# Marcus Bellfield, Managing Partner at Bellfield & Cho — read of Lawfirm AI, May 26 2026

> 14 years plaintiff-side, currently running 8 attorneys across personal injury and employment in Phoenix. I am the intake bottleneck. I know it. My paralegal knows it.

## How I got here

Bad Monday. Three voicemail leads went cold while I was stuck in a four-hour deposition. Got home, put Zoe to bed, opened my laptop at 9pm, and searched "law firm intake automation" out of frustration. This was page two of results. Clicked because the meta description said "calls that come at 9pm get answered." That is literally the problem I had six hours ago.

## What I clicked first

The hero pulled me in. "Solo lawyers, stop losing prospects to the night shift" — that line is doing real work. It names the moment exactly. I have lost leads to the night shift. I know what it feels like to see a 9:43pm voicemail from someone who already called three other firms by 8am the next morning.

But then I hit "Calls that come at 9pm get answered, conflicts checked, and retainers drafted while you sleep." I slowed down here. Conflicts checked how? Against what database? If this thing runs a conflict check wrong and I onboard a client I shouldn't, that is a bar complaint, not a billing error. This sentence is doing a lot of heavy lifting with zero explanation of the mechanism.

## Where I paused

The proof section. Three case studies, real-sounding numbers: "Mid-market firm (12 attorneys): 40% intake time reduction in 4 weeks." "Recovered 280 billable hours/year." "Flagged 8 missed renewal dates in backlog." These are good numbers. The 280 billable hours one I can do math on. At $250/hour associate rate that is $70,000/year. For $199/mo. That math works on its face.

Then I scrolled lower and found this: "Honest disclosure: we don't have live customers on this idea yet."

I had to read that twice. So where did those case studies come from? They are presented as real results from real firms. Then the page says there are no customers. That is not a footnote. That is a contradiction sitting in the same scroll. I do not know how to reconcile those two things and the page does not explain it.

## What I distrusted

"99% Accuracy on standard clauses." That number is sitting in the hero with no methodology, no definition of "standard clause," no testing context. I have seen Harvey, I have seen CoCounsel, I have seen the Clio AI stuff. None of them lead with a single accuracy percentage like it is settled fact. That number made me more suspicious, not less.

Also: the whole framing shifts partway down the page. Above the fold it is a product. Below the fold it is a "dossier" you can "unlock for $5" and "adopt the build for $99-$199." There is a score: "74/100 Adoptability." A "Fermi" estimate of "-$22,000 Year-1 take-home." A "1 in 8 meaningful success odds."

Wait. Is this a product I can subscribe to, or a business-idea kit I can buy to BUILD this product? Because those are completely different things and I cannot tell. If you are selling me the idea of building this company, say that at the top. If you are selling me the software, get rid of the Fermi odds section. Right now it reads like a used car lot where half the cars are for sale and half are on a poster about starting a used car lot.

## What would convince me

I want one video — not the 30-second explainer teaser, an actual screen recording — of the intake flow running on a real call. Show me a caller saying "I was in a car accident last Thursday" and show me the AI transcribing it, classifying it as PI, flagging urgency, and dropping a summary into a Clio matter. That would take five minutes to record and would answer more questions than all the copy on this page.

On the conflict check claim specifically: tell me how it works. Does it query a database I populate, or does it try to cross-reference something automatically? That one sentence is either the most valuable feature on the page or a liability bomb. I cannot tell which.

## What I'd ask in an email reply

1. The page says "conflicts checked" as part of the overnight intake flow. Can you walk me through how the conflict check actually works, and what happens if it returns a false negative?

2. Those three case studies — the 12-attorney firm, the solo-plus-3, the in-house team — are those real clients or are those projected scenarios from the Fermi modeling? The honest disclosure at the bottom says there are no live customers yet, but the case studies read like reported results.

3. What does the actual handoff look like the next morning? I wake up, there is a "qualified lead in my queue" — is that an email, a Clio task, a dashboard I log into? What does 7am actually look like for me?

## Verdict: on-the-fence

The problem is real, the night-shift framing is sharp, and $199/mo is not a serious objection if the intake hours claim is accurate. But the case-study-versus-no-live-customers contradiction is a trust problem, and the page identity crisis (product vs. idea kit) means I am not sure what I would even be signing up for.

---
*Memo by skeptic persona, generated 2026-05-26. Studio breaks own self-grading loop.*