You are not there to close. You are there to price and de-risk. This deal sets the valuation anchor for the follow-on cash round — that is its biggest power and its biggest trap.
1 · The three outcomes you want
- Scope. A written list of what "resources" means. Roles, hours, duration, exclusions.
- Value. How they price those hours, and what equity they want. This is the valuation lever.
- Fit. Proof they can run a life-safety product. Not "we have 40 devs" — named people, named skills, an on-call answer.
2 · What ELA actually needs
The product is built. The need is maintenance, on-call, and steady feature flow. The honest package:
| Role | Load | Why |
|---|
| Senior React Native / Expo engineer | Full-time | The mobile app is the product. RN 0.83 + Expo SDK 55, EAS/TestFlight releases. |
| Senior TypeScript backend engineer | Full-time | Node 22 API, Next.js admin, Postgres 17, dispatch orchestration. |
| iOS-native Swift/WatchKit capability | Fractional | The Apple Watch app is native Swift. Niche — must be a named person, not assumed. |
| DevOps / SRE | Fractional | Small infra (one VPS), but life-safety uptime expectations. |
| 24/7 on-call rotation | Shared | The real ask. A member under arrest at 03:00 is a P0. Needs ≥3 people who know the system. |
| UI designer | On-demand | They lead with design. Use for website + marketing assets. Accept it; don't overvalue it. |
DECIDEDAsk for: 2 full-time seniors + fractional Swift + fractional DevOps + a 24/7 on-call rotation. Your "two or three personnel" instinct is right — the on-call rotation and the Swift/watch skill are what a generic shop will miss.
Say plainly: this codebase is developed AI-natively (agent tooling, strict CI gates, visual regression, runbooks). Their engineers must work with that toolchain. Ask how their devs use AI tools today.
3 · Map to your budget lines
| Budget line | They cover? | Note |
|---|
| Dev resources | Yes — the core offer | Negotiate hard here. |
| Design | Yes | Website + ad/brochure assets for any marketing firm you hire. |
| Outsourced tech support (L1) | Maybe — ask | Member support ≠ dev. Different skill, language, hours. |
| Outsourced call center | Almost certainly no | UAE-hours, arrest-context ops desk. Keep in cash budget. |
| Server / cloud services | Almost certainly no | Small anyway: one VPS + usage-based vendors. Low hundreds/month at launch. Cash. |
| Equipment / licenses | No | Test devices, Apple org account, background-geolocation license, SMS/email usage. Cash. |
| Contingency | No | Cash by definition. |
DECIDEDTheir offer replaces the people lines — which dominate the budget — but you still need cash for ops, licenses, call center, and marketing. Say that in the meeting so nobody pretends this deal alone funds launch.
4 · Questions to ask
A — Scope of the offer
- What roles do you offer? List them.
- How many hours per month, per role?
- For how long is the commitment? 12 months? 24?
- What is excluded? Hosting? Licenses? Support? Say what is not included.
- Do you offer L1 member support? In English? At UAE hours?
- Can we increase or decrease the team month to month?
B — Deal mechanics (the valuation lever)
- How do you value one hour of your work in this deal? Show the rate card.
- Is that your list price or your cost? Services-for-equity is routinely priced at 2–3× cash cost. Push to a rate you would actually pay in cash.
- What total value do you commit? Over what period?
- What equity do you want for it?
- Does the equity vest as work is delivered, or up front? Insist: vest on delivery, clawback if they stop.
- If we are unhappy with a person, can we swap them? How fast?
- If we are unhappy with the firm, can we convert the remainder to cash or wind down? On what terms?
- Do you want a board seat or information rights? Ideally: information rights only.
C — People and continuity
- Name the people you would assign to us. Not the bench — the people.
- Can we interview them and reject them?
- Do the same people stay on our account? For how long?
- What is the seniority mix? We need seniors, not a pyramid.
- What working-hours overlap do you guarantee with UAE (GMT+4)? Romania is GMT+2/+3 — good, but confirm.
- Who manages the team day to day — us or your PM?
D — Capability fit (make them prove it)
- Which React Native apps in production do you maintain today? Name two.
- Have you shipped through EAS and TestFlight? Who has?
- Who has written a native watchOS app in Swift? Name the person.
- Node + TypeScript + Postgres — which production systems?
- Have you run a system where downtime harms a person, not just revenue?
- How do your engineers use AI coding tools today?
E — On-call and incidents (the hard one)
- Will you staff a 24/7 on-call rotation for our production system?
- What response time do you commit for a P0 at 03:00?
- Who carries the pager in week one? Name them.
- What is your incident process? Postmortems?
- How many months until your team handles a P0 without me? Commit to a date.
F — Security, IP, compliance
- All work is work-for-hire. All IP assigns to our company. Confirm in writing.
- No subcontracting without written consent. Confirm.
- Who gets production access? Named individuals only; least privilege.
- Do you run background checks on staff with production access?
- Do you carry professional indemnity insurance?
- Do you build for any competitor? Will you tell us if you start?
G — References
- Give two client references — products you have maintained 12+ months, not built and left.
