01 / 16
Winning by Design × Snowflake

Embedding Revenue Architecture into Snowflake's GTM Agents.

A working conversation on how your 3,700-person global GTM organization transitions to an agentic, consumption-first operating model.

April 2026 · Confidential
Act 1 · Context 02 / 16
Who We Are

We are the global standard for Revenue Architecture.

Winning by Design helps GTM organizations design predictable, compounding revenue engines. We wrote the book on Revenue Architecture. Our work is the operating language for thousands of teams across every segment.

Act 1 · Context 03 / 16
What We Do

We design the machine that produces recurring revenue.

In a consumption model, the sale is the starting line, not the finish line. Revenue is re-earned through recurring customer impact. We architect every stage of that journey so impact compounds.

Frameworks

Revenue Architecture

Six models connecting GTM strategy to financial outcomes. Revenue, Data, Mathematical, Operating, Growth, GTM.

Language

SPICED

Situation, Pain, Impact, Critical Event, Decision. The operating language across the full customer lifecycle.

Method

Growth Governance

How leadership teams run the revenue engine with rigor. Coaching, forecasting, QBRs, board reporting.

Act 1 · Context 04 / 16
What We Believe

Three principles shape every deployment.

01

Recurring impact drives recurring revenue.

Customers who achieve measurable impact expand. Customers who do not, churn. Every GTM motion must be architected around impact delivery.

03

The manager is the multiplier.

Adoption lives in the 1:1, the forecast call, and the QBR. If managers do not coach it, the field will not run it. Leadership owns adoption.

Act 1 · Context 05 / 16
The First Requirement

A shared language across the organization. Everything else builds on top of it.

Agents, CRM, and coaching systems only work when they speak the same language as the humans around them. The first requirement of agentic GTM is not better software. It is a common language that every seller, CSM, manager, marketer, and agent understands the same way.

Without a common language

Every system amplifies confusion.

Agents scored on one rubric. Managers coaching to another. CRM fields capturing a third. Each layer is sharper. The signal is noisier.

With a common language

Every layer compounds.

Agent prompts, CRM fields, manager 1:1s, QBR templates, and coaching rubrics all speak SPICED. Clarity replicates. Signal compounds.

Act 1 · Context · Handoff 06 / 16
SPICED

The common language that spans the full customer journey.

Unlike qualification frameworks built for the sales cycle only, SPICED works from first discovery call through implementation, renewal, and expansion. The impact definition created on the first call is the same definition used at year three.

S
Situation
Context of the customer's world today
P
Pain
The problem they are trying to solve
I
Impact
Measurable outcome they must achieve
CE
Critical Event
Business urgency, not artificial deadline
D
Decision
Path to commit across the buying committee

SPICED is machine-readable by design. Every dimension maps to a CRM field, a coaching rubric, and an agent prompt. It is the only GTM language built to operate inside an agentic system as naturally as it operates in a human conversation.

Act 2 · Your Seven Questions · 01 07 / 16
Question 01

What does AI-first deployment actually look like?

It begins with the common language. Then the language runs in three places at once. The prompt library every rep can use today. The scoring logic embedded in your internal agents. The certification that gates enterprise rollout.

Surface

SPICED Prompt Library

Every rep builds a personal SPICED assistant in Gemini, Claude, Copilot, or ChatGPT on day one. No new infrastructure required.

Agents

Embedded in Your Stack

SPICED scoring logic, CRM field mapping, and coaching flows inside Snowflake's own agents. We guide. Your team builds.

Quality Gate

Agent Certification

Before enterprise go-live, we certify your agents score SPICED correctly. No scale without calibration.

We run this internally. Our agent is called Jack. Jack is trained on the full Revenue Architecture IP and scores our own customer calls every week. Happy to demo live.

Act 2 · Your Seven Questions · 02 08 / 16
Question 02

Where is this deployed today?

Two honest answers, because they matter differently. SPICED as the operating language is everywhere. Agentic SPICED, embedded in a customer's own internal agents at enterprise scale, is the emerging motion. Snowflake would be the flagship.

Today

SPICED as Language

Deployed across 2,000+ client organizations. Global Payments is a flagship train-the-trainer example where frontline managers became the reinforcement system.

Flagship Motion

SPICED inside Your Agents

Pioneer deployment at enterprise scale. We run this internally with Jack today. Snowflake is positioned as the flagship case study under the Section VII co-marketing terms.

The outcome-aligned structure of our proposal reflects this honestly. We share the risk of a pioneer motion. We are not pitching you on a finished platform. We are partnering with you to define the new category.

Act 2 · Your Seven Questions · 03 09 / 16
Question 03

How do you enable 3,700 people across a global organization?

Top-down and systems-first. Methodology adoption fails when it starts with the field. We invert that sequence. Leaders first, agents second, managers third, field absorbs through the work.

PHASE 1 · WKS 1-3
ELT + SVPs

Alignment workshop. Bowtie as the shared revenue model.

PHASE 2 · WKS 4-12
AI + IT Team

Agent training guidance. Prompt architecture and scoring.

PHASE 3 · WKS 2-6
RevOps + Finance

GTM Diagnostic. Baseline across five selling motions.

PHASE 4 · WKS 6-14
Frontline Managers

Certification in remote cohorts of 20 to 25. Six parallel streams.

PHASE 5 · WKS 10-12
Facilitators

Enablement team certified. Snowflake owns onboarding forever.

