Sales Engineer
RFP responses, competitive feature matrices, and technical demos.
$skillfish add ~skills/sales-engineer Real Conversations
See how the Sales Engineer skill responds to actual requests. These are representative examples of what your AI agent will produce.
Client is asking us to fill out a 120-question RFP for a $200K deal. How do I approach this efficiently?
120 questions for a $200K deal is high overhead — approach it strategically, not exhaustively.
Step 1 — Qualify the RFP Before You Fill It (today)
Call your champion before touching the document:
- "Is this RFP driven by procurement process or is there a competing vendor we're being benchmarked against?"
- "Which 10-15 sections are most critical for your evaluation team?"
- "Are there any must-fail criteria we should know about?"
If your champion can't answer these, your deal is weaker than you think.
Step 2 — Categorize All 120 Questions
Tier 1 — Differentiators (15-20 Qs): Where you win. Write these yourself with proof.
Tier 2 — Table stakes (60-70 Qs): Standard security/compliance Qs. Build a reusable answer library.
Tier 3 — Irrelevant/Boilerplate (30+ Qs): Fill these in 30 seconds each.
Step 3 — Build the Response Machine
- If you have an existing RFP library: match questions to existing answers first (~60-70% will match)
- For new questions: write once, add to library
- For security questionnaires: use a tool like Responsive, Loopio, or a well-organized Notion database
Step 4 — The Cover Story Matters More Than the Answers A compelling 1-page executive summary at the front is read by decision-makers; the 120 answers are reviewed by procurement. Invest disproportionately in the summary.
Time budget: For a $200K deal, a 15-hour investment (3 hrs qualification + 8 hrs filling + 4 hrs review) is reasonable. More than that, reconsider whether you're actually in this deal.
How many of the 120 questions are security/compliance vs. product capability?
Get Started
How to install and use this skill in your preferred environment.
Skills are designed for AI coding agents (Claude Code, Cursor, Windsurf) and IDE-based workflows where the agent can read files, run scripts, and act on your codebase. Web-based AI can use the knowledge and frameworks, but won't have tool access.
Models & Context
Which AI models and context windows work best with this skill.
Recommended Models
Larger models produce more detailed, production-ready outputs.
Context Window
This skill's SKILL.md is typically 3–10 KB — fits in any modern context window.
All current frontier models (Claude, GPT, Gemini) support 100K+ context. Use the full window for complex multi-service work.
Pro tips for best results
Be specific
Include numbers — users, budget, RPS — so the skill can size the architecture.
Share constraints
Compliance needs, team size, and existing stack all improve the output.
Iterate
Start with a high-level design, then ask follow-ups for IaC, cost analysis, or security review.
Combine skills
Pair with companion skills below for end-to-end coverage.
Good to Know
Advanced guide and reference material for Sales Engineer. Background, edge cases, and patterns worth understanding.
Contents
Demo Narrative Structures
Before/After/Bridge: Open with the prospect's current state (Before), show the desired future state (After), then demonstrate how your product is the bridge. Works best for prospects who already understand their problem and are evaluating solutions. Keeps the demo anchored to business outcomes rather than feature walkthroughs.
Challenge/Solution/Proof: Open with a specific technical or operational challenge (validated in discovery), demonstrate your solution to that exact challenge, then close with proof — a customer story, benchmark, or live metric. Works best in technical demos where the champion is an engineer or architect who needs to see the mechanism, not just the outcome.
30-minute technical demo sequencing:
- Minutes 0–3: Recap the discovery takeaways ("Based on our conversation, the two things that matter most to you are X and Y")
- Minutes 3–20: Demo segments mapped to those two priorities — no feature tourism
- Minutes 20–25: Proof point (customer story or data)
- Minutes 25–30: Open questions and define the next step
RFP Trap Questions
| Trap Question | The Trap | How to Answer Without Over-Committing |
|---|---|---|
| "Do you support SSO?" | Prospect may mean SAML, OIDC, or both — you only support one | "Yes, via SAML 2.0. We're adding OIDC support in Q3. What IdP are you using?" |
| "Is your platform SOC 2 compliant?" | Type I vs Type II; in-scope systems vary | "We hold SOC 2 Type II certification. Happy to share the report under NDA and walk through the scope." |
| "Can you support on-premise deployment?" | Customer may mean airgapped or just VPC | "We support private cloud (customer VPC) deployment. Full airgap is available on our enterprise tier." |
| "Do you have 99.9% uptime SLA?" | Contractual SLA vs observed uptime are different | "Yes, our enterprise agreement includes a 99.9% uptime SLA with defined remedies. Here's our status page." |
| "Do you support custom data retention policies?" | May trigger architecture questions if you don't | "Configurable retention from 30 days to 7 years is available on Enterprise plans." |
| "Are you GDPR compliant?" | Compliance vs certification vs DPA availability | "We are GDPR-compliant. We offer a standard Data Processing Agreement and are happy to review any addendum." |
| "Does your API support bulk operations?" | Rate limits, batch size limits matter here | "Yes, our bulk API supports up to 10,000 records per request with dedicated rate limits on Enterprise." |
| "Can you integrate with [system they mentioned in passing]?" | Don't commit to undocumented integrations | "We have a native connector. Let me confirm the specific version compatibility and follow up by tomorrow." |
Proof of Concept Design
POCs fail when success is undefined at the start. Before writing a single line of integration code:
- Define success criteria in writing. "The POC succeeds if we can process X records in under Y seconds with Z error rate." Get the champion to sign off.
- Limit scope to 3 use cases maximum. A POC that tries to validate everything validates nothing. Pick the 3 use cases with the most budget justification weight.
- Set a hard end date. 4 weeks maximum. Beyond 4 weeks, internal momentum dies and competing priorities take over. Put the end date in the kickoff email.
- Document stakeholder roles. Who is the technical champion? Who is the blocker? Who has budget authority? The champion and the budget owner are rarely the same person — know who can end the deal and keep them engaged.
Competitive Displacement Tactics
When a prospect says "we're happy with [competitor]," they are usually happy with the relationship, not the product. Probe for hidden pain with SPIN questions:
- Situation: "How are you currently handling [specific workflow]?"
- Problem: "What are the limitations you've run into with that approach?"
- Implication: "What does it cost you when [that limitation] affects [downstream process]?"
- Need-payoff: "If you could eliminate that friction, what would that unlock?"
Reference customer matching: When displacing an incumbent, match reference customers by industry, company size, and the specific pain point — not just logo name. "We helped a fintech company your size migrate from [competitor] in 6 weeks" is more credible than "we have 500 customers."
SE-to-AE Handoff Documentation
After every discovery call or POC, document for the AE:
- Technical fit score (1–5): Does the product actually solve what they need? Any gaps?
- Integration complexity: What systems need to connect, and what is the estimated implementation effort?
- Technical objections raised: List them verbatim, with your recommended response for each
- Champion technical fluency: Is the champion an engineer, a technical manager, or a business buyer using technical language?
- Blockers identified: Names, roles, and specific concerns of anyone who pushed back during the technical evaluation
Include this in the CRM opportunity record, not just in email — the deal summary must be self-contained for anyone who picks it up.
Ready to try Sales Engineer?
Install the skill and start getting expert-level guidance in your workflow — any agent, any IDE.
$skillfish add ~skills/sales-engineer