Session 2 // AI and Practice Management

Session 2:

AI & Practice Management

Reflections on AI use in the IP profession - which parts of the work may

be handled by AI, and which parts will continue to require experience

and human expertise in the near future.

Ezgi

Baklacı

Gülkokar

Moroğlu

Arseven, Türkiye

AI ADOPTION IN IP PRACTICE

www.morogluarseven.com

April, 2026

E Z G I B A K L A C I G Ü L K O K A R , L L . M .

P A R T N E R

Opening and

Introduction

Inevitable, and

Already Here

Openin

AI is Already Here:

•

As of 2026, AI is integrated into all major
aspects of IP practice worldwide

•

Thomson Reuters survey of

1,500+

professionals in 27

countries

confirms

is now an integral part of the legal
profession

•

40% of organizations

say AI is being

adopted in their practice

•

IP prosecution tasks are well-suited to AI:
structured, repetitive, and rule-based

AI Across IP Services

AI Use in IP Practice

Patent Services:

•

AI-powered prior art search across patents and academic literature

•

Claim drafting and construction assistance

•

Freedom to operate (FTO) analysis

•

Patent portfolio management and monitoring

Workflow Automation:

•

Filing:

Automated trademark application preparation and submission

•

Deadlines:

AI-driven deadline tracking and status reports

•

Responses:

Office action response drafting with AI assistance

•

Evidence:

Digital evidence capture and authentication tools

Online Infringement:

•

Trademark monitoring.

•

Image and text recognition to detect counterfeit listings

•

Ability to scan millions of listings across platforms

•

Enforcement: take-down notices

Trademark Services:

•

Clearance searches across classes and jurisdictions

•

Watch services and infringement monitoring

Legal AI Assistance Tools

AI Use in IP Practice

Why AI is Inevitable

Advantages

for IP Work

Advantages of AI

Speed

•

AI processes documents in minutes vs. hours of manual billable work

•

Competitors using AI may undercut on price and turnaround

Competition

Client Expectations

•

Clients will seek experts for complex, high-judgment matters

Costs & Fees

•

Change in fee structures??

•

Predictability

•

Critical thinking and legal judgement

How AI Supports Legal Work Today

AI in IP Practice

Document Review & Summari

zation

Quickly recalls case history from prior correspondence .
No manual re-reading.

The more we structure our workflows & define our standards, the more AI allows us to work on strategies that require a

lawyer's judgment.

Tabular Review Tool

Extracts specific data points from multiple documents simultaneously
puts data into a structured, comparative tables.

Dedicated AI project per matter; database fed with articles, opinions
& court decisions for targeted research.

Matter Projects & Legal Research

Automated Document

Drafting

Create AI workflows .
LInclude case specific details for solid first drafts .

Predefined rules and fallback positions applied consistently by AI
across every negotiation.

Playbooks

K E Y TA K E AWAY

Risks & Disadvantages

What Could Go Wrong?

Risk &

Disadvantages

Hallucination:

•

AI can fabricate irrelevant citations.

•

Generates technically plausible but legally incorrect claim language

•

Risk of professional misconduct.

•

All AI output must be verified before filing or sending

Change in Billing

Structure

•

AI will lower the

time

spent on

some

tasks.

•

Clients

may question

the billable

hour

fee

structure

Threat to

Current

Jobs:

•

AI automizes the tasks of junior associates, paralegals, and support staff

•

Traditional IP career paths are being compressed

De-skilling Risk:

•

Cognitive and Skill Erosion

•

Overreliance may reduce lawyers' deep legal reading

•

Less developed legal instincts

Legal and Ethical Obligations

Risk &

Disadvantages

Confidentiality:

•

Most critical risk: be aware of each AI tool's data and privacy standards

•

Know if your vendor uses client data to train models

•

Enterprise agreements offer stronger protections than consumer products

•

Shadow AI

•

Supervising attorney is responsible for output

•

No AI output should leave the office without human review

Compliance

with Legal Expectatioons:

•

Bar guidance is evolving.

