B:Side Advisors
Industry / Professional services

Professional-services AI, without the hand-wave.

The work is billable hours against a calendar. AI that saves a partner three hours a week and a paralegal ten is not a rounding error. It is margin. Here is how we scope it for firms of 5 to 80 people.

Christopher Myers2026 / Industry brief
Scene / 01
A 14-attorney employment firm · Aurora, CO

It is 4:40 on a Thursday. A partner has spent the morning on a deposition. Her inbox has 214 unread messages, 11 flagged urgent. Her paralegal is on a call with opposing counsel. An intake form came in at 2:11 p.m. that has not been triaged.

Two conflict checks are waiting. One is from a referral that came in over the weekend and is already cooling. The firm's matter management system, which they pay $9,600 a year for, has a well-built intake module. Nobody has had the time to configure it.

This is where AI lives or dies for small firms. Not in the courtroom. In the 4:40 p.m. inbox.

What's actually broken right now.

The problem is rarely the tool. The problem is that nobody has had 90 minutes and the right operator to set it up.

Most small firms already pay for 80% of the software they need. Clio, Smokeball, MyCase, PracticePanther, Karbon, Canopy, Ignition. These are mature products with intake flows, automation, and integrations that would cut admin time meaningfully. The adoption gap is not a product problem. It is a time problem.

The billable hour trap is that the only person with authority to reconfigure the workflow is also the person with the highest opportunity cost of reconfiguring it. A partner who bills at $450 an hour will not spend a Saturday mapping triggers in Clio.

The second trap is the pilot trap. A managing partner subscribes to a new AI tool, runs a pilot for three months, sees modest results, and cancels. The failure is almost always in the glue. The tool does not talk to the calendar. The intake bot does not fill the matter. The doc assembler outputs PDFs the filing system cannot read. AI inside a firm is 10% model and 90% plumbing.

Sources referenced
  • Clio Legal Trends Report 2024 Time utilization data: only ~2.9 billable hours per 8-hour day for the average firm.
  • AICPA PCPS MAP Survey Admin-time and realization rates across small-firm accounting practices.
  • McKinsey Global Institute, "The state of AI" 2025 Adoption vs. measurable ROI gap across professional-services verticals.

Five workflows AI actually changes for a small firm.

These are not hypotheticals. These are the five scopes we quote most often for firms in this industry. Every one is a fixed-price sprint, typically between 4 and 10 weeks.

Workflow 01

Turn inbound inquiries into qualified matters in under 20 minutes.

What we'd build

An intake flow that captures the prospect, runs a conflict check against your matter database, categorizes the request, drafts a first response in your voice, and schedules a paid consult when appropriate. Your intake coordinator reviews and sends, instead of writing from scratch.

Vendors we'd evaluate
  • Clio Grow
  • Lawmatics
  • Karbon
  • Intake1

Vendor-neutral. No reseller margins.

Workflow 02

Compress first-draft document work from hours to minutes.

What we'd build

A document assembly layer over your existing templates (NDAs, engagement letters, demand letters, routine motions). It pulls structured data from the matter, fills the template, flags anything that needs a human judgment call, and leaves a clean first draft for review.

Vendors we'd evaluate
  • Harvey
  • Spellbook
  • Gavel (formerly Documate)
  • Lexis+ AI

Vendor-neutral. No reseller margins.

Workflow 03

Clear time-entry and billing without the Friday marathon.

What we'd build

A passive time-capture layer that watches calendar, email, and document activity, drafts narrative entries tagged to matters, and lets the timekeeper confirm or edit at end of day. Hooks into your existing billing system. Kills the 6:30 p.m. Friday reconstruction.

Vendors we'd evaluate
  • Ajilitee
  • TimeSolv + automation
  • Karbon Time
  • Ignition

Vendor-neutral. No reseller margins.

Workflow 04

Answer client questions faster, without pulling attorneys into triage.

What we'd build

A client-facing response layer trained on your firm's prior correspondence and practice-area knowledge. Routine questions get a drafted reply for the assigned timekeeper to review. Non-routine questions get escalated with matter context pre-attached. Response SLAs improve without adding headcount.

Vendors we'd evaluate
  • Supio
  • ZenDesk + GPT
  • Lawmatics
  • Fireflies for intake

Vendor-neutral. No reseller margins.

Workflow 05

Turn matter outcomes into a real pipeline of referrals.

What we'd build

A post-matter workflow that asks for reviews at the right moment, drafts a LinkedIn post for the timekeeper to approve, and maintains a referral-source dashboard. Most firms have a referral problem. It is usually an operations problem, not a marketing problem.

Vendors we'd evaluate
  • NiceJob
  • EvenUp
  • Karbon CRM
  • Clio Grow + automation layer

Vendor-neutral. No reseller margins.

