Winning by Design * R&D Roadmap

SPICED Pulse is not five projects.
It is one loop.

The dashboard measures the gap. The course and the agent close it. The rep changes behavior. The dashboard measures the improvement. Five workstreams from today's call are four positions on one flywheel, plus the manager coaching around it.

1
connected loop, not a portfolio
5
workstreams, each an arc of the loop
3
product surfaces on one data model
Mon
Nathan engineers against the map
The flywheel

Measure. Diagnose. Upskill. Act.

Every turn of the loop raises the number. The manager runs the loop on their existing cadence. The rep never leaves it.

Measure the dashboard Diagnose the gap Upskill course + agent Act rep behavior Manager on cadence
Measure
Real calls scored on rolling SPICED deal state. The gap is visible, grounded in verbatim customer lines.
Diagnose
The pattern surfaces: Impact, Critical Event, Decision Process. Not a per-call gotcha, a team habit.
Upskill
The per-rep agent hands the playbook section and a ready prompt. The course teaches the rep to run AI to get there. This arc is WS5.
Act
The rep asks the impact question on the next call. Behavior changes where the deal is won.
The five workstreams (WS), mapped to the loop

One owner, one next action, each.

Each is a workstream, labeled WS1 through WS5. Read them as arcs of the same flywheel, not a to-do list.

WS1 · Measure + Diagnose

Own the SPICED Pulse product definition

Spec, price, features, packaging. The forcing function that gives Nathan a target Monday. Sourced from the existing Momentus PRD, not reinvented.

Owner: Dan drafts · Shari names it customer-facing · Nathan engineers
WS2 · The rails under all of it

Map to bowtie, then data model, then agents

Formalize the live analysis schema as the SPICED data model on the bowtie, then define the agents and their entry, exit, and handoff. This is the M1 protocol thesis made concrete.

Owner: Dan + Dr. Dan Patterson architect · Nathan builds · feeds the open SPICED spec
WS3 · Act + Measure, on us first

Dogfood on WbD's own team

Point the pipeline at our own Gong calls. Use the dashboard in one real 1:1. Make it useful to us before we sell it. This is the proof point and the case study.

Owner: Dan runs it · Shari operationalizes the cadence
WS4 · Upskill + Act

The dashboard-to-action per-rep agent

The dashboard diagnoses; the agent acts. It reads the rep's gap, hands the playbook section and a ready prompt, follows up on cadence, and is how we actually move the lagging number.

Owner: Dan scopes · ties to Kai / 1mind
WS5 · The arc that closes the loop

GTM AI course: teach the rep to upskill with AI, toward the outcome

This is the piece that connects everything. The dashboard names the gap; the course teaches the rep how to use AI, including which frontier model for which task, to actually close it. Same outcome the dashboard measures, reached by the rep's own hands. WS5 is not a separate content project, it is the Upskill arc of the flywheel pointed straight at the number.

Owner: Dan + Patrick McGrann · Shari for naming (frontier model, not AI of choice)
Product definition · WS1

Three surfaces, one data model.

Surface 1

SPICED coaching card

Per rep, four parts: what the customer said, what the rep said, why right or how to improve, the next action. Grounded in the verbatim line.

Surface 2

Deal Pulse

Their deals, each with rolling SPICED state, confidence, and risk. Warm-but-doomed deals flagged, not scored healthy.

Surface 3

GTM Pulse overview

Team altitude: adoption, segments, funnel, the consistency gap between top reps and the rest.

Manager workflow

  • 1Opens on the existing pipeline review, not a new ritual
  • 2Sees the team pattern first: where SPICED breaks as a habit
  • 3Drills to one rep, sees the exact moments and the better response
  • 4Gets the coaching plan: playbook section plus a ready prompt for the 1:1
  • 5Comes back weekly because the number moved, or did not

Rep workflow

  • 1Sees own gaps as growth, with a manager-override on false positives
  • 2Reads the customer line, own response, and the ideal response
  • 3Gets the agent's prompt to practice the impact question
  • 4Runs it on the next real call
  • 5Sees the gap close on the dashboard next cycle
Price · hypothesis only

Working shape: per-seat per month plus a token-metered analysis cost, sized against the roughly one to three dollars per learner we modeled on the course. Packaged standalone or bundled with the refreshed training. These are hypotheses to test with the MDs, not settled numbers, and the routing economics that justify a price are not measured yet.

WS2 · the vertical slice

One stage, one rep, engineered end to end.

Not the full map at low resolution. A worked example Nathan can build Monday, grounded in a real scored account.

Bowtie stage: Discovery to Qualified
Rep: John Haire · Account: CBRE · live score from the Momentus pipeline
DimensionScoreStateGap
Impact2.0weakRep never pushed pain to a business consequence
Critical Event1.5weakNo compelling event pinned
Decision Process2.2weakApprover path unmapped
Situation3.5solidno gap
Customer said

"The only way we can process money is physical check or ACH. We don't do credit card payments. And if the money was lost, it's going to come back on me."

Rep said

"I'm so glad you said that. We have a demo just on Momentus payments, our native payment channel."

Ideal response

"That is a real risk. If a payment slipped through on a 25,000 dollar buyout, what would that cost in lost revenue or rework, and how often does that near-miss happen?"

Entry
Deal reaches Discovery to Qualified with an Impact score below 2.5.
Agent action
Hands John the Impact-quantification play plus a ready prompt to draft the impact question for CBRE.
Exit
Next call shows an asked impact question and a quantified consequence, Impact rescored above 3.
Handoff
Manager (Josh Klatz) gets the before and after in the pipeline review, coaches or clears it.
Before Monday *

Ship two things. Hold the rest.

The whole win is walking in with one clear system Nathan can engineer against. That is WS1 and the WS2 slice. WS4 (the agent) and WS5 (the course) build on the data model, so they follow, not lead. Everything else waits.

1. The one-page product definition
Three surfaces, both workflows, flagged price. Sourced from the Momentus PRD.
2. The WS2 vertical slice
One stage, one rep, entry, exit, handoff, one agent action. The worked example above.