# Oction Labs — Cold Email Sequences **Brand voice:** Deadpan, concrete, precise. No exclamation marks. No hype adjectives (revolutionary, cutting-edge, transformative). Sentence case. Short sentences. Concrete claims with unrounded numbers. Last line earns wit. **Structure rules:** - Subject line: specific, not spammy - Opening: one sentence, personalized or specific pain - Body: 2-3 short paragraphs max - CTA: one clear ask - Sign-off: professional, not sales-y - No attachments in first email - No generic pleasantries --- ## 1. Engineering (Geotechnical/Consulting) ### Touch 1: Intro / Pain **Subject:** Your senior engineers spend 5 hours a week searching for answers they already wrote **Body:** ``` Ten engineers hunting through shared drives for prior calcs and standards interpretations. Five hours a week each. At $250 an hour loaded, that is roughly $650,000 a year of billable capacity spent searching instead of engineering. We deploy private AI that ingests your reports, RFIs, and design standards into a sovereign knowledge base. Your team asks in plain language. The system retrieves the exact passage with the source cited. Worth a 15-minute call to see if the economics line up for your firm? Brandon Gill Founder, Oction Labs brandon@octionagency.com ``` ### Touch 2: Value / Proof **Subject:** 74% correct vs 5-10%: the difference retrieval makes **Body:** ``` Most AI invents answers. In our measured evaluation, retrieval-grounded systems scored 74% correct on technical questions versus 5-10% without grounding. Zero percent fabrication on trap questions. For engineering firms, that means a senior engineer can ask "how did we detail the pile cap on the 2023 riverfront job and why" and get the answer from your own drawing register in seconds. Not a guess. A cited passage they can verify. We are onboarding two engineering clients this quarter. One slot remains. Reply if you want to see it run on your own documents. Brandon Gill Founder, Oction Labs ``` ### Touch 3: Breakup / Soft Close **Subject:** Closing the loop **Body:** ``` I have reached out twice about private AI for engineering knowledge management. No response usually means timing, priority, or the problem is not acute enough yet. I will stop following up now. If you are losing time to document hunts or have a principal retiring with 30 years of unwritten knowledge, the offer stands. No discovery fees. No snapshot phase. Six to eight weeks to a live system. If not, best of luck with the practice. Eight agents. One organism. Brandon Gill Oction Labs ``` --- ## 2. Legal (Midsize Firms) ### Touch 1: Intro / Pain **Subject:** Associates spend 6-8 non-billable hours a week hunting precedents **Body:** ``` Every associate searching closed matters for clause language and memos they know the firm wrote. Every departure taking precedent knowledge with them. The bleed is roughly $22,000 a month in recoverable time for a 30-lawyer firm. We built a system that queries your own documents and cites the source passage. No hallucinated cases. No invented citations. In our measured evaluation: 0% fabrication on trap questions. Solicitor-client privilege stays intact because nothing leaves your building. Private deployment on your infrastructure. Open to a brief call this week? Brandon Gill Founder, Oction Labs brandon@octionagency.com ``` ### Touch 2: Value / Proof **Subject:** The AI citation problem (and the fix) **Body:** ``` Lawyers fear AI because it hallucinates citations. A dead case. A fabricated statute. Professional negligence exposure. Our system grounds every answer in your own documents. It retrieves, not generates. Your memos. Your opinions. Your clause banks. With the source document, page, and paragraph attached for human verification. The same question always returns the same grounded answer. Reproducible. Auditable. Defensible. We have one Standard deployment slot open next quarter. Reply if you want to see how it handles your actual precedent bank. Brandon Gill Founder, Oction Labs ``` ### Touch 3: Breakup / Soft Close **Subject:** Final follow-up: precedent retrieval **Body:** ``` I have written twice about private AI for legal precedent retrieval. Silence usually means the problem is not urgent or the firm has other priorities. Closing the loop. If you are watching associates burn non-billable hours on document hunts or facing a partner departure that will walk knowledge out the door, the system exists. If not now, perhaps when the next senior associate leaves. Eight agents. One