Connector Corner: Transform Unstructured Documents with Agentic Automation

Unstructured data accounts for up to 90% of enterprise information, yet most remains untapped due to the limits of traditional document processing tools. UiPath’s new Unstructured Complex Documents (UCD) module changes that enabling prompt-based data extraction using Generative AI.

In this session, you'll see how to apply UCD across use cases like contract analysis and multi-table layouts, then connect that data into end-to-end automations using UiPath Agent Builder.

What You’ll Learn

How to use UiPath’s new UCD module to extract data from complex, unstructured documents
How to configure and apply prompt-based extraction to real business scenarios (e.g., contracts, multi-table layouts)
How to use this extracted data in end-to-end agentic workflows with UiPath Agent Builder

1. The UiPath word mark, logos, and robots are registered trademarks owned by UiPath, Inc. and its
affiliates. UiPath (R) is a registered trademark in the United States and several countries across the
globe. See TMEP 906. ©2025 UiPath. All rights reserved.
Transform Unstructured Documents
with Agentic Automation

2. 2
Meet the UiPath Team
Alex Dombroski
Speaker
AI/ML Specialist
Erik Bohman
Speaker
Director AI/ML Sales Engineering
Diana Gray
Host
Senior Community Marketing Manager

3. Setting the Scene
The document and communications
challenge that is limiting growth, and
scalability across the enterprise.
UiPath IXP (Demonstration)
Your introduction to UiPath IXP and the
products core capabilities for data extraction
and processing of communications and
documents.
IXP for Agentic Automation
How IXP fits into the Agentic Automation
framework to enable seamless end-to-end
automations.
IXP & Agents (Demonstration)
Use the IXP capabilities to extract relevant data
from structured, semi-structured, unstructured,
and high complexity documents, trigger Agent
review & analysis.
Content

4. 4
Unstructured and complex data
accounts for most enterprise data
•Businesses that unlock their unstructured data can identify
new insights, extend automation into new areas, and
outperform their competition.
•For decades, businesses have worked mostly with structured data.
•Unstructured data has long been out of reach for automation,
creating inefficiencies and limiting end-to-end process automation.
Charting a path to the data-and-AI-driven enterprise of 2030 - McKinsey & Company
90% Unstructured Data 10% Structured Data

5. UiPath IXP
UiPath IXP (Intelligent Xtraction & Processing), is a multi-modal data
classification and extraction experience that unlocks enterprise data.
UiPath IXP brings together our existing, and category-leading IDP capabilities (Document Understanding and Communications Mining) with a new
prompt-driven capability for unstructured, and high-complexity documents – we call this Generative Extraction for unstructured & complex documents.
UiPath IXP
Multi-modal data classification and extraction that unlocks enterprise data at speed and scale​
Communications
Mining
Generative Extraction for
unstructured & complex documents
Document
Understanding

6. Our vision is to continue to expand modalities
UiPath for enterprise data processing
Documents Emails
Tickets &
Feedbacks
Charts &
diagrams
Chats & Calls
Images &
Video
Multi-modal data extraction capability to provide structured data for analytics and for automations
TODAY TOMORROW

7. Act
Validate extractions if required, initiate
automation and downstream processing
Human in the loop
Validate extraction results if
required or in case of
inaccuracies and exceptions
UiPath Platform
Route extracted data to
downstream systems for
further processing
3
Understand & extract
Extract the information you need
from communications and documents
2
Receive
1
Upload documents and
connect communication channels
✓Multiple languages
✓Various formats
✓Handwriting
✓Signatures
✓Skewed & low-quality scans
✓Checkboxes
✓Tables
✓Long documents (100+ pages)
How IXP works

8. How IXP works
Act
1
Understand & extract
Upload documents and
connect communication channels
Extract the information you need
from communications and documents
Validate extractions if required, initiate
automation and downstream processing
Extract relevant data from
structured and semi-structured
documents
Extract key intent, sentiment and
context data from messages and
communications
Extract relevant data from high
complexity, varied unstructured
documents
✓Multiple languages
✓Various formats
✓Handwriting
✓Signatures
✓Skewed & low-quality scans
✓Checkboxes
✓Tables
✓Long documents (100+ pages)
Human in the loop
Validate extraction results if
required or in case of
inaccuracies and exceptions
UiPath Platform
Route extracted data to
downstream systems for
further processing
2 3
Receive

