> Markdown version of [/videos/157-intelligent-automation-using-machine-learning?t=1398](https://www.wearedevelopers.com/videos/157-intelligent-automation-using-machine-learning?t=1398). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# Intelligent Automation using Machine Learning

Are brittle selectors breaking your automated workflows? Discover how machine learning models replace hardcoded logic to deduce rules, handle exceptions, and drive definitive business actions.

- **Speakers:** Boris Krumrey, Andreas Palfi, Radu Pruna
- **Event:** WeAreDevelopers LIVE
- **Published:** May 26, 2021
- **Duration:** 2:11:02
- **URL:** https://www.wearedevelopers.com/videos/157-intelligent-automation-using-machine-learning

## Summary

The transition from rule-based execution into high-power automation fundamentally shifts how enterprises digitally operate. While traditional development dictates that programmers manually hardcode logic paths, introducing ML models redefines automations as dynamic workflows capable of deducing rules from historical data. This evolution allows software bots to visually parse interfaces utilizing deep learning-powered computer vision rather than relying on brittle underlying selectors, eliminating the limitations of simple screen scraping.

Bridging the gap between raw data science and tangible business actions remains a crucial hurdle in enterprise AI deployment. Integrating ML workflows directly into business operations—such as embedding human-in-the-loop validations via Action Center—solves this disconnect. This architecture actively mitigates standard model degradation because exceptions, like an illegible handwritten invoice or an ambiguous taxonomy, are flagged for human review and seamlessly routed back into AutoML pipelines for continuous retraining. Pushing process discovery further, task mining passively monitors user keystrokes to algorithmically surface high-value automation opportunities before a developer writes a single line of code.

"A machine learning model on its own is not very useful; you need a way to integrate it into an end-user application." This philosophy is showcased in predictive accounts receivable modeling, which ingests structured SAP records to forecast delinquent invoices. Instead of deploying stagnant analytics, the platform translates predictions into low-code app interfaces, equipping finance and sales teams with proactive remediation triggers like automated communication sequences. Ultimately, scaling intelligence necessitates decoupling the data scientist's environment from the developer's execution layer, delivering a unified ecosystem that converts probabilistic forecasts into definitive operational capacity.

**Keywords:** robotic process automation, machine learning lifecycle, computer vision UI parsing, intelligent document processing, human-in-the-loop ML validation, continuous AutoML retraining, process execution mining, accounts receivable prediction, predictive cash flow modeling, low-code enterprise applications, unstructured data extraction, model degradation management, SAP invoice integration, exception handling routing

## Chapters

1. **Introduction to UiPath and robotic process automation** (00:17) — The platform evolved from basic screen scraping algorithms to intelligent interface understanding using complex computer vision models.
1. **How machine learning approximates reality through mathematical functions** (08:05) — Algorithms estimate the best mathematical function from historical data sets to describe relationships between inputs and outputs.
1. **Deducing rules from data instead of manual programming** (17:22) — Machine learning models automatically learn decision rules from data characteristics instead of requiring explicitly coded logic statements.
1. **Overview of supervised, unsupervised, and reinforcement learning** (19:03) — The major categories of machine learning solve discrete business use cases through structured classification and regression approaches.
1. **Core stages of training and managing machine learning models** (23:18) — Real data science efforts heavily involve data processing pipelines and engineering tasks rather than simply tuning state-of-the-art algorithms.
1. **Moving machine learning models into production automation flows** (27:51) — Integrating prediction engines into user applications enables operational teams to act upon predictions and continuously retrain models.
1. **Using automation as the foundation for digital transformation** (34:27) — Companies scale digitization efforts by deploying automation flows that generate the metadata required for subsequent artificial intelligence initiatives.
1. **Building an end-to-end platform for automation and AI** (43:53) — Extending rule-based robotics with integrated cognitive capabilities enables robust task optimization across complex enterprise software environments.
1. **Processing unstructured data through intelligent document understanding** (54:20) — Combining robotic process automation with continuously trained models allows accurate data extraction from customized forms and handwritten papers.
1. **Uncovering automation opportunities via process and task mining** (57:33) — Machine learning tools analyze system logs and desktop interactions to visualize poorly standardized manual tasks that need optimization.
1. **Operationalizing custom and pre-built machine learning models securely** (60:11) — Connecting various narrow intelligence tools via application programming interfaces expedites secure model deployment for non-technical software developers.
1. **Break and automated assistant feature recorded demonstrations** (70:55) — Recorded system demonstrations show a locally running automation assistant handling resume screening and new employee onboarding flows.
1. **Selecting context-appropriate performance metrics for prediction models** (77:49) — Applying highly imbalanced training sets requires robust accuracy metrics like F1 scores rather than naive overall capability measurements.
1. **Predicting overdue invoices to improve accounts receivable collection** (81:32) — Classifying historical customer payment behaviors helps financial teams flag late payments before their assigned scheduling due dates.
1. **Transforming tabular metrics into meaningful business value dashboards** (88:52) — Structuring complex gradient boosted output into monetary confusion matrices enables operational stakeholders to prioritize financial payment collection efforts.
1. **Building a comprehensive accounts receivable cash collection application** (98:39) — A backend virtual machine triggers system scripts that pull client invoice data to predict delinquencies via deployed machine-learning backends.
1. **Executing independent automation workflows outside browser environments** (111:07) — Connected endpoint robots execute localized backend processes rather than utilizing client-side web browser compute processing constraints.
1. **Identifying appropriate business use cases for deploying artificial intelligence** (119:14) — Ideal automation targets involve high-volume, variable workflows where traditional deterministic rule-based logics become practically impossible to maintain accurately.
1. **Calculating return on investment and comparative software advantages** (124:07) — Dedicated evaluation portals synthesize operational effort requirements against monetary benefits while supporting both localized and cloud-deployed orchestration software methods.

