Legal AI Has a Mid-Market Problem

If you run a law firm in Lake Forest, Wilmette, or Evanston with 5 to 20 attorneys, the two most-discussed legal AI vendors are probably not built for you. Harvey was designed for the Am Law 100. Spellbook was designed for contract lawyers who live in Microsoft Word. Neither one quite fits the mid-market practice that handles real estate closings, business disputes, estate plans, and employment matters under one roof. This article puts the two side by side and explains where the gap is, so you can make a faster, better-informed decision.

Harvey Is Built for Enterprise

Harvey is the most well-funded legal AI vendor in the market. The company raised $160 million in a Series F round in late 2025, led by Andreessen Horowitz, pushing its valuation to roughly $8 billion, according to TechCrunch's coverage of the round. More than 142,000 legal professionals across 1,300 organizations use the platform today, including more than 50 Am Law 100 firms. DLA Piper expanded to 5,000 licenses in early 2026.

What Harvey actually does is closer to a research and analysis layer for large, complex matters. You can ask it to analyze hundreds of documents at once, surface precedents across case law, draft complex filings, and run purpose-built agents for end-to-end legal workflows. The platform integrates with legal databases and handles multi-jurisdiction research at a scale that most boutique practices never need. According to Harvey's profile on Wikipedia, it is tailored specifically for leading law firms and Fortune 500 legal departments, not general practitioners.

The pricing reflects that positioning. A detailed Harvey pricing review at LawXY AI puts the platform at roughly $1,200 per lawyer per month, with a minimum of 20 seats and a 12-month commitment. That floor sits around $288,000 per year before any custom training or implementation fees. For a 10-attorney estate and business firm in Lake Forest, that number ends the conversation before it starts. If a vendor quotes you Harvey for a small practice, ask why you need 20 seats and what implementation looks like on your team.

Spellbook Is Built for Contracts

Spellbook takes the opposite approach. It works inside Microsoft Word, which is where most transactional lawyers already spend their day. Rather than replacing your research workflow, Spellbook focuses on making the document in front of you better, faster.

The core capability is AI redlining. Spellbook reads the contract you have open, flags missing clauses, surfaces risky terms, and suggests alternative language, all using Word's native Track Changes so nothing looks foreign to the other side. It benchmarks terms against a library of more than 2,300 contract types by sector and jurisdiction, so when a counterparty pushes unusual indemnification language, you get a fast read on whether the ask is market. The platform also handles multi-document workflows for M&A data rooms and financing packages.

More than 4,500 legal teams now use Spellbook, according to the company. It has SOC 2 Type II certification and runs on a mix of leading AI models, including GPT-5 and Claude, under the hood. Enterprise pricing is no longer publicly listed, but public reporting places it around $350 per user per month with a six-month minimum commitment. If your practice is heavy on commercial contracts, business acquisitions, or real estate documentation, Spellbook fits well. For a closer look at where Evanston attorneys lose billable time, the data suggests contract work and follow-up are the two biggest drains, which is exactly where Spellbook helps. The constraint is that it is a Word-only tool. If your work is research-heavy or litigation-focused, that boundary matters.

The Mid-Market Gap

The gap between these two tools is real and worth naming directly. Harvey prices out most firms under 20 attorneys. Spellbook covers contracts but does not do deep legal research. Neither addresses the full range of what a general practice or boutique litigation firm handles on a given day.

The tools that fill the mid-market space most naturally are CoCounsel, Clio AI (Vincent), and Lexis+ AI. According to a guide on Harvey alternatives for smaller law firms, CoCounsel Core starts around $225 per month and draws on Thomson Reuters' existing legal database, so research returns actual case law rather than results trained on the general web. Lexis+ AI layers AI drafting and research onto an existing Lexis subscription, which many mid-market firms already carry. Both handle the research-heavy work that Spellbook does not.

What to Do This Week

Here is where I think most firms get this wrong: they start with the vendor and work backward to the use case. Start with the use case. Write down the three most time-consuming document and research tasks in your firm. Not all legal work, just the tasks that create the most friction week after week. Then test whether the tools above handle those three things before you sign anything.

For contract-heavy practices doing commercial leasing, business acquisitions, or financing work, Spellbook is worth a seven-day trial. For firms that need deep legal research as part of daily work, CoCounsel or Lexis+ AI is the better spend. For firms already on Clio, Vincent is the logical first test because there is no new platform to learn. If you want to see how Lake Forest firms are running AI without expensive API builds, the short answer is that the consumer tools available today handle more than most attorneys expect before they try.

If you want to test AI-assisted legal drafting before committing to any vendor, here is a starting prompt worth running on a real clause from a current matter:

Claude does not replace a specialized legal AI platform for high-volume contract work. But it is a capable first test for a firm that wants to see whether AI drafting assistance fits before paying for a dedicated product. The firms that gain the most are usually the ones that run a real matter through a free tool first, learn what AI can and cannot do in their practice, and then buy the right product with that knowledge in hand.

The legal AI market is moving fast enough that any pricing comparison is worth rechecking every six months. What does not change is the evaluation framework. Match the tool to the work. If your firm does 30 contract reviews a month, Spellbook is worth the cost. If you do 30 research memos, CoCounsel is the better spend. If you do both at meaningful volume, you may need a combination, or a more general workflow built on top of Claude or another foundation model. The AI services Bace Agency provides to North Shore law firms often start exactly here: figuring out which tools to buy, which to skip, and which workflows to automate before spending on enterprise software.

If you want to work through this decision for your specific practice, take the AI Readiness Quiz at Bace Agency to get a baseline, or schedule a free 30-minute AI audit and we will walk through the tools that make sense for your size and work mix.

Frequently Asked Questions