5 · Deal traps
TRAPInflated service valuation → distorted priced round. If they book services at list rate for equity, your implied valuation rests on a number nobody would pay in cash. The follow-on investor re-derives it and either reprices you down or walks. Fix: cash-equivalent rates — or structure as a convertible that prices at the next round. That may serve you better than "setting the valuation" with a services deal, which is a weak anchor.
TRAPService-provider equity is a cap-table smell. Future investors dislike a dev shop holding a big, unvested stake. Fix: vesting tied to delivered milestones, clawback on non-delivery, modest stake.
TRAPCaptive-team risk. If they are your whole engineering org and a shareholder, firing them is a crisis. Fix: contractual swap and wind-down rights; keep the right to hire in-house later.
TRAPHours ≠ outcomes. "40 devs available" is worth nothing. The commitment is named people, fixed allocation, on-call SLA, delivery milestones.
TRAPThe 40-dev bench cuts both ways. Great burst capacity, but bench economics push them to staff whoever is free. Continuity clauses are the defense.
6 · Website and web app — the stance
- DECIDEDMarketing website: yes, for launch. App Store review, trust, legal pages, paid-acquisition landing. Small, static. Good first test project for the partnership — their designers can shine here.
- DECIDEDMember web app: no. The emergency trigger is inherently mobile — background location, watch relay, biometrics, push. A browser cannot do the core job, and a web version adds a whole new attack surface to a life-safety product. Only future web candidate: a member account portal (billing, family management), post-launch.
- The admin portal already is a web app. If they push web work, channel it to the marketing site and admin improvements.
7 · Introducing the codebase
Reframe the worry. You are not a founder who can't answer code questions — you run an AI-native development process with unusually strong documentation, CI gates, and runbooks. The evidence below proves it. Most partners inherit far worse from human teams.
- Tomorrow (no NDA yet): share the map, not the territory. The corrected end-to-end flow diagram, the one-pager below, the architecture line: "Server-authoritative — client renders, server decides. React Native mobile, native Swift watch app, Node/TypeScript API, Next.js admin, Postgres, single hardened UK-hosted VPS."
- No repo access tomorrow. Repo access comes after NDA + term-sheet intent, as their technical due diligence.
- Their tech DD is your DD on them. Ask their engineers for a written assessment of the repo: what they'd improve, what worries them, how they'd staff it. Specifics = keeper. "Looks fine, we can start Monday" = they didn't read it.
Deflection lines for deep technical questions:
- "The repo answers that better than I can from memory — that's what the DD phase is for."
- "Every subsystem has a runbook; dispatch and duress flows are life-safety-critical with their own drill procedures."
- "The right people to talk code are your engineers and my toolchain, in DD."
You never need to bluff. The docs are the answer.
8 · Evidence pack
8.1 — Codebase one-pager (counted from the git tree, 2026-08-13)
- Shape: monorepo, 3 apps + 11 shared TypeScript packages. Mobile (RN 0.83 / Expo SDK 55 / React 19), native Swift WatchKit watch app, framework-free Node 22 API with Zod contracts at every boundary, Next.js ops console.
- Size: ~223,000 lines TS/TSX (566 files) · ~3,500 lines Swift · 140 SQL migration files in 66 append-only directories · ~1,000 commits in 3.2 months.
- Quality gates: 162 test files incl. 69 API contract tests · 25-step CI gate (lint, typecheck, contract + breaking-change checks, PII-redaction check, security greps) · visual-regression harness covering all 31 mobile screen states.
- Docs: 771 files — 18 numbered incident runbooks, go-live checklist, pentest report (2026-05), compliance pack (data-flow map, subprocessors, DSAR workflow), App Store submission pack. Every doc opens with a 7-line greppable summary.
- Infra: one hardened UK VPS — Caddy + auto-TLS, Postgres 17, systemd services, self-hosted analytics/error/log stack. Prod runs in a documented safe mode today; flipping it is the documented launch gate.
- Pending licenses/creds (your "licenses" cash line, itemized): background-geolocation license (launch blocker), SMS/WhatsApp sender approval, email domain setup, KYC vendor account, maps token, push key, live payment mode, off-box backup storage, Play Console fee.
- Honest flags for DD, not to hide: single committer (bus factor of one — exactly what this deal solves); one duplicate iOS directory to clean up; some vendor integrations code-complete but unverified against live sandboxes.
8.2 — Flow diagrams
- OPENEnd-to-end flows diagram: was stale (old dispatch semantics, old pricing, banned tier phrasing); a correction pass is running now. Show it tomorrow only after the redeploy is confirmed.
- DECIDEDBirds-eye journey diagram: do not show tomorrow. All 54 screenshots predate the retheme and the dispatch lane states the inverse of shipped behavior. Full regeneration scheduled after the meeting (~1 day).
9 · Open questions — decide before the call
- OPENWhat equity range will you discuss, and priced round vs convertible? Don't improvise it live.
- OPENSole dev partner, or one of two (them + your own hires when the cash round lands)? Changes the continuity asks.
- OPENIs L1 member support in scope for them, or does it stay in the UAE call-center cash line?
- OPENBring the business-plan numbers — the people-cost line they replace is the anchor for valuing their offer.
Source: docs/business/2026-08-14-dev-partner-investor-brief.md · unlisted URL, no auth — do not add secrets to this page.