## Related Moments

- [Transitioning artificial intelligence into operational business environments](https://www.wearedevelopers.com/videos/111-detecting-money-laundering-with-ai) (from "Detecting Money Laundering with AI")
- [Enhancing case matching with integrated machine learning tools](https://www.wearedevelopers.com/videos/86-rpa-in-the-public-sector) (from "RPA in the Public Sector")
- [Introduction to the speakers and process automation transformation journey](https://www.wearedevelopers.com/videos/1433-transforming-task-management-integrating-camunda-with-scalable-frontend-solutions) (from "Transforming Task Management: Integrating Camunda with Scalable Frontend Solutions")
- [Building intelligent applications and enhancing developer productivity experiences](https://www.wearedevelopers.com/videos/1096-architecting-the-future-leveraging-ai-cloud-and-data-for-business-success) (from "Architecting the Future: Leveraging AI, Cloud, and Data for Business Success")
- [Introduction to artificial intelligence driven development](https://www.wearedevelopers.com/videos/347-mlops-and-ai-driven-development) (from "MLOps and AI Driven Development")
- [Mapping AI integration across the software development lifecycle](https://www.wearedevelopers.com/videos/1539-agentic-devops-how-ai-powered-automation-transforms-software-delivery-on-github-and-azure) (from "Agentic DevOps: How AI-Powered Automation Transforms Software Delivery on GitHub and Azure")

## Related Articles

- [MLOps – What’s the deal behind it?](https://www.wearedevelopers.com/magazine/125-mlops-what-s-the-deal-behind-it)
- [MLops – Deploying, Maintaining And Evolving Machine Learning Models in Production](https://www.wearedevelopers.com/magazine/115-mlops-deploying-maintaining-and-evolving-machine-learning-models-in-production)
- [MLOps And AI Driven Development](https://www.wearedevelopers.com/magazine/82-mlops-and-ai-driven-development)
- [Coffee with Developers - Maria Apazoglou - Making AI understandable for all in production](https://www.wearedevelopers.com/magazine/475-coffee-with-developers-maria-apazoglou-making-ai-understandable-for-all-in-production)

## Related Jobs

- [Machine Learning Engineer](https://www.wearedevelopers.com/jobs/ext/588393-machine-learning-engineer) at **Twilio**
- [Machine Learning Engineer](https://www.wearedevelopers.com/jobs/ext/1355348-machine-learning-engineer) at **TWILIO**
- [AI Software Engineer (Germany)](https://www.wearedevelopers.com/jobs/48317-ai-software-engineer-germany) at **Sunhat**
- [Data Scientist](https://www.wearedevelopers.com/jobs/ext/1351648-data-scientist) at **Almedia**
- [Machine Learning Engineer](https://www.wearedevelopers.com/jobs/ext/1597388-machine-learning-engineer) at **ZEISS Group**
- [Staff, Machine Learning Engineer (L4)](https://www.wearedevelopers.com/jobs/ext/1202639-staff-machine-learning-engineer-l4) at **Twilio**