Digital Workforce Services Oyj · OMXHEX:DWF
Digital Workforce Services Oyj — UiPath Event (June 2025)
43 slides from Digital Workforce Services Oyj's conference presentation, Jun 5, 2025
Slide 1: Key Facts at a Glance
Key Facts at Glance Highest Strong Industry Scores in Expertise Vision Innovation # Clients 9 Offices 200+ Process automation Healthcare Public sector Finance Manufacturing specialists 250 5000+ Software Pricing flexibility and assets and IP transparency Large customers globally Processes automated
Slide 2: We are a business automation partner
We are a business automation partner We take our customers business process, and we digitalise the data and orchestrate it through the process, we embed agents to automate work and we modernize the Why orchestrate & user experience by building end user interfaces. automate processes? You cut down Yo
Slide 3: Building AI Agent Programs
Building AI Agent Programs Digital Workforce Key Design Principles For Self-funding automation programs WHY DIGITAL WORKFORCE • Trusted by 250+ enterprise customers for over 10 years • Built the largest automation programs globally Modular, Scalable Pilot, Change Management • 200+ AI & automation ex
Slide 4: Challenges to drive AI Agent adoption
Challenges to drive AI Agent adoption • Understand where to start, use cases • Understanding best practices to build the first agent into production and then scale agents • How to embed & orchestrate agents into processes • Talent is scarce • Different approaches with all have pros and cons • Cultur
Slide 5: ROM Learnings, and what is missing to scale Agents
ROM* Learnings, and what is missing to scale Agents ROM Learnings, remodel for Agents What is missing to scale agents ▪ The ROM brings clear governance and ownership with defined ▪ Data governance & LLM model oversight: RPA are deterministic flows, AI roles and responsibilities agents (can) work wit
Slide 6: AI Agent Accelerator (AAA) Program
AI Agent Accelerator (AAA) Program Agent Operating Model (5-10x ROI) Key Design Principles Business Value Delivered Strategic Alignment & Governance 100-Day Ramp Up (500K€ value) Pilot delivered Modular, Scalable (1-3 FTE saved) Technology Architecture Month 0-1 Month 3 Month 5 Month 12 1. Start sma
Slide 7: AI Agent Accelerator Program
AI Agent Accelerator Program 100-Day Agent Operating Discovery 2-3 Ramp-Up Model Assessment weeks 500K€ value by deploying the first wave 5-10x ROI in 12 months of agents & Centre of Expertise Framework to Operationalize Agents ✓ Build Center of Expertise and Strategy ✓ Significant business impact A
Slide 8: Digital Workforce 100-Day Ramp-Up
Digital Workforce 100-Day Ramp-Up Objectives of the 100-Day Ramp Up: Internal Agent capability & value delivered ▪ 500K€ value Implement first wave of agents (2-5-10-20 agents) ▪ Agent Strategy & Roadmap built Define strategy & roadmap ▪ Centre of Expertise built Responsibilities, core team trained
Slide 9: Agent Operating Model
Agent Operating Model Operations Delivery Team Head of Automation / Chief Agent Officer Delivery Team Operations Teams Project Leader Project Leader Teams Designer Designer Process SME Developer Developer Process SME Testers Process Analyst Head of COE Delivery Head of COE Delivery Process Analyst T
Slide 10: Agentic AI: Identifying a suitable use case
Agentic AI Identifying a suitable use case Niko Lehtonen A I Solution A rchit ect , D ig it al W orkf orc e [email protected] 10 Digital Workforce Services
Slide 11: Target identification
Target identification 11 Digital Workforce Services
Slide 12: Target Identification 1/3
Target Identification 1/3 Tangible Business- Clearly defined Business Value critical E2E process The bigger the Worthy of The agentic target, the C-suite system can bigger the oversight handle the impact full process 12 Digital Workforce Services
Slide 13: Target Identification 2/3
Target Identification 2/3 Previously impossible Requires regularly Requires data or unfeasible to scheduled or analysis, forecasting automate continuous processing or value prediction New technology brings Agents can work LLM-based solutions new opportunities non-stop or whenever are great at provid
Slide 14: Target Identification 3/3
Target Identification 3/3 Involves structured Involves more than Involves multiple and unstructured data one app or system data sources Agents can Usually a sign of a Typically very time- handle large more complex process consuming for amounts of data humans 14 Digital Workforce Services
Slide 15: Feasibility assessment
Feasibility assessment 15 Digital Workforce Services
Slide 16: Feasibility assessment 1/2
Feasibility assessment 1/2 Needs data Process length Has critical Business case structuring and complexity decision points Savings vs. Agents can High impact Agents' insights project costs really show potential, many can be used to and running their value here opportunities support costs for agents
Slide 17: Feasibility assessment 2/2
Feasibility assessment 2/2 Unclear process Overwriting data Customer-facing? flow required Difficult to describe Needs to be done Can get very to the agent carefully complex to implement 17 Digital Workforce Services
Slide 18: Agentic system example
Agentic system example 18 Digital Workforce Services
Slide 19: Agentic system example: Loan application processing
Agentic system example Loan application processing Customer fills Customer Customer loan contacted contacted Tool use application Tool use Application details Additional CRM Information Recommen- checked (e.g. loan details updated NO analysed dation made type & amount) needed? Analysis YES Tool use
Slide 20: Agentic system example: Loan grant and payment collection
Agentic system example Loan grant + payment collection (Case management) Payment Payment YES Case due received wrap-up ? process NO Debt collection sub-process 20 Digital Workforce Services
Slide 21: Learn about Agentic AI – for free!
