Knowledge base
Manage the chatbot's knowledge corpus. Active chunks live in Cloudflare Vectorize and surface in real-time chat responses.
- blogThe 7-specialist review pipeline that keeps AI features from breaking in production/ The seven specialists
[The 7-specialist review pipeline that keeps AI features from breaking in production — The seven specialists] Each specialist is a focused review prompt with a checklist. I run them as Claude subagents from the main session — each gets the diff plus the relevant context (schema,…
- blogThe 7-specialist review pipeline that keeps AI features from breaking in production/ The seven specialists
[The 7-specialist review pipeline that keeps AI features from breaking in production — The seven specialists] The trap this catches: AI features that work in dev because the dev DB has 50 rows and collapse in prod because they introduced an unindexed full-table scan. ### 3. AI …
- blogThe 7-specialist review pipeline that keeps AI features from breaking in production/ The seven specialists
[The 7-specialist review pipeline that keeps AI features from breaking in production — The seven specialists] What's the worst-case latency? What's the request budget per user action? Are model calls in a critical render path, or async? Are there caches, and are they invalidated…
- blogThe 7-specialist review pipeline that keeps AI features from breaking in production/ How does the seven-specialist pass actually work?
[The 7-specialist review pipeline that keeps AI features from breaking in production — How does the seven-specialist pass actually work?] Concretely, in the workflow: 1. PR is opened with the AI feature changes. 2. Pre-commit hook runs the unit tests and type-checks. 3. Author …
- blogThe 7-specialist review pipeline that keeps AI features from breaking in production/ How does the seven-specialist pass actually work?
[The 7-specialist review pipeline that keeps AI features from breaking in production — How does the seven-specialist pass actually work?] The whole pass takes 4–8 minutes wall-clock for a typical feature PR<sup class="sidenote-mark">①</sup>. Author wait-time is roughly the time …
- blogThe 7-specialist review pipeline that keeps AI features from breaking in production/ Why seven specialists, not three or fifteen?
[The 7-specialist review pipeline that keeps AI features from breaking in production — Why seven specialists, not three or fifteen?] Three is too few — the lenses overlap enough that you miss whole categories. Fifteen is too many — diminishing returns and review fatigue set in. …
- blogThe 7-specialist review pipeline that keeps AI features from breaking in production/ What it has caught for me
[The 7-specialist review pipeline that keeps AI features from breaking in production — What it has caught for me] A non-exhaustive list from the last twelve months: - A voice agent that quoted stale eligibility caps because its tool was reading from a 12-minute cache (perf + …
- blogThe 7-specialist review pipeline that keeps AI features from breaking in production/ How do you get started with this review pipeline?
[The 7-specialist review pipeline that keeps AI features from breaking in production — How do you get started with this review pipeline?] Copy the seven names. Write a one-page checklist for each one yourself — what does "architecture" mean in your stack, what does "observabilit…
- blogAnatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision/ What this means for HEI operators evaluating vendors
[Anatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision — What this means for HEI operators evaluating vendors] The HEI market in 2026 sits at the inflection the consumer-mortgage-tech market hit around 2018: a wave of "AI" tools that look impres…
- blogHow we cut a home equity platform's lead-to-close cycle in half
[How we cut a home equity platform's lead-to-close cycle in half] This is a field report. Names are abstracted; the work is real. The company runs a multi-program home equity portfolio — HEI, HEA, SLB, shared-equity, plus an alt-finance product. Series Seed–B. Twelve people. Wh…
- blogHow we cut a home equity platform's lead-to-close cycle in half/ What did the funnel look like before we started?
[How we cut a home equity platform's lead-to-close cycle in half — What did the funnel look like before we started?] The funnel had four problem areas, in priority order: **1. Inbound response.** Leads were calling the main line. The main line was a human SDR. After 5 p.m. or b…
- blogHow we cut a home equity platform's lead-to-close cycle in half/ What did the funnel look like before we started?