•

excuse

for

hallusination

or AI error.

•

Updated engagement letters.

Competence:

•

Duty of competence includes technological competence (ABA Model
Rules)

•

Lawyers must understand the tools they use, including AI

Professional Responsibility :

Responsible AI

Adoption

Framework & Closing

Possible

Framework:

•

Verify everything. No AI output leaves without human review

•

Check security compliance

•

Update engagement letters to address AI use and confidentiality

•

Train the team. Tools are only as good as the practitioners using them

AI Native

Firms

•

AI-

Native Law Firms: built from ground up with AI at core of

operations

Smaller teams with fewer traditional hierarchy layers

•

Some AI

native firms are being invested in by major law firms

Still early to assess long-term viability of this model

Trends

Framework & Closing

Embrace the Wave

Closing Thoughts

Fax

Internet

Laptops

Mobile

Stay Informed

Watch developments closely. Observers will be

left behind. Commit to continuous upskilling.

Integrate Thoughtfully

Embed AI as multiplier of your expertise. It is not

a replacement of expertise.

Preserve What Defines Us

Judgment, creativity, and ethical standards

remain the irreplaceable core of the profession.

The professionals who will thrive are those who embrace change thoughtfully while preserving the judgment and ethics that def

ine

our profession.

April, 2026

w w w . m o r o g l u a r s e v e n . c o m

AI for an IP Law Firm

Real-w orld im plem en tation in sights for IP profession als

Control the FOMO / Hype

Tools, Not Substitutes

AI agen ts en han ce efficien cy — they don 't replace
profession al judgm en t or legal strategy.

Don't Confuse the Two

Learn in g a tool ≠ learn in g the law .

Real Adoption Costs

Train in g an d chan ge costs m ay n ot justify short-term
adoption .

Efficiency, Not Strategy

AI accelerates execution — strategy rem ain s hum an .

Options for an IP Firm

General AI Assistants

Co-pilot-style chat in tegrated in to tools like MS Word — easy to adopt,
m in im al setup.

Specialized AI Tools

• Pa ten ts: Patlytics, Edge, Solve

In telligen ce, DeepIP

• Tr a d em a r k s: Markify,

Tradem arkNow , Clarivate RiskMark

• Con tr a cts/Cop yr igh t: Harvey,

CoCoun sel, Ivo AI, Con tractPodAi

• Wor k flow: Clio, Litera, Zapier

Key

Ben efi ts

• Easy to adopt, m in im al setup
• Draftin g, sum m arization

quick

research

Considerations

• Data & privacy con cern s
• Mitigated via en terprise-grade

subscription s

No sin gle m odel fits all — iden tify task-specific tools.

Specialized AI Tools

Recommended for Operations:

Bil ling

Tim ekeepin g
autom ation , draft
description s, guidelin e
flags.

Docketing

Auto-create dockets;
extract deadlin es from
docum en ts.

Workflow

Task assign m en t,
trackin g, real-tim e
visibility, bottlen ecks.

Administration

Gen eral support
fun ction s.

Using AI for Patent Drafting

Optin g out isn 't r eally an option — m ultiple tools available, each w ith th eir ow n pr os an d con s.

Advantages

Implementation

Common AI Drafting Issues

Redundancy / Verbosity

"...a processor configured to process data, wherein the processor processes the data by performing data processing..." —Circular an d
redun dan t lan guage; Adds n o legal value

Hallucinated Elements

"...a blockchain verification module..." — Not presen t in specification ; Violates w ritten description requirem en t

Over-Generalization

"...configured to perform any suitable analysis..."

Too broad

�

m ay fail en ablem en t, Exam in er m ay object:

un due breadth

Antecedent Inconsistency

"user input" ≠ "input data"

— frequen t AI issue; m ust be con sisten t

Improper Functional Claiming

"means for analyzing data using AI" — triggers m ean s-plus-fun ction un in ten tion ally

Patent Drafting Workflow

GǾMŅP FÕMÒÖ Î ÞPÕÒŌÑ RÒPOŎÞP ĖH

Draft broad in depen den t claim s

coverin g the in ven tion . Then add

depen den t claim s that specify

com pon en ts, ran ges, con cen tration s,

con dition s, an d other refin em en ts.

