowledge broker. Second response: spot-on with zip-code-specific marketing details. Outcome: wrote the offer at the banquet, UC the next day instead of booking another showing.

The Alicia story is the textbook example of the improvement loop — the first answer was wrong, she argued, the second answer landed, the deal closed. That whole sequence becomes corpus next week.

Buyer-Agent "Pain Grade" Reports — Pre-Rollout (2026-05-20)

First concrete preview of the team transaction corpus being used for per-opposing-agent intelligence, surfaced by Fred 2026-05-20 in #direct-from-fred. Currently DM-only, in advance of the Ares / Elysium / Pantheon rollout.

What it produces: a profile of a named opposing buyer's agent based on their historical deal pattern across the team's closed transactions. Two dimensions Fred named:

  • Concession size pattern — average and tail of seller concessions extracted across their closed buyer-side deals
  • Termination rate — how often their deals fall out before closing

Example contrast Fred shared (anonymized): one agent extracts deeper/bigger concessions less frequently; another extracts smaller concessions more frequently. Same outcome category, different negotiation cadence — and that cadence shapes how Jackie should frame her own offer.

Tactical use cases (per Fred's 2026-05-20 post):

  • Live competitive-offer framing. Before submitting, know how the opposing buyer's agent typically negotiates concessions, and frame the offer either to undercut their playbook or to look more professional to the listing agent.
  • Pre-listing-appt prep. When the seller will interview other agents, know the competitor's deal-pattern profile before walking in.

Distribution: ad-hoc, DM Fred for the report. Will become self-serve once Ares (listing appt prep) and Pantheon (agent dashboard) ship. Per CLAUDE.md, Ares is private kitchen-table tool — competitor-by-name and steering/laundering material allowed; Elysium is the public-safe sibling. This pain-grade report is the Ares-side capability.

Related