[How we cut a home equity platform's lead-to-close cycle in half — What did the funnel look like before we started?] **3. Enrichment gap.** Lead intake captured name, email, phone, property address, and self-reported home value. Underwriting needed all of that plus county tax re…
- blogHow we cut a home equity platform's lead-to-close cycle in half/ Component 1 — Voice intake
[How we cut a home equity platform's lead-to-close cycle in half — Component 1 — Voice intake] We wired Retell + Claude + Deepgram into the existing phone number. The agent picks up inside 1.5 rings, qualifies on six fields (state, age, property type, estimated home value, estim…
- blogHow we cut a home equity platform's lead-to-close cycle in half/ Component 1 — Voice intake
[How we cut a home equity platform's lead-to-close cycle in half — Component 1 — Voice intake] **Impact:** Time-to-first-response on inbounds dropped from a median of 17 minutes (during business hours) to 22 seconds. Off-hours response went from "next-day callback" to "in-call q…
- blogHow we cut a home equity platform's lead-to-close cycle in half/ Component 2 — Eligibility config
[How we cut a home equity platform's lead-to-close cycle in half — Component 2 — Eligibility config] We pulled all the eligibility logic out of the calculators and into an admin-driven settings table. State whitelists, LTV caps, CLTV caps, ownership rules, minimum equity, FICO f…
- blogHow we cut a home equity platform's lead-to-close cycle in half/ Component 3 — Agentic enrichment
[How we cut a home equity platform's lead-to-close cycle in half — Component 3 — Agentic enrichment] We built a Claude agent with web-search tools that takes a lead's address and fills out the underwriting profile. County tax record, lien status, AVM (we use three different sour…
- blogHow we cut a home equity platform's lead-to-close cycle in half/ Component 4 — Post-acceptance contract gen
[How we cut a home equity platform's lead-to-close cycle in half — Component 4 — Post-acceptance contract gen] Replaced the Word-template DocuSign flow with a templating system reading from the same admin config. Program names, terms, jurisdictional language all live in one plac…
- blogHow we cut a home equity platform's lead-to-close cycle in half/ What broke during rollout, and how did we fix it?
[How we cut a home equity platform's lead-to-close cycle in half — What broke during rollout, and how did we fix it?] The voice agent confidently quoted a wrong CLTV cap on one call in week three. The cap had been updated in the admin config; the voice agent's eligibility-check …
- blogHow we cut a home equity platform's lead-to-close cycle in half/ What stuck six months in?
[How we cut a home equity platform's lead-to-close cycle in half — What stuck six months in?] All four components are live, six months in. The lead-to-close median is at 7 days, down from 14. The same team — twelve people, no new headcount — is processing roughly 2.2× the lead v…
- blogYou can't hire a senior AI engineer fast enough — here's the math
[You can't hire a senior AI engineer fast enough — here's the math] If you're running a $2–10M revenue company in real-estate finance, you already know you need AI infrastructure. Not a chatbot. Voice intake. Lead enrichment. CRM consolidation. Sequence orchestration. The work s…
- blogYou can't hire a senior AI engineer fast enough — here's the math/ What does a senior AI engineer actually cost in 2026?
[You can't hire a senior AI engineer fast enough — here's the math — What does a senior AI engineer actually cost in 2026?] [Base salary for a senior AI engineer in the US sits around $220K](https://www.glassdoor.com/Salaries/senior-ai-engineer-salary-SRCH_KO0,18.htm). Add 25–35…
- blogYou can't hire a senior AI engineer fast enough — here's the math/ How long is the senior AI engineer hiring cycle?
[You can't hire a senior AI engineer fast enough — here's the math — How long is the senior AI engineer hiring cycle?] [The 2025 surveys put senior AI engineer search at 4.5–6 months](https://www.acceler8talent.com/resources/blog/ai-engineer--salary---market-rates-2025-2026/) fr…
- blogYou can't hire a senior AI engineer fast enough — here's the math/ Where are the senior AI engineers who can actually ship?
