The Data Edge: AI, Procurement and FM

Stephanie Wiechers

Welcome to the Data Edge: a podcast series on data management, AI, and data quality.

Facilities management and procurement run on data. But most of it is unstructured, siloed, and AI-ready in name only.

The Data Edge is the podcast for FM directors, heads of procurement, CFOs, asset managers who want to know what good data actually looks like. and what it unlocks.

From invoice data and web quotes to comprehensive spend analysis, each episode covers the real work of turning raw procurement data into something you can actually act on — think data quality, AI readiness, and platforms like Microsoft Fabric doing what they're supposed to do.

Host Stephanie Wiechers speaks with industry leaders across construction, infrastructure, and hard services about the data decisions that drive operational performance.

Because better data isn't an IT problem. It's a competitive advantage.

Hosted by Stephanie Wiechers, CEO of Pearstop

Play

Open in

25 episodes
weekly
Avg 15 min
English

Episodes 25

Newest Oldest Longest

All20 Unplayed20 In progress0 Played0

Counted on this page — what you have heard stays on this device, so it is not something the list can be paged by.

September 20 · 20 min
How AI Is Actually Used in Construction & FM (Real Examples) | Daniel Roth on The E3 Effect
Most AI conversations in FM and construction stay stuck at the pitch stage. This one doesn't. Daniel Roth has spent 27 years running transformations across 15 industries, and he walks through three of them in detail: a construction camera taught to spot its own blind spots, a contact centre that grew 68% while cutting costs 57%, and an ERP migration that exposed a quarter of an organisation's spend with no category at all. 0:00 Intro: why most transformation problems are visibility problems, not strategy problems 1:10 The camera that couldn't see the job site 8:10 68% growth, 57% lower cost in a healthcare contact centre 11:03 A quarter of the spend, uncoded 17:00 From 1% of calls monitored to 100% 19:30 Why Daniel crossed "execute" off the cover of his book What you'll walk away with: How a blocked construction camera turned into an AI object-recognition product, then a security business What actually drove a 57% cost reduction alongside 68% volume growth, and why it wasn't one tech project Why "just put something in that field so the record saves" is how a quarter of an organisation's spend ends up with no category How Daniel's team uses two AI engines to check each other, with a human still in the loop The one question most procurement teams can't answer about their own spend data Daniel Roth is the founder of Pangea Effect and author of The E3 Effect, a framework for building high-performing teams around enable, empower and endeavor rather than execution. In this episode he and Stephanie Wiechers of Pearstop cover AI adoption in construction and facilities management, procurement data quality, spend classification, ERP migrations, and how contact centres use AI for quality monitoring at scale. The Data Edge is for FM, construction and infrastructure teams who want to get time and money back with their data. New episodes weekly. More on getting your own spend data into a state where you can act on it: pearstop.com
- Transcript
September 10 · 21 min
Why a UK School Wastes 3 Meals Per Student | Nicky Rifat on Data-Driven Waste
A school can be producing three meals of food waste for every student it feeds, and still have all the data sitting in its catering contract the whole time. This episode is about why that data goes uncollected, unseen, or just unused. 0:00 What we're covering 3:22 The data ladder — why collecting data isn't the hard part 7:41 Real case study: cutting general waste by making it visible 15:19 Keeping humans in charge: why the AI should prompt, never decide 18:49 The £78k school problem, and why the easy sale was the wrong one Three questions worth asking before your next waste review: 1. Is your waste data periodic (a monthly email) or real-time, and would real-time actually change what your team does differently? 2. Has anyone on your team seen a full breakdown of what's driving your general waste volume, or only the total? 3. If a supplier promised to "save you money" on waste, could they show you where that saving actually comes from? Nicky Rifat, founder of Green Space Innovations, walks through his data ladder — collect, visualise, analyse, change — and why most FM and facilities teams stall between the middle two steps. Covered: a real case study of a building that cut general waste and lifted recycling purely by making waste visible to occupants, why "you'll save money" is too broad a promise for any waste or sustainability vendor to make, the case for keeping a human in the loop on waste-stream classification rather than letting AI decide alone, and why a school generating £78,000 a year in food waste didn't need a bigger digester first, it needed a change in catering policy. The Data Edge is for FM, construction and infrastructure teams who want to get time and money back from their data. New episodes weekly. More on getting your procurement and waste data into a state where you can act on it: pearstop.com
