Practical AI for small professional firms in Texas
I find where your staff’s time leaks, build the AI workflows that fix it, and keep them running safely.
Free 30-minute AI time-leak call
Based in Houston, serving firms across the Texas Triangle (Houston, Dallas–Fort Worth, Austin, San Antonio).
Who it's for
Firms of 2 to 50 people who sell their expertise.
Built for law firms, CPA and bookkeeping firms, and insurance agencies. The best fit is a firm of 8 to 25 people.
A good fit
- An owner or managing partner who can make the decision.
- An office manager or operations lead who will own the tools day to day.
- Microsoft 365 or Google Workspace, and cloud software for your practice.
- You are curious about AI and worried about doing it wrong.
A good time to call
- You lost a staff member, cannot hire, or are paying overtime.
- A seasonal crunch is coming, like tax season or renewals.
- Staff are pasting client information into free AI tools.
- A client asked, "Do you use AI on my file?"
- Your regulator published AI guidance and you have not acted on it yet.
How it works
One free call, then three steps.
Every firm follows the same path. You can stop after any step.
30 minutes
I ask five questions about where your staff's time goes. You leave with one quick win you can do yourself.
One to two weeks
A short survey, a few staff interviews, and a review of your software and where your client data goes. You get your top opportunities ranked by hours saved, a risk snapshot, and a 90-day roadmap with a fixed price for the first build.
Two to six weeks
One to three workflows per build. The scope is written down with acceptance tests before I start. Every build includes a recorded staff training session, a runbook and 30 days of fixes.
Monthly, optional
I monitor the workflows, update them when vendors change things, and meet with you each quarter about what to do next.
Law firms
Less time on intake and first drafts.
A lawyer still reviews everything before it leaves the firm.
What I build
- Intake and conflict-check prep
- First drafts of engagement letters, demand letters and discovery responses
- Deposition and record summaries
- Billing narrative cleanup
The rules I build around
Texas Ethics Opinion 705 and ABA Formal Opinion 512 ask lawyers to understand the AI tool they use, keep client information confidential, verify its output, and supervise it like a nonlawyer assistant. I build each workflow around those four points.
CPA and bookkeeping firms
Stop chasing documents and retyping receipts.
Your staff get the hours back before the next busy season.
What I build
- Document request reminders and organizer intake
- Receipt and bank statement extraction
- Engagement letter generation
- Month-end close checklists
- Client email drafts
The rules I build around
Section 7216 limits where tax return information can go, and the FTC Safeguards Rule requires tax preparers to oversee their vendors, AI vendors included. I use US-hosted business tiers that do not train on your client data.
Insurance agencies
Faster renewals and service replies.
A person at your agency approves each message before it is sent.
What I build
- Renewal outreach
- Certificate of insurance requests
- Quote comparison summaries
- Policy document answers for producers
- Service email triage
The rules I build around
TDI Bulletin B-0003-26, issued in June 2026, covers independent agencies. It expects AI governance, human review before a consequential decision is acted on, and procedures you can produce on request. The assessment shows where your agency stands against it.
Demos
See it work.
Two short walkthroughs, recorded with made-up data.
Law: intake email to engagement letter draft
An intake email arrives. A draft engagement letter is ready for a lawyer to review a few minutes later.
Bookkeeping: receipt to bookkeeping entry
A receipt comes in. The details are read from it and a draft entry is ready for the bookkeeper to approve.
How I handle your data Draft for attorney review
Your client data stays in your systems.
These rules apply to every engagement, and you get them in writing.
- Your data stays in systems you own. I keep no copies on my own devices.
- You own every account and API key. I am invited as a user. My access is removed at the end of the engagement and I confirm that in writing.
- I build with test or made-up data wherever possible.
- I use AI vendors on business tiers that do not train on your data, and only vendors approved for the kind of data involved. Privileged data needs your written approval first.
- I document every automation: what data it touches and where that data goes.
- A person at your firm approves anything before it goes to a client.
- If I suspect any data exposure, I tell you within 24 hours.
Prices
Fixed prices, stated up front.
Start with an assessment sized to your firm. No hourly billing.
Quick Scan
$1,500
For firms of up to 8 people. A half-day working session, your top 5 opportunities and a one-page roadmap.
Readiness Assessment
$4,500
For firms of 8 to 30 people. Staff interviews, your top 10 opportunities ranked by hours saved, a risk snapshot and a 90-day roadmap.
Fixed-price builds
The assessment ends with a fixed price for your first build, in writing, before any work starts. Half of the assessment fee is credited toward a build signed within 30 days.
Founding-client offer
The first three firms get a reduced price on their first build. In return I ask for a case study, a testimonial and a referral call. Ask about it on the call.
About Rick
I'm Rick Gossett.
I have worked in operations and business software for 14 years. I started as a software implementation consultant. I installed systems at customer sites, gathered requirements and trained the people who used them.
After that I managed technical support teams for eight years.
Much of that work has been replacing manual reporting and paperwork with automation. What I learned is that a tool only helps when the people doing the work trust it and use it.
I started Gossett Advisory to do that work for small firms in Texas. You work with me directly, from the first call through the build. I have a business degree in Management Information Systems and Entrepreneurship from Baylor University.
Questions
What firms ask first.
Can't we just use ChatGPT ourselves?
You can, and most firms already do. What is usually missing is anyone measuring the results or controlling where client data goes. The assessment shows you both in about two weeks.
Is our client data safe?
Your data stays in accounts you own. I use business tiers that do not train on your data, a written data-handling policy, and a human review step before anything goes to a client.
Is it expensive?
Every build is quoted against the staff hours it saves, so you can see the payback before you sign. You can also start with the $1,500 Quick Scan.
AI makes mistakes. What then?
It does. That is why every build has a human approval step and acceptance tests. The AI does the drafting and the chasing. Your people keep the judgment.
We don't have time for a project.
The assessment takes three to four hours of your team's time in total. I do the build.
What if you disappear?
You own the accounts, the prompts and the runbook. The monthly care plan is optional, and after the first three months it is month to month.
Why you over a big agency?
I am based in Houston and I visit in person. The work is built for firms your size and priced in fixed numbers. I also know the AI guidance your regulator has published.
Find out where your firm's time is leaking.
Thirty minutes, five questions, and one quick win you can do yourself.
Book a callFree 30-minute AI time-leak call