Sales Play Executor · AI ROI

One prospect. Two prompts. One CFO-tuned outreach.

A two-stage AI-assisted workflow that turns any AI ROI conversation from generic outreach into CFO-grounded, sourced, house-of-brand-safe engagement. Stage 1 researches the specific human against public signals and produces a persona essay. Stage 2 uses that essay to write outreach with a marginal-cost hook, the cost-per-unit narrative, and a concrete next step — with every ROI claim traceable to Nucleus Research or public reference customers.

⚡ Wave 1 · Get in the room

How it works

Two stages. One outreach.

BYOI — bring your own AI. Paste the prompts into Claude, ChatGPT, or Gemini. Your API keys and prospect data never leave your browser.

🎯Pick a persona

Three ROI archetypes: CFO, VP Infrastructure, or FinOps Lead. Each reads a different line on the invoice and answers to a different KPI.

→

📝Generate persona essay

Stage 1 prompt — researches the specific prospect against public signals (LinkedIn, earnings-call commentary, industry press) and produces a grounded intelligence brief.

→

✉️Generate outreach

Stage 2 prompt — feeds the essay + the cost-per-unit narrative into a marginal-cost message with cold, follow-up, and warm-intro variants. Every ROI claim sourced.

Step 1 · Pick the persona

Three ROI archetypes.

Every cost axis in the model answers to a different persona. Choose the one closest to your prospect's role and P&L accountability.

STEP 2 Prospect signals

Provide as much or as little as you have. The Stage-1 prompt will explicitly demand grounding and flag any gap it cannot fill from public sources.

Prospect name *

Job title optional

Company *

Industry / P&L context optional

LinkedIn URL optional

Paste the URL. The prompt tells the AI to note if it cannot access it directly.

Additional context you already know optional

Outreach variant

The concrete ask

STAGE 1 Persona Essay Generator

Paste this into Claude, ChatGPT, or Gemini. The AI will research the prospect against public signals (LinkedIn, earnings calls, analyst commentary) and produce a persona essay you use as grounding context for Stage 2.

Prompt 1Persona Essay Generator

STAGE 2 Outreach Message Generator

After Stage 1 produces the persona essay, paste it into Stage 2 (replaces the essay placeholder). The result is a CFO-tuned outreach with hook, marginal-cost bridge, and concrete next step — pre-loaded with the AI ROI play context and every ROI claim traceable to Nucleus Research or public reference customers.

Prompt 2Outreach Message Generator

Guardrails — do not skip these

Every prompt above enforces these rules. The seller reviews the AI output against them before sending. Any output that fails a rule goes back into Stage 2 with a correction.

🛡️ House-of-brand

✓Matter-of-fact CFO tone — no marketing language.
✓Marginal-cost framing — every claim ties to a P&L unit.
✕Never use "leverage". Never open with "I hope this finds you well". Never use exclamation marks.
✕Never invent a specific competitor's cost failure. Reference publicly-reported FinOps patterns only.

📊 ROI-claim discipline

✓Every ROI claim must trace to Nucleus Research or a public reference customer. State the source.
✓Approved numbers: 427% ROI (Nucleus) · 14% year-3 cost reduction · $0.0009 vs $0.0686 per query · 62× query volume.
✕Never fabricate a projected ROI for the prospect's specific business.
✕Never quote a claim without its source in the same sentence.

📐 Naming (Guide D044422)

✓First reference always full: Teradata Autonomous Knowledge Platform
✓Use Teradata Cloud, Teradata Factory, Teradata Fabric, Active Compute / Elastic Compute.
✕Never AK Cloud, AI Factory, Artemis, ClearScape Analytics (as product name).
✓In-Database features: "features included in Teradata Database."

✍️ Seller self-check

?Does the hook reference a specific P&L moment for this person?
?Is every ROI number sourced in the same sentence?
?Is the next step tied to a P&L outcome, not a product demo?
?Does the tone read as CFO-native, not marketing?

Worked example

CFO · Insurance · DACH

Illustrative output showing what "good" looks like — an anonymized persona essay and the resulting outreach message. Names, companies, and public references are fictional composites; ROI numbers are from Nucleus Research (real).

STAGE 1 OUTPUT Persona essay

Michael Chen has been CFO at a top-15 European reinsurer for four years, having spent the prior decade at a Big Four audit firm and then as Deputy CFO at a systemically-important European bank. His public commentary — quarterly earnings calls, two industry keynotes in the last six months — consistently returns to a single theme: that the "AI capex" line on his P&L has grown 340% over 24 months without a matching movement in the ROIC line, and that his board is asking pointed questions about it. He writes on LinkedIn about the disconnect between AI POC pilots showing dramatic uplift and production workloads whose unit economics remain opaque. He has publicly criticized vendor pricing models that scale by usage without predictable ceilings — his phrase, in his October keynote, was "we bought a variable, and we needed a fixed." He is not known to have expressed a technology preference or vendor loyalty in public. His CTO reports into him, not laterally — which is unusual and means CTO priorities pass through his P&L filter. Public gap: no known conversation on re-tokenization cost (A2) or contention overhead (A5); his focus appears to be inference (A1) and idle capacity (A4).

STAGE 2 OUTPUT Cold outreach (Day 1)

Subject: "We bought a variable, we needed a fixed" — Nucleus 427% Michael, Your framing at the October keynote — "we bought a variable, and we needed a fixed" — has been sitting with me. Especially the sub-point about POC uplift not translating to production ROIC. That distinction is doing more work than most vendor slides acknowledge, and Nucleus Research's finding of 427% ROI across Teradata customers is one of the few analyst-grounded data points that lands on a marginal-cost basis rather than a POC-uplift one. We recently published a short piece called The Cost of More — built on published reporting from the battery industry: the flagship that had everything except the price of one finished unit, and the rival that prices every layer of a valid cell, week by week, and earned the margin to scale. Its companion model then walks the six axes of an enterprise AI invoice and which of them are architecturally addressable rather than merely negotiable. It reads as a companion to the "we needed a fixed" argument you have referenced publicly. No product pitch; just the invoice map. Nucleus separately measured 14% year-3 cost reduction under consumption pricing, which speaks directly to your ceiling concern. Worth a 30-minute conversation? Two dates that could work: [date A] or [date B]. Happy to send the parable link ahead of time so we can skip the introduction and go straight to the two axes I suspect are on your radar (inference and idle capacity). Best, [Seller name] — Guardrails self-check: ✓ hook tied to public statement · ✓ cost-per-unit framing referenced · ✓ every ROI number sourced (Nucleus) · ✓ concrete next step with dates · ✓ CFO tone · ✓ naming compliant

Wave 5 backlog

EXTENSIONExtend Executor to 9 personas total. Wave 1 ships CFO · VP Infrastructure · FinOps Lead (three primary ROI archetypes). Wave 5 adds Head of Data Platform · Chief Procurement Officer · Deal Desk Lead for the second ring — plus the same pattern applied to the AI Strategy and AI Sovereignty Executors. Total across the three plays: 27 persona variants.