- Transcript
S2 · E1
September 3 · 9 min
How to Handle a Supplier Price Increase Letter (FM & Construction) | Cleaning, Hard Services Example
A supplier increase letter is one of the few procurement decisions where you're expected to reply without ever seeing the numbers behind the number. This episode gives you those numbers. 0:00 What we're covering 0:50 Why the percentage in the letter is almost always too high 1:25 The maths, using a cleaning contract 3:20 Why hard services works completely differently 4:35 The four questions to send back 5:10 Construction: check the contract even allows it 6:25 The one question most teams can't answer 7:50 What to take away The four questions to send back: 1. Which rate or index does our contract actually tie the increase to, and does the letter match it? 2. What proportion of this contract is labour, and are you applying the increase to the whole contract value or only to that proportion? 3. How many of the staff on our site are paid at the statutory minimum today, and how many are already above it? 4. Is this a permanent change to our baseline price, or does it come back down if costs move the other way? Your cleaning contractor, M&E provider or subcontractor has sent a price increase letter, and the percentage in it looks non-negotiable. It usually isn't. Stephanie Wiechers of Pearstop walks through how to check whether a supplier price increase is actually justified, using worked examples from commercial cleaning, hard services maintenance and construction. Covered: why a rise in the UK legal minimum wage only justifies an increase on the labour portion of your contract, how the maths changes when you move from a labour-heavy cleaning contract to an M&E contract loaded with parts and materials, why construction increases need checking against the fluctuations provision and the named index in the contract, and the percentage-based management fee that quietly rises with the total. Also covered: why most procurement teams can't answer the one question that would save them the money, and what that says about their contract and spend data. The Data Edge is for FM, construction and infrastructure teams who want to get time and money back with their data. New episodes weekly. More on getting your procurement data into a state where you can answer that question: pearstop.com
- Transcript
August 20 · 14 min
Why 3 quotes?
Most FM and construction teams still rely on multiple quotes to decide whether a job is fairly priced. But what happens when the real pricing knowledge lives in the heads of just one or two experienced people? In this episode of The Data Edge, Stephanie Wiechers looks at the hidden cost of that system: slower approvals, unnecessary contractor work, weaker supplier relationships and teams waiting days—or even weeks—for decisions that should take minutes. Using a real-world facilities management example, Stephanie explores how historical invoice data can be turned into a practical cost baseline. Instead of requesting three quotes every time, teams can compare a new quote against prices they have already paid for similar work and quickly identify what looks reasonable, what looks unusual and what genuinely needs senior review. The episode also includes a simple exercise you can do with your own invoice data to start building a baseline without new software. In this episode: Why the three-quote process creates unnecessary friction How pricing knowledge becomes an operational bottleneck How historic invoices can reveal a fair-price range Why cost baselines can improve supplier relationships A practical way to build your first baseline in under an hour The Data Edge is for facilities management, construction and infrastructure teams looking to get more value from their data. #FacilitiesManagement #Construction #Procurement #DataAnalytics #CostManagement #FM #InvoiceData #SupplierManagement #ConstructionData #TheDataEdge
- Transcript
August 13 · 18 min
AI in hard FM – a real story of how to do it well
One number. Nearly a month to find it. The problem wasn’t missing data. It was data that didn’t agree with itself. In this episode of The Data Edge, Stephanie Wiechers breaks down why messy supplier and invoice data quietly undermines procurement decisions, and how to build a cost baseline you can actually trust. Because better decisions don’t start with another dashboard. They start with better data. #TheDataEdge #Procurement #DataManagement #FacilitiesManagement #AI