PHASE 6 · WKS 17-18
Agent Certification

Quality gate before enterprise go-live.

For the 3,400 ICs, three mechanisms run in parallel. An asynchronous IP library on your LMS. The manager coaching cadence in every 1:1 and QBR. And the agent itself, which surfaces the missing SPICED element in real time. The agent is the enablement.

Act 2 · Your Seven Questions · 04 10 / 16
Question 04

How do we manage the transition from MEDDPICC to SPICED?

Pick it, stick it, and evolve together. MEDDPICC is a qualification scorecard built for the sales cycle. SPICED is the operating language across the full customer journey. We work with your MEDDPICC investment rather than replacing it, so the transition is additive instead of disruptive.

What Stays

Respect what works.

MEDDPICC qualification fields and talk tracks that are proven stay in place. No rip-and-replace. No parallel double-entry.

What Evolves

Add the journey language.

SPICED extends coverage to post-sale: onboarding, impact realization, renewal, and expansion. The consumption model is the right side of the Bowtie. MEDDPICC does not serve that terrain. SPICED does.

We have seen this work when leadership coaches SPICED personally in forecast calls and QBRs inside the first 60 days. We have seen it stall when organizations ran it as a training event and never changed the management cadence. We design the deployment to land the cadence change first.

Act 2 · Your Seven Questions · 05 11 / 16
Question 05

How do we leverage the methodology with AI beyond launch?

Launch is week 18. Everything interesting happens after week 18. Perpetual methodology is a closed loop. Every call scored. Scoring data retrains the next agent version. Coaching dashboards updated weekly. IP updates flow into the agents continuously during the contract.

LOOP 01
Score Every Call

Agent scores SPICED dimensions on every customer conversation.

LOOP 02
Retrain Quarterly

Scoring outcomes retrain the agent on your actual customer voice.

LOOP 03
Coach on Pattern

Managers see SPICED progression across their team weekly.

LOOP 04
IP Flows In

New playbooks and role IP flow into agents during the contract.

LOOP 05
Benchmark

Anonymized cross-customer benchmark becomes the network effect.

You are not buying a 2026 methodology. You are buying continuous R&D that runs across thousands of calls per week. That is the compounding.

Act 2 · Your Seven Questions · 06 12 / 16
Question 06

AI automation versus human touch. Now, and in 12 to 24 months.

The honest map. Today, AI is strong at preparation, scoring, inspection, and drafting. Humans still own the moments that make or break the deal and the relationship. In 18 months, the management layer shifts before the rep layer does.

AI owns today

Preparation, scoring, inspection, drafting

  • Discovery prep and account research
  • Post-call SPICED scoring and summary
  • Next-best-action recommendations
  • Deal inspection and forecast anomaly detection
  • Email and follow-up drafting
  • Onboarding checklist adherence in CS
Humans own today

Trust, judgment, novel context

  • The moment of commit with the economic buyer
  • Executive sponsorship and board relationships
  • The difficult retention conversation
  • The novel impact thesis not in training data
  • Pressure-testing a Critical Event with a CFO
  • Reading the room when it contradicts the transcript

In the 12 to 24 month horizon, AI moves from post-call coaching to in-ear coaching during the call to running portions of the call autonomously. The management layer feels this first and hardest. Reps six to twelve months behind.

Act 2 · Your Seven Questions · 07 13 / 16
Question 07

How do you use AI to build perpetual methodology enablement content?

Enablement content is no longer authored. It is extracted from the work. Your team stops being content producers and becomes curators, coaches, and system designers. Same headcount. Ten times the leverage.

01
Calls Recorded

Gong, Zoom, or direct through your agents.

02
Scored

SPICED agent scores every transcript.

03
Exemplars Selected

High and low performers flagged and anonymized.

04
Content Generated

SPICED good vs bad clips, role-play scenarios, coaching cases.

05
LMS Updated

Enablement team curates and publishes.

06
Agent Retrained

Exemplars feed the next agent version.

Your enablement function evolves into growth governance. The operating cadence takes over from the workshop.

Act 3 · Co-Creation 14 / 16
Let's Structure This Together

Four pillars we want to shape with you.

This is a new motion for both of us. We have a proposal on the table. We want to use the back half of today to refine the offering with you across four pillars.

01
Adoption Signal

What counts as adoption for Snowflake? Buyer response, rep behavior, or agent accuracy. We need your ground truth.

02
Accountability

The fourth leg most organizations miss. Classroom, roleplay, content, and adoption measurement. The agent layer is the accountability mechanism.

03
Calibration

What does good look like at Snowflake, specifically. First 90 days we score your existing call library to set the baseline before anyone is trained.

04
Compounding

How the investment gets more valuable in year two and year three without Snowflake paying more. The benchmark is the moat.

Act 3 · Close 15 / 16
What We Ask For Today

Three commitments before we leave this call.

01

AI + IT Working Session

Schedule the next conversation with your AI and IT team to walk through the agent architecture and SPICED embedding approach.

03

Refined Offering in 14 Days

We return with the offering structured around the four pillars you helped shape today. Specific. Dated. Ready to sign.

16 / 16
Thank you

We designed this for Snowflake.
Help us finish it.

Matt Kaminski, Managing Director · matt.kaminski@winningbydesign.com · 415.609.7708

Winning by Design · winningbydesign.com · Confidential