9. XX4987BBB
Reference: x
Investment Position
Not confirmed16▲
Not confirmed15▼
6,000.0000
Reference: x
Number of Units
19
Reference: x
Market Value
Brokerage Account > Investments
6097.50
Reference: x
Market Value
USD
Example document:
Report including detailed, complex tables, charts and graphs
IXP: understands and extracts the information you need
Spotlight on Unstructured Complex Docs
Extract data, no training required

10. Current alternatives do not
meet the needs of the business
Manual processing (eg Doing Nothing)
Limits business growth and scalability
Labor-intensive processing limits the ability
to scale efficiently and capture market opportunity.
Results in poor customer experience
Complex, unstructured data mandates human
decisioning, slow onboarding, and servicing. This can
result in a poor customer experience.
Increase risk
Higher chance of data input errors, missed information,
and incorrect procedures, increases risk.
Not cost effective or scalable
Long-term and hidden costs
Building an in-house solution requires substantial time and costly
expertise. Third-party GenAI models require AI specialists and
engineers to fine-tune the chosen model to business requirements.
Maintenance, security, and governance
When you build a solution, you are solely responsible for complex AI
model and platform governance. Every system component requires
critical decisions and specialist skills, increasing technical debt and risk.
Talent acquisition and retention
With AI talent in high demand recruitment can be time-consuming and
costly. Talent attrition can also make a system non-performant or non-
viable in the long term
Build your own solution (eg With LLMs)
Simple to demo, hard to deliver and maintain

11. IXP DEMO

12. UiPath journey to Agentic Automation
2018
RPA
UI Automation
Computer Vision
API Integration
Low-code Apps
Orchestration
Governance
Cloud
Discovery
Analytics
Continuous
Testing
Intelligent
Document &
Communication
Processing
Generative AI
Powered
Automation
UiPath LLMs
UiPath GenAI
Activities
UI-based automations
Agents, Robots,
People
2020
API-based
automations
2022
Structured data
understanding
2023
Unstructured data
understanding
2024
GenAI
automations
2025
Agentic
processes
ACTIVITIES
WORKFLOWS
PROCESSES

13. 13

14. Robots
rule-based
Agents
goal-based
KR - 50
JP
Summit -
30
JP RT -
20
JP
Partner -
40
ANZ
Summit -
30

15. 15
Refinement
ExecutionPlanning
Start
End
Tools
Automations
APIs
Other Agents
Escalation
HITL
Conversational
Context
Enterprise
Knowledge
Documentation
Models
Instructions
Start
End
Robot
Human
Robot
Robot
Robot
Decision
Robots
rule-based
Agents
goal-based

16. 18

17. AGENT DEMO

18. We are entering the agentic era, where AI agents, automation, and people combine
to perform more complex autonomous business workflows
Documents/messages are
processed using IDP and these
outputs are sent to an Agent as
inputs to execute some agentic
reasoning/logic
Agent has access to IDP project as
a tool/index to do Q&A (e.g., how
many contracts were processed in
Q4, what was theaverage contract
amount, etc.)
Agent receiving IDP model
output as an input
Agent querying IDP as a
repository to complete a task
Agent selecting to use an IDP
model for a task
Agent has access to IDP models as
tools to process documents/
messages as part of the agent run…
Available now Available soon Available soon
How IXP & Agents work together
Capabilities available now and more coming soon

19. Legal Copy: Es quate in preperu mquaspitatia nobitiu menduntiam non nem facimen estrum faceaturenia voluptist, sum, voluptatiis et quodit aces del imi, quidell aborepello beat quatatius, sam
im el et officiis estrum facepedio quae nonserum, ex et que doluptatius, que cuptureped quidi tentoresto moluptati odionecea nia quis ilit ma videbitesci ra poruntin cum con nost, sedis ex explab
imporem. Hillore peria velis quae et mod quia sed ma volo temossitio que as sinci dem qui occulpa si ommolore susanda nihitis ellamet officiisit volupta temporumqui bero et quis undis eos elique
pario. Et moluptae solore velicil itassi omnihiciis entempor re, voluptiunt occumquam, a verae pre debitat exerum voloriandi alit porehenihit, omnim eum cum venihit enihill enihillit, aut qui am
dollam re duntio tene sunt.