Seek suggestion s on claim s from AI

Fin alize Claim s

Prepare Draft Drawings and/or Prepare

Term Bank

Iden tify the list of elem en ts, techn ical term s,

chem ical or biological com pon en ts, an d

abbreviation s used throughout the

disclosure.

Create a Draft Strategy/Outline

Set the sequen ce for the Figures /

Em bodim en ts an d specification section s so

the disclosure follow s a con sisten t draftin g

structure.

Select a Template for the Specification

Specification

Create a tem plate tailored to the specific

jurisdiction , clien t, an d/or techn ical dom ain

before draftin g begin s.

GǾMŅP Ĭ ÑŃPÒŎŌŒ RÒPO Ĭ ŐÑŃÒŅÒŃ

ĨǾŎÖ ŐPŒ

Use prom pts to keep term in ology con sisten t

w ith the term ban k or claim s, an d to rely

on ly on in form ation from the disclosure

docum en t to preven t hallucin ation s.

Best Practices for AI-Assisted Patent Drafting

Preparation is Key

• Have a draftin g plan /outlin e

◦

Claim scope

◦

Flow of draft an d section s

◦

Poin ts to be covered in each section

• Iden tify specific poin ts for research:

◦

Stabilizin g agen ts w hich are relevan t for tyre
com position s

◦

Iden tify the characterization m ethods used for
m echan ical properties of a com posite

Before Using AI

• First draft the claim s yourself w ithout an y AI in put to

preven ts bias (review bias, an chorin g bias)

• Usin g AI as startin g poin t m ay com pletely m iss the

strategy

Prompting Best Practices

Back gr ound Sect i on

In stead of "Dra ft a ba ckground section", use:

"Draft a Backgroun d section lim ited to 200-250 w ords.
Neutral, factual lan guage. No specific problem s
highlighted."

Detailed Description — Software

In stead of "Dra ft deta iled description of the Dra wings", use:

"Draft detailed description for [figure/embodiment],
in troducin g each elem en t an d describin g in teraction s
en ablin g claim s [numbers]."

Claims — Software

"Draft three con sisten t in depen den t claim s: (1)
apparatus, (2) m ethod, (3) n on -tran sitory CRM. Form al
paten t lan guage, con sisten t term in ology, avoid
un n ecessary lim itation s."

Cl ai ms — Li fe Sci en ces

In stead of "Dra ft cla ims for the invention", use:

"Draft U.S. m ethod of treatm en t claim s for [gestation al
diabetes], m ax 20 claim s, at least on e in depen den t.
Om it dosage in in depen den t claim — defin e in
depen den ts."

Tips to Reduce AI Hallucinations

Min im izin g AI-gen erated in accuracies is crucial for high-quality paten t draftin g.

Leverage Specialized AI Tools

Prioritize AI tools specifically design ed for
paten t draftin g over gen eral-purpose AI
assistan ts (e.g., specialized paten t tools vs.
ChatGPT). These tools are often train ed on
relevan t datasets, reducin g the likelihood of
gen eratin g irrelevan t or in correct
in form ation .

Provide Comprehensive Input

Upload detailed techn ical disclosures an d
specification s. Avoid relyin g solely on an
In ven tion Disclosure Form (IDF) w hich m ay
lack the gran ularity n eeded for AI to
gen erate accurate an d com plete con ten t,
preven tin g factual errors.

Utilize Specific Revision Prompts

After in itial drafts, use precise prom pts to
correct or exclude iden tified problem atic
con ten t. This guides the AI to refin e its
output based on your specific requirem en ts
an d observation s.

Example:
"Do not include ‘blockchain verification
module’ or any reference to ‘quantum
computing’ in the specification."

"Revise the specification to remove the
advantages associated with ‘proprietary
algorithms’."