- Transcript
July 23 · 9 min
Find the REAL Money Leak
Somewhere in your contract, money may be hiding—and even the person managing it might not know where to find it. In this episode of The Data Edge, PearStop founder and managing director Stephanie Wiechers explores how poor invoice visibility can quietly reduce contract profitability across facilities management, construction and infrastructure businesses. Stephanie introduces “Mark,” an experienced contract manager overseeing a cleaning contract across 40 sites. Although his supplier totals and monthly reports appeared correct, they failed to reveal the real problem: identical consumables being purchased from different suppliers at different prices for more than a year. The issue was not fraud, negligence or bad management. It was simply hidden inside thousands of invoice line items that nobody had the time or tools to compare properly. This episode explains why supplier-level reporting is often insufficient, how invoice line-item data can expose cost leakage, and why recurring inefficiencies can be more damaging than one visibly bad month. Stephanie also shares a practical three-step invoice audit that contract managers, procurement teams and FM professionals can carry out immediately. Discover how better spend visibility, invoice data management and supplier price comparison can help protect margins and improve contract performance. Learn more about PearStop and access free resources through the links below.
- Transcript
July 16 · 17 min
No duplicate records - One number!
A supplier calls ahead of a contract renewal and asks for a price increase. Before responding, the procurement team needs one simple figure: how much are we currently spending with this supplier across every site, service and contract? It takes them almost a month to find the answer. In this episode of The Data Edge, Stephanie Wiechers, CEO of Pearstop, explains why procurement teams across facilities management, construction and infrastructure struggle to produce reliable spend data, even when all the information technically exists. The problem is rarely a lack of data. It is that supplier names, invoice descriptions, contracts, quotes and spreadsheets do not agree with one another. Stephanie breaks down a practical four-step process for turning fragmented procurement information into a trusted cost baseline: 1. Clean and consolidate duplicate supplier records. 2. Build a practical procurement spend taxonomy. 3. Categorise spend from the top down. 4. Use contextual data to interpret unclear invoice lines. You will also learn why dashboards cannot solve poor data quality, why excessive categorisation causes projects to collapse, and how organising 95% of spend confidently is often more valuable than chasing unreliable perfection. With clean procurement data, contract renewals become faster, supplier negotiations become more informed, and hidden costs such as fuel surcharges, duplicate suppliers and inconsistent pricing become easier to identify. Pearstop helps procurement teams transform messy invoices, quotes and spreadsheets into structured, decision-ready spend intelligence. Subscribe to The Data Edge for practical insights on procurement data, spend analysis, supplier management and AI-powered data transformation. #Procurement #SpendAnalysis #SupplierManagement #FacilitiesManagement #ConstructionData #ProcurementTechnology #DataQuality #CostReduction #SpendManagement #Pearstop
- Transcript
S1 · E9
July 2 · 12 min
What's Hiding in Your Data?
In this episode of The Data Edge by Pearstop, Stephanie Wiechers takes facilities management, construction and infrastructure teams inside the hidden world of procurement data. From supplier totals to invoice lines, this episode breaks down why “clean data” is rarely as clean as it looks, and how line-by-line spend visibility can help businesses protect margins, reduce waste and make smarter procurement decisions. Using a real-world facilities example, Stephanie explains how everyday site purchases like toilet cleaner, paper products, PPE, transport costs, packaging and fuel surcharges can reveal major gaps in spend control. For CFOs, procurement leaders and operations teams, this episode shows how structured procurement data can support supplier benchmarking, contract compliance, tender preparation and better cost management across the UK, Netherlands, France and global facilities markets. If your organisation manages multiple sites, suppliers, invoices or cost centres, this episode will help you understand what is really happening inside your procurement data and why better data quality creