Does Billables.ai read my email bodies?

Yes. Billables.ai's core feature is AI-generated time-entry narratives, which requires reading email content and document text to summarize the work. Their public materials are transparent about this — it's the product. If you don't want email bodies or document contents ingested, you need a metadata-only tool instead.

Is using an LLM on privileged communications a bar-ethics problem?

It's not settled, and state bars are actively debating it. ABA Formal Opinion 512 (2024) frames generative AI use as requiring competent supervision and informed client consent in specific contexts. The safer posture for a privilege-paranoid practice is to keep the content on your device and let AI operate only on metadata — which is what ClaimHour does. Check your jurisdiction's most recent opinion before adopting any content-ingesting tool.

What's the cost comparison?

Billables.ai public pricing as of April 2026 starts at $50/user/month with higher tiers for multi-matter and integrations. ClaimHour Starter is $29/mo, Pro is $59/mo. Comparable tier-for-tier once you factor the scope difference.

Can ClaimHour still draft narratives for me?

It drafts suggested narratives from metadata (e.g. '35-min call with opposing counsel Re: [matter]'). You finalize or rewrite before the entry becomes billable. We don't use LLMs trained on your content to produce billing descriptions — that's a deliberate product choice, not a limitation.

Alternative guide · Updated April 2026

Billables.ai alternative: metadata-only capture for privilege-sensitive solos

Billables.ai is a strong product for lawyers who are comfortable sending email and document contents through a generative-AI pipeline. If you aren't — because you handle criminal defense, family law, immigration, or any practice where "AI summarized my privileged comms" is a phrase you never want to say to a client — here's what to use instead.

TL;DR

Billables.ai reads the content of your emails and documents to draft AI narratives for billable entries. ClaimHour never reads content — only metadata (call duration + counterparty, email counts, document edit time). Both solve the same "I forgot to log it" problem. ClaimHour's tradeoff: you write the narrative (we draft a stub), we don't write it for you from the underlying text. For privilege-paranoid practices that's a feature, not a limitation.

What Billables.ai actually does

Billables.ai connects to your email (Gmail or Outlook via OAuth), your calendar, and optionally a document store. It ingests the message and document contents, passes them through an LLM, and produces a proposed time entry: a duration estimate, a matter guess, and a drafted narrative like "Reviewed opposing counsel's motion in limine and drafted response brief (1.2)." You approve, edit, or reject. Approved entries export to your billing system.

The product is well-built and solves a real problem. Our disagreement isn't with the execution — it's with the data model. Reading email bodies and document contents is a strong signal, and it's also a big trust ask for attorneys who work in practice areas where client communications are privileged by default and clients expect them to stay on the attorney's machine.

The privilege question, in plain English

American Bar Association Formal Opinion 512 (2024) discusses generative AI and attorney duties. The headline: a lawyer using AI to process client information has duties of competence, confidentiality, and informed consent, and must evaluate the AI provider's data-handling practices against those duties. Several state bars — California, New York, Florida — have issued similar guidance.

None of those opinions bans AI use. They raise the due-diligence bar. If you use Billables.ai, you need to verify their data-handling, get informed client consent where applicable (Formal Opinion 512 is case-specific on when consent is required), and document your reasoning. That's workable for many firms. It is extra overhead that some solos simply do not want to carry.

Metadata-only capture sidesteps most of the question. The duration of a phone call, the subject line of an email, the number of minutes spent editing a Word document — these are facts about your work, not about your client's communications. They're what the old pen-and-paper timesheet captured. An LLM summarizing metadata is categorically different from an LLM summarizing a privileged letter.

Feature-by-feature comparison

Feature	Billables.ai	ClaimHour
Email ingestion	Full body + subject + attachments	Counts only (sent/received per contact per day) + subject keywords
Document capture	Full content for AI narrative	Edit-time bursts only — no content read
Phone calls	Not a primary signal	Metadata (duration, counterparty, direction)
Data residency	Cloud — content sent to their AI pipeline	Local-first — metadata lives on device; only approved entries sync
AI narrative drafting	Yes — from content	Stub only ("Call w/ [contact] Re: [matter]") — you write the narrative
Starting price	$50/user/mo	$29/user/mo
Exports	CSV, QuickBooks, integrations	QuickBooks IIF, LawPay, FreshBooks, CSV

Who Billables.ai is genuinely right for

Transactional practice areas where bodies aren't typically privileged or sensitive (most corporate, real estate, estate planning routine matters).
Firms willing to do the AI-vendor due diligence and log it for bar-compliance purposes.
Attorneys who hate writing entry narratives and value AI-drafted descriptions more than metadata-only capture.
Practices with a general counsel or compliance officer who already vets SaaS tools against privilege and confidentiality rules.

Who ClaimHour is a better fit for

Criminal defense — where anything you write down about a client's statement could conceivably be subpoena-adjacent.
Family law — where emails from clients often include financial, custody, or health-related information.
Immigration — where client communications may touch asylum claims or undocumented status.
Solos who just don't want AI anywhere near their client communications, regardless of the vendor's promises.

None of those practice areas is incompatible with AI in principle. ClaimHour's position is simpler: a tool that never sees content cannot accidentally leak content. The smaller surface area is the pitch.

The trade-off, stated plainly

ClaimHour asks you to do 15–30 seconds of narrative writing per captured entry. A typical solo has 6–10 captured entries per day, so that's about 2–4 minutes of typing a day. Billables.ai writes those narratives for you. If saving those minutes is worth the AI-ingestion trade-off, Billables.ai is a fair pick. If it isn't, ClaimHour is narrower on purpose.

How to evaluate

Read ABA Formal Opinion 512 and your jurisdiction's most recent AI guidance. ~30 minutes.
Read both vendors' privacy and data-handling policies. Make note of: where data is stored, how long, whether the AI provider retains it, and whether model training on your data is opt-in or opt-out.
Run a 14-day trial of one. Track whether the AI narratives are accurate enough to post as-is or whether you're rewriting them anyway (if you're rewriting, the AI isn't saving time).
Decide on privilege posture, not on feature velocity.

How ClaimHour fits

We built ClaimHour for the solo who read about Billables.ai, liked the idea, and then hit "wait, it reads my client emails?" and closed the tab. Join the waitlist for early access. You keep your email client, your billing tool, and your privilege posture; we add the capture layer you've been missing.

Get early access