Key Change

Allocate sufficien t tim e for review an d be m en tally prepared that the review tim e m ay be sign ifican tly lon ger than
con ven tion .

Watch out for “review bias”

Key Takeaways - Patents

Do not give AI tools to juniors and young professionals without supervision.

Define claim scope and fallback positions.

Understand trade secret vs. patent disclosure strategy.

Plan divisionals, continuations, embodiment disclosure.

Verify all AI-generated content.

Thank You

Question s? We'd love to con n ect.

+91 11 40200200 |

em ail@obhan m ason .com

w w w .obhan m ason .com

IP OFFICES, CLIENTS, AND OURSELVES

AI’s Impacts:

Brett Slaney

CPST Intellectual Property

April 20, 2026

AI and the IP Offices - general



Most Offices have an “AI agenda”



Searching (including image searching)



Classification



Office Action Generation (more than boilerplate selection)



Translation



Transcription (e.g., EPO oral proceedings)



Important: will we have access to the same tools?



Future: will we see fully autonomous AI examiners?

AI and the IP Offices - issues



AI-generated prior art – IDSs, disclosure dates, are they enabled?



Disclosing use of AI to patent office – in limited circumstances



Existing duties of candor and good faith sufficient (for now)



Possible - If AI significantly contributed to the conception of the invention, it

may be necessary to disclose this to ensure that any human inventor's

contribution is clearly identifiable and qualifies as an inventor



https://www.federalregister.gov/documents/2024/04/11/2024-07629/guidance-

on-use-of-artificial-intelligence-based-tools-in-practice-before-the-united-states-

patent

AI and the IP Offices



FICPI Resolution from Naples, October 2025: “Human-centric use of AI in the IP

System”



WELCOMING the idea of a human-centric approach to the use of AI, where it is

a tool to assist a human being;



ACKNOWLEDGING that AI used responsibly may help in the preparation of

applications for IP rights, in the examination of IP rights, and in arriving at final

decisions on IP rights with high quality and in a very efficient way;



URGES IP Offices, administrative bodies and Courts to be transparent about

their use of AI in conducting searches and rendering opinions and decisions;



FURTHER URGES IP Offices and IP courts to use AI responsibly in a human-

centric manner, such that substantive decisions are always made by at least one

competent human examiner or human judge.

AI and our Clients



FICPI Resolution from Naples, October 2025: “Human-centric use

of AI in the IP System”



URGES IP Attorneys to be transparent with their clients about their

use of AI in providing services to them;



Some clients want us to use AI



Some clients expressly forbid us from using AI



Many clients appreciate that AI should fit somewhere in the

workflow



Have the discussion before a decision is made for you!

AI and our Clients - example



Begin with an invention disclosure (prepare or get from client) –

traditional/AI



If necessary, get background for technical field -



Draft a few claims and a few diagrams (+ terms defined) -

traditional



Have AI tool prepare full specification, additional drawings, additional dependent

claims or claim types -



Review and revise –

traditional



Use AI proofing to catch manual errors -



If appropriate some post analysis – e.g., prior art, detectability, freedom to

operate, other? –

traditional + AI



Conclusion: cut out tedious drafting without losing control or key inputs



But…do we split these tasks with our clients?

The Future: Impact on Profession



Balancing costs of tools with reduction in time or overhead – find the sweet spot



But, if we are replacing the role of trainee with AI, how do we train new professionals?



How do they become competent enough to do the hard stuff?



Can AI help with training?



Will the profession increase or decrease in size?



Easier to draft – more filings, more prosecution



Easier to draft – do not need as many professionals



Will companies be able to reduce reliance on firms and/or in-house attorneys?



Is this desirable?



Compare to impact of other leaps in technology



More patents, more prior art (thickets) … but better search tools



If patents become less expensive and more accessible does amount of work increase, decrease,

stay the same?

AI’s Impacts:

IP Offices,

Clients, and Ourselves

Thank You!

Questions and Discussion

Brett Slaney
April 20, 2026

Seminar Cocktail Reception