# Pre-Call Brief: Legal Firms

**Vertical:** Small and mid-size legal practices
**Last updated:** 2026-07-22
**Oction Labs**

---

## 1. Target Buyer Profile

| Attribute | Specification |
|-----------|---------------|
| **Firm size** | 10-100 lawyers |
| **Job titles** | Managing Partner, Firm COO/Executive Director, General Counsel (in-house) |
| **Decision authority** | Managing Partner or partnership vote; COO runs the evaluation |
| **Sub-verticals** | Full-service firms, boutique practices, notarial/conveyancing, corporate legal departments |

---

## 2. Primary Pain Points

1. **Associates burn non-billable hours** searching precedents, past memos, and closing binders for language that already exists.
2. **Hallucinated citations** — the number-one AI fear in the profession; a fake case citation is malpractice.
3. **Knowledge walks out** with every departing partner; matter history and client context evaporates.
4. **Solicitor-client privilege anxiety** — any data sent to external AI vendors creates privilege and confidentiality exposure.

---

## 3. Bleed Calculation Example

~$22,000/month baseline (illustrative starting point).

*Confirm live: How many associates spend how many hours per week on precedent and memo retrieval? What is their billable rate? What is the opportunity cost of that time?*

---

## 4. Key Questions to Ask

1. "How do your associates find precedents and past opinions today?"
2. "What happens to matter knowledge when a partner leaves the firm?"
3. "Has your firm evaluated AI tools? What stopped you from adopting them?"
4. "How do you handle privilege and confidentiality when evaluating new technology?"

---

## 5. Proof Points to Share

- **0% fabrication on trap questions** in our measured test set — the headline for lawyers
- **74% correct retrieval** with grounding vs ~5-10% without
- **Cited answers** — every response references the source document, not invented sources
- **Sovereign deployment** — nothing sent to external AI vendors; privilege analysis stays simple
- **Full audit trail** — supports professional responsibility obligations

---

## 6. Objections Likely to Hear

| Objection | Response |
|-----------|----------|
| "AI makes up cases. We can't risk it." | "Our measured evaluation: 0% fabrication on trap questions. Every answer cites a source document you can verify." |
| "Law society rules don't allow AI." | "They require competence and confidentiality. On-premise deployment with full audit trail meets both." |
| "Partners won't trust it." | "Partners don't have to trust the AI. They verify the cited source — just faster." |

---

## 7. Relevant Case/Example

A 25-lawyer commercial practice has 12 years of memos, opinions, and clause banks. A junior asks, "What language did we use for non-compete enforceability in Ontario technology transactions?" The system returns three relevant precedent clauses with the source memos cited. The lawyer reviews, adapts, and owns the advice — in 5 minutes instead of 2 hours searching binders.

*Illustrative scenario. Not a client result.*

---

## 8. Recommended Pricing Tier

| Firm Profile | Tier | Pricing |
|--------------|------|---------|
| 10-30 lawyers, single office | Standard | $25,000 initial + $10,000/month |
| Regional multi-office, 50+ lawyers, or corporate legal department | Enterprise | $200,000 initial + $20,000/month |

Add-ons to mention: Expertise interview program ($2,000/person) for departing partners.

---

## 9. Talk Track Opening

"Your associates are losing billable hours to precedent hunting. We install a private system that retrieves your own memos and opinions with verified citations — no hallucinated cases, no data sent to external vendors. Can I ask how your lawyers find precedent language today?"

---

## 10. Competitors in This Space

- **Traditional:** Document management systems (iManage, NetDocuments), internal knowledge management teams
- **Legal-specific AI:** Harvey, Casetext, CoCounsel — emphasize on-premise sovereignty vs cloud-hosted
- **Generic AI:** ChatGPT, Claude — emphasize the privilege and fabrication risks they create

---

*eight agents. one organism.*