Learn about Agentic AI – for free! 21 Digital Workforce Services
Slide 22: Thank You!
Thank You! 22 Digital Workforce Services
Slide 23: Thank You!
Thank You! 23 Digital Workforce Services
Slide 24: Agentic AI POCs: Unlocking Automation with UiPath
Agentic AI POCs Unlocking Automation with UiPath Mateusz Zybert Te ch Le ad at Dig ita l W or kfor ce [email protected] 24 Digital Workforce Services
Slide 25: 4 POCs
4 POCs Email Triage Agent Confluence Agent UiPath Agentic AI Ticket Handling Agent Build Documentation Agent 25 Digital Workforce Services
Slide 26: Email Triage Agent
Email Triage Agent Scope of the agentic process Input • The agent automatically analyses new emails New mail arrived in the Mailbox arriving in the mailbox • Classifies them as invoice, sales opportunity, spam, other, or unable to categorise. • Forward the email to the proper person Classified mail
Slide 27: Email Triage Agent
Email Triage Agent Why this is a good use Doubts and case for an AI agent considerations Emails might be large and Repetitive & high volumes exceed the allowed input/output size. Mails do not come in pre- defined format; classifying If email gets associated them would be very hard incorrectly by the
Slide 28: Confluence Agent
Confluence Agent Scope of the agentic process Input • Agent prompted by the user via Teams to search Teams message the internal Confluence Knowledgebase and return a summary based on the user's message. • The aim is to give a concise summary of the Search on appropriate knowledge base page results.
Slide 29: Confluence Agent
Confluence Agent Why this is a good use Doubts and case for an AI agent considerations Agent Builder is relatively The output of the straightforward, and with Confluence integration the UiPath Training, little service is quite difficult experience is needed to to parse and provide an begin building
Slide 30: Ticket Handling Agent
Ticket Handling Agent A new ticket arrived on FreshService Scope of the agentic process Input • Agent automatically analyses incoming tickets in Historical ticket data FreshService and • Classifies them as either Incidents or Service Requests Search on Confluence • Assigns them to the correct subgro
Slide 31: Ticket Handling Agent
Ticket Handling Agent 31 Digital Workforce Services
Slide 32: Ticket Handling Agent
Ticket Handling Agent Input Ticket Prompt Associator Contexts Agent Output 32 Digital Workforce Services
Slide 33: Ticket Handling Agent
Ticket Handling Agent All decisions are based on Agent outcomes All data sourced from the Agent Full integration via Integration Service (IS) 33 Digital Workforce Services
Slide 34: Ticket Handling Agent
Ticket Handling Agent We have a change in P001 concerning the Säkraplan 4 offer. We have changed the offer in Lumera and therefore need to update the robot to ensure it selects the correct offer in these cases. Säkraplan 4 should now select the offer called SäkraPlan Fondrabatt in Lumera. This refer
Slide 35: Ticket Handling Agent
Ticket Handling Agent Why this is a Doubts and Future good use case Future Use considerations Improvements for an AI agent Relies on past tickets having been correctly Populate context based on Can be easily adjust to any Repetitive & high volumes associated and being completed cases queries/tickets
Slide 36: Building Documentation Agent
Building Documentation Agent Robot Code (DevOps) Scope of the agentic process Input • The agent retrieves code from DevOps Orchestrator data • The agent pulls relevant data directly from the Orchestrator. • The RPMD template is located within the agent’s Analyze input execution context. • The agent
Slide 37: Building Documentation Agent
Building Documentation Agent • Tool in UiPath Studio Orchestrator • Template stored in AI Agent • UiPath Process execution • DevOps API Storage Bucket • JSON into DataTable • Dynamic search support • Tool in UiPath Studio • Storage Bucket used for • Table format from conversion by AI • Multiple Orch
Slide 38: Building Documentation Agent
Building Documentation Agent Agent loads input data Agent analyzes the template 38 Digital Workforce Services
Slide 39: Building Documentation Agent
Building Documentation Agent 39 Digital Workforce Services
Slide 40: Building Documentation Agent
Building Documentation Agent 40 Digital Workforce Services
Slide 41: Building Documentation Agent
Building Documentation Agent 41 Digital Workforce Services
Slide 42: Building Documentation Agent
Building Documentation Agent Why this is a Doubts and Future good use case considerations Improvements for an AI agent More comprehensive Manual documentation is time- documentation filling Documentation gaps remain consuming Auto-update of docs after code LLMs struggle with full code context change
Slide 43: Sum Up
Sum Up POCs: General Impressions: • Selected solutions are promising • Building agents requires a different • The timeframe is too short to mindset than traditional development measure gains • Input data must be rationally • Scalability to other systems is optimized relatively easy • At the start, y
Digital Workforce Services — UiPath Event (June 2025) on 6ix