The small-firm edge over AmLaw 200.

Large firms are paralyzed by conflicts, compliance review cycles, and partner-committee approvals. You have none of that.

An AmLaw 200 firm has 1,400 attorneys, a general counsel, a compliance committee, a technology committee, and a 14-month procurement cycle. Their AI strategy is a 30-page deck that gets debated for a year before a pilot runs.

Your firm does not have any of that. You can decide to try a new intake flow on Monday and ship it by the following Friday. That speed advantage compounds. A firm that runs three AI-assisted workflows by year-end is not 3% more productive. It is structurally different.

Second, the data is tractable. AmLaw firms have thirty years of matter data in six different systems. Yours fits in one. The cost to get AI working well on a small corpus is a fraction of what it is at scale. You are not fighting yesterday's decisions.

Mid-post · 30-minute scoping call

Want a 30-minute scoping call tailored to your practice?

Bring your current stack (Clio, Karbon, Outlook, whatever it is) and the two workflows you'd most want to compress. We'll give you a rough ROI range and a straight answer: audit now, wait a quarter, or skip us.

Three things AI won't fix for a small firm.

Every firm has operational problems AI is genuinely wrong for. Being honest about these is the first sign your advisor is worth paying.

01

If your data is stuck in email.

If matter-critical information lives in a partner's Gmail rather than your matter system, AI is the wrong first move. The work is data hygiene. Do that quarter, then talk to us.

02

If the firm has no consistent templates.

Document assembly depends on reasonably consistent source templates. If every attorney drafts engagement letters from memory, the first sprint is codifying templates, not buying AI.

03

If ethics review is deferred.

Client confidentiality and unauthorized-practice concerns are not AI objections. They are obligations. We won't deploy anything that puts your bar status at risk, and we won't pretend the risk is zero.

How we'd work with your firm.

A Readiness Audit runs two to four weeks. We sit with the managing partner, the operations lead, and your highest-volume timekeeper, map the matter lifecycle end to end, and identify the five workflows with the strongest ROI signal. You walk out with a roadmap, a readiness score, and a prioritized sequence. Fixed price, $3,500 to $8,500.

If the audit points to a clear first sprint, we scope it in writing. Written acceptance tests, a weekly demo, and a fixed band between $12,000 and $45,000. We build inside your environment (not a sandbox), train the team, and hand over a playbook your office manager can maintain.

Past that, Managed is a monthly retainer that keeps the workflows running, measures baseline against usage, and ships one to three new automations per quarter. Month-to-month, 30-day exit, no termination fee. For most small firms, two workflows in Managed cover the retainer in saved admin time alone.

Questions firms actually ask.

The questions operators in this vertical actually ask on the first call.

01How do you handle client confidentiality and privileged communications?
We scope data residency on day one. For most firms this means hosting inference inside your existing stack (Microsoft 365 Copilot, Clio-native AI features, or a private OpenAI / Anthropic deployment where your matter data never leaves your tenant). If your state bar has issued specific AI guidance (as California and Florida have), we map the engagement to it before we build.
02We are on Clio / MyCase / PracticePanther. Does that change the scope?
It changes the tool choices, not the shape of the engagement. The core question is always the same: which workflows inside your matter lifecycle have the highest time cost and the lowest risk to automate. The answer depends on the workflow, not the practice management system.
03Can you guarantee a number of billable hours recovered?
No, and be skeptical of anyone who does. We commit to written acceptance tests at kickoff (e.g. "the intake bot correctly categorizes and drafts a response for 90% of standard matters over a 30-day window"). If we do not pass them, we keep building.
04What about conflicts and engagement letters across multi-office firms?
We handle these as first-class objects in the scope. Conflicts checks almost always require integration with two or three systems: your matter database, your CRM, and often an internal spreadsheet nobody has touched. Part of the sprint is consolidating that, not working around it.
05We are pre-adoption. Where do we even start?
With the audit. Two to four weeks, a fixed price under $8,500, and a written roadmap with ROI ranges. If the audit concludes that AI is the wrong move for you right now, we say that. The audit fee credits 100% against your first sprint within 90 days.
06Do you work with solo practitioners?
Sometimes. Below five total people, the math usually argues for using off-the-shelf tools (Copilot, Harvey for Solo, Clio Duo) without a custom build. We will say so during the scoping call. We turn down engagements that are not a fit.
End of post · Next step

Your firm runs on hours. Get some back.

Thirty minutes, a scoping call, a straight answer. If AI is not the right move for your firm this quarter, we'll tell you. If it is, we'll map out the first sprint with you.

What the 30 minutes delivers
  • 01A short list of AI opportunities specific to your shop.
  • 02A rough ROI range and a sense of which to build first.
  • 03An honest answer: audit now, wait a quarter, or skip us.
Free · 30 minutes · No deck