organism. Brandon Gill Oction Labs ``` --- ## 3. Healthcare (Clinics/Operations) ### Touch 1: Intro / Pain **Subject:** Policy questions that interrupt your practice manager 20 times a day **Body:** ``` Front desk asking billing codes. Clinicians asking protocol exceptions. Admin staff re-answering the same questions. For a mid-size clinic, the bleed runs about $15,000 a month in interrupted workflow and after-hours catch-up. We deploy private AI that holds your protocols, billing rules, and regulatory guidance in a sovereign knowledge base. Staff get cited answers. The practice manager gets their day back. PHI never touches a third-party model. Runs on your network. Full audit trail for compliance. Worth 15 minutes to see the economics for your clinic? Brandon Gill Founder, Oction Labs brandon@octionagency.com ``` ### Touch 2: Value / Proof **Subject:** Privacy by architecture, not policy **Body:** ``` Most AI platforms pool data across customers. That is a non-starter for health records. Our deployments are single-tenant by design. Your data stays on your infrastructure. No shared models. No training on client data. Encrypted at rest and in transit. Every access logged. It is the difference between "we have a privacy policy" and "the data physically cannot leave." We are onboarding one clinic group this quarter. If you are facing a privacy audit or expanding to a new site, reply for a demo on your own policy documents. Brandon Gill Founder, Oction Labs ``` ### Touch 3: Breakup / Soft Close **Subject:** Closing the loop: clinic operations **Body:** ``` I reached out about private AI for clinic policy and operations. No reply typically means the bleed is not acute enough to prioritize. Final note. If your admin team is drowning in repeat questions or you are staring down a privacy audit, the system is built for exactly that. Sovereign deployment. Cited answers. Full audit trail. If the timing is wrong, file this away for when it is not. Eight agents. One organism. Brandon Gill Oction Labs ``` --- ## 4. Mining (Site Operators) ### Touch 1: Intro / Pain **Subject:** Site engineers inheriting decades of undocumented history **Body:** ``` New site engineer starts. The drill logs are in a filing cabinet. The geotech reports are PDFs scattered across drives. The closure documentation is in the head of a geo who retired last year. We deploy sovereign AI that ingests decades of site records into a private knowledge base. Runs on site, offline-tolerant. New staff query "what did we commit to on tailings monitoring in the 2019 amendment" and get the clause cited. Resource data never leaves the fence. Commercially sensitive. Regulator-ready. Open to a call to discuss your site knowledge challenges? Brandon Gill Founder, Oction Labs brandon@octionagency.com ``` ### Touch 2: Value / Proof **Subject:** Site knowledge that survives turnover **Body:** ``` Remote sites have unreliable connectivity. Regulators ask questions that take weeks to answer. Senior geos retire and decades of context walk out. We build systems that run on your infrastructure, at the site, with nothing leaving the fence. Drill logs, geotech reports, permit conditions, and inspection history made queryable with grounded citations. A new engineer inherits the site's full memory on day one. We have one multi-site deployment slot remaining this year. Reply if you are acquiring a new site or staring down a wave of retirements. Brandon Gill Founder, Oction Labs ``` ### Touch 3: Breakup / Soft Close **Subject:** Final note on site knowledge systems **Body:** ``` I have written twice about private AI for mining site knowledge management. No response usually means the problem is not pressing or procurement cycles are long. Closing the loop. If you are facing a regulator information request that will take weeks to answer or inheriting decades of undocumented records from an acquisition, the system exists. For now, best of luck with the operation. Eight agents. One organism. Brandon Gill Oction Labs ``` --- ## 5. Government (Municipal) ### Touch 1: Intro / Pain **Subject:** FOI requests that take weeks because the knowledge walked out **Body:** ``` Election cycles and retirements wipe institutional memory. The staff who knew where the 2014 variance decision lived have moved on. FOI requests sit in queues because finding responsive records takes days. We deploy sovereign AI for municipal departments. Bylaws, council minutes, policies, and procedure manuals made queryable with cited, reproducible