Smart Start Beyond the Buzz · Session 01 of 05
Golden Gate ALA
April 23, 2026
1-7 jump · navigate

Why you're here.

Open · The Hook
Why You're Here
Leadership said: "Figure out AI."
You're here to figure out where to actually start.
The Cost of Waiting Just Changed
Clients have stopped asking if you're using AI.
They're auditing how you use it, how you govern it, and what results you produce.
Fortune 10 Healthcare Client

To its outside counsel, January 2026:

  • No more manual doc review above 5,000 documents
  • No more manual deposition prep above 1,000 documents
  • No more manual privilege logs above 100 entries
  • AI usage documented in every billing narrative
From my client work, last 90 days. Firm and client names withheld.
Fortune 50 Retailer Client

To its immigration firms, November 2025:

  • AI capability is now a screening criterion
  • Before the RFP is even issued
  • Pre-RFP questionnaire on tools, governance, outcomes
  • 12-month AI roadmap required in response
From my client work, last 90 days. Firm and client names withheld.
"We're exploring AI" is no longer an answer. The burden of proof has moved to the firm.
The Shift
60 minutes. Four things. What AI is · How to assess readiness · Governance essentials · A live sentiment survey that shapes Sessions 2-5.

What AI is. What it isn't.

Foundations · Pillar One
What AI Isn't
  • Not a tool you buy
  • Not a procurement decision
  • Not a one-time project
What AI Is
A capability your firm builds.
The AI Software Stack
LAYER 1 Foundational Models Anthropic · OpenAI · Google
LAYER 2 Software Adding AI Your DMS · Copilot · legal tech
LAYER 3 You and Your Firm the decisions, the work, outcomes
Lock-in isn't at the model layer. It's at the software layer.
Chat vs. Agent · A Preview

Chat

You explain your world every time.

Agent

It already knows your world.

What Matters Most Right Now
  • Client expectations are escalating
  • Competitors are building capability, not buying tools
  • Data governance maturity matters as much as tool adoption
  • Quantified outcomes are becoming table stakes

Assess where you actually stand.

Assess · Pillar Two
Before You Buy Anything

Most AI rollouts fail because they start with technology or policy.

They should start with a survey.

What Surveys Consistently Reveal
  • Staff use AI more than leadership thinks
  • Sentiment is overwhelmingly positive or cautiously optimistic
  • Data security is the universal top concern
  • Attorneys, staff, and admins want very different things
Pre-Session Readiness Snapshot · anonymized mid-size firm
85%
use AI at least weekly in their work
40%
are uncomfortable with the firm's current AI policy
The gap between usage and policy comfort is where the work begins.
Then Audit What You Already Have

AI features you're likely paying for already:

  • Copilot in Microsoft 365
  • Gemini in Google Workspace
  • iManage AI features
  • NetDocuments AI
  • Litera · Harvey · Clio
  • Zoom AI Companion
Audit before you buy.

The STEPS Framework.

Five stages · In order · Always
Today
S
Survey
Ask before you deploy.
Today
T
Technology
Audit, then match tools to needs.
Session 2-4
E
Effort
Train for capability, not attendance.
Session 5
P
Policy
Enable, don't just restrict.
Session 5
S
Success
Measure what matters. Iterate.

Today we focus on the first two. Sessions 2 through 5 live inside the rest. Finance · Marketing and BD · HR · IT.

Why This Order Matters

Most failed rollouts invert it.

They start with Policy ("here's what you can't do"), then Technology ("here's the tool, figure it out"), and only then Survey ("what are you doing with it?"). People get restricted before they understand. They get tools before anyone asked what they needed. Adoption stalls.

Signals It's Working
  • People ask questions during training
  • Teams share use cases with each other
  • Usage grows after the first month
  • Policy feels helpful, not restrictive
  • Requests come in for additional tools

Govern it, or it governs you.

Govern · Pillar Three
Policy That Enables · Not Just Restricts

Most firm AI policies read like legal documents.

That's why they get ignored.

Useful policy answers three questions

  1. When is AI appropriate?
  2. What quality standards apply?
  3. Who do I ask when I'm not sure?
A Scenario You've Already Seen
Client joins your Zoom.
Their AI notetaker joins with them.
Now what?

Firms with a policy handle this in 30 seconds.

Firms without one fumble it every time.

Four Risk Categories You Need to Name

Until you name them, you can't govern them.

Privilege Waiver

Confidentiality

Accuracy of Output

Discovery Exposure

Model Rule 1.1
Attorneys have an affirmative duty of technological competence.
This isn't optional.
Reference: NYC Bar Formal Opinion 2025-6
Who Owns AI in Your Firm?

If no one owns it, the gap lands on operations.

Which, for most of you, means it lands on your desk.

A small working group works.

IT Operations Marketing One Practice Group Rep One Named Owner

Your turn.

Live Sentiment Survey
Live Sentiment Survey

What does this room think?

Your answers shape Sessions 2 through 5.
Launch Poll
What We're Asking
Question 1
How often do you personally use AI tools in your work?
Question 2
What's your top concern about AI in the firm?
Question 3
Where would you want AI to help most in your role?
Question 4
How confident are you in your firm's current AI policy?
Question 5
One word: how do you feel about AI right now?
After The Poll
Pull two or three sharp threads live. Full readout goes to registrants.

What's next.

Series Roadmap + One Action
Four Role-Specific Sessions Follow
30 min
One Thing to Do This Week
Put it on the calendar with your IT lead.
Ask one question.
"What AI features are we already paying for that we're not using?"
That's the first page of your working group brief. Everything else flows from this conversation.
Thank you. Questions? Drop them in the chat. Spencer X. Smith · AmpliphI · spencerXsmith.com