[You can't hire a senior AI engineer fast enough — here's the math — Where are the senior AI engineers who can actually ship?] The senior AI engineers who can ship production systems end-to-end — voice agents wired into dialers with drift detection, Claude-powered enrichment wit…
- blogYou can't hire a senior AI engineer fast enough — here's the math/ The alternative
[You can't hire a senior AI engineer fast enough — here's the math — The alternative] Fractional senior engineering. $7,500–$15,000 per month. Six-month minimum so the work has time to compound, but no equity, no recruiting cycle, no FTE conversion drag. You get the same stack …
- blogYou can't hire a senior AI engineer fast enough — here's the math/ The order of operations
[You can't hire a senior AI engineer fast enough — here's the math — The order of operations] If you're certain you'll need a full-time senior AI engineer eventually, the fractional path isn't a substitute — it's a head start. You ship the production stack now, with a partner wh…
- blogThe 7-specialist review pipeline that keeps AI features from breaking in production
[The 7-specialist review pipeline that keeps AI features from breaking in production] The most common failure mode I see in shipped AI features is not "the model is wrong." It's "the model was right in the demo, and then something around the model failed in production." [Researc…
- blogAnatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision
[Anatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision] This is a field report. Names are abstracted; the work is real. The same multi-program home equity operator covered in the [previous field report](/blog/lead-to-close-half/) runs four HEI …
- blogAnatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision/ What does the architecture look like?
[Anatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision — What does the architecture look like?] The current state of the engine, in one sentence: per-program eligibility rules live in an admin-editable settings table, the calculators are determi…
- blogAnatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision/ What does the architecture look like?
[Anatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision — What does the architecture look like?] This took a deliberate choice early. The pull toward "just throw it in a JSONB column" is strong at every Seed-stage shop. The pull is also wrong. JS…
- blogAnatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision/ How do four investor overlays run on one decision graph?
[Anatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision — How do four investor overlays run on one decision graph?] Each investor overlay evaluates independently against the same applicant snapshot. The engine returns all per-investor verdicts pl…
- blogAnatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision/ How do four investor overlays run on one decision graph?
[Anatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision — How do four investor overlays run on one decision graph?] The same engine runs the proprietary sale-leaseback product, with one wrinkle: a state-specific LTV override. The national cap is …
- blogAnatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision/ Why is there no LLM in the decision graph?
[Anatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision — Why is there no LLM in the decision graph?] LLMs are great at parsing unstructured documents. They are great at normalizing inbound data: a lead form that arrived as a free-text email, a v…
- blogAnatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision/ Why is there no LLM in the decision graph?
[Anatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision — Why is there no LLM in the decision graph?] A deterministic engine reading from a versioned settings table gives you all three for free. An LLM-in-the-loop gives you none of them. You can …
- blogAnatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision/ What broke
[Anatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision — What broke] Four production fixes hit the eligibility layer in a twelve-day window after the per-investor admin tabs went live in mid-May. All four were caught by the architecture's loggin…
- blogAnatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision/ What broke
[Anatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision — What broke] **Admin LTV cap wasn't actually the source of truth.** The admin Settings UI exposed an LTV cap as an editable knob. Three different code paths (a primary calculator, a shared …
- blogAnatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision/ What broke
[Anatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision — What broke] **Investor proceeds cap silently bypassed in the partner wizard.** The admin-editable per-investor proceeds cap wasn't being applied in the partner-facing proceeds-range formul…
- blogAnatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision/ What broke
[Anatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision — What broke] **Empty admin array crashed every request.** Production had one program's `ineligible_ownership_types` set to an empty array — an intentional admin choice to allow LLC ownershi…
- blogAnatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision/ What this means for HEI operators evaluating vendors
[Anatomy of an HEI eligibility engine: four investor overlays, zero LLMs in the decision — What this means for HEI operators evaluating vendors] Three questions to ask any vendor selling you an "AI eligibility engine": First: show me the audit row for a decision made 90 days ag…