answers. Full audit trail for every retrieval. Data stays resident. No third-party AI dependency. Aligned with data sovereignty requirements. Worth a conversation about your department's knowledge continuity? Brandon Gill Founder, Oction Labs brandon@octionagency.com ``` ### Touch 2: Value / Proof **Subject:** Consistent answers for consistent administration **Body:** ``` The same policy question should return the same grounded answer every time. That matters for consistent public administration. Our system retrieves from your own documents, not a public model's guess. The answer includes the source document, page, and paragraph. Your staff verify instead of trust. FOI support with a full audit trail of what was retrieved. Institutional memory that does not turn over when the election cycle does. We are working with one provincial agency this quarter. If you have a modernization budget or FOI backlog, reply to discuss scope. Brandon Gill Founder, Oction Labs ``` ### Touch 3: Breakup / Soft Close **Subject:** Closing the loop: municipal knowledge systems **Body:** ``` I have reached out twice about sovereign AI for municipal knowledge management. Silence usually means procurement cycles, budget timing, or competing priorities. Final word. If you are facing a wave of retirements in a long-tenured department or a FOI backlog that keeps growing, the system is purpose-built. Data residency. Full audit trail. No third-party dependency. If now is not the time, best of luck with the administration. Eight agents. One organism. Brandon Gill Oction Labs ``` --- ## 6. Logistics (3PL/Freight) ### Touch 1: Intro / Pain **Subject:** Dispatchers burning hours on "where is my freight" and "what does this surcharge mean" **Body:** ``` Ops staff chasing shipment status across email threads. CS reps escalating to managers for contract questions that are buried in rate sheets. Every interruption costs margin you do not have. We deploy private AI that holds your carrier contracts, SOPs, and accessorial schedules in a sovereign knowledge base. "What is our detention rate with carrier X and when does it start" answered from the actual clause, cited. Rate sheets never leave your network. Your competitive crown jewels stay inside the fence. Open to 15 minutes to see the economics for your operation? Brandon Gill Founder, Oction Labs brandon@octionagency.com ``` ### Touch 2: Value / Proof **Subject:** The margin is in the details you cannot find **Body:** ``` 3PL margins live in contract nuances. Detention rules. Accessorial schedules. Claims precedents. If your staff cannot find them in under 30 seconds, you are leaving money on the table. We build systems that query your own contracts and SOPs with cited answers. Custom integrations pull TMS exports into the knowledge base on schedule. The AI does not replace your TMS. It makes everything around it findable. Claims and incident retrieval by lane, carrier, or customer. Better decisions. Better carrier scorecards. Reply if you are facing margin compression or a key ops veteran just gave notice. Brandon Gill Founder, Oction Labs ``` ### Touch 3: Breakup / Soft Close **Subject:** Final follow-up: freight operations **Body:** ``` I have written twice about private AI for logistics knowledge management. No reply typically means the bleed is not acute enough to act. Closing the loop. If customer SLA penalties are mounting from slow answers or you are merging two ops teams after an acquisition, the system exists. Sovereign deployment. Cited answers. Your contracts never touch a third-party model. If not now, perhaps when the next ops veteran leaves. Eight agents. One organism. Brandon Gill Oction Labs ``` --- ## Usage Notes **Cadence:** Send Touch 1, wait 5-7 business days, send Touch 2, wait 5-7 business days, send Touch 3. Stop after Touch 3. **Personalization:** Customize the opening sentence with a specific trigger (recent hire, retirement announcement, merger news, regulatory filing) when possible. Do not use scraped data until verified contact established per lead quarantine rules. **No case studies:** Until first clients close, reference "our measured evaluation" (74% correct retrieval-grounded, 0% fabrication on trap questions) rather than client outcomes. **Pricing anchor (if asked):** Standard from $25,000 CAD + $10,000/month. Enterprise from $200,000 CAD + $20,000/month. Floors, never quote below. **Sender:** Brandon Gill, Founder (brandon@octionagency.com). Alternative: Shamir Cheema, CRO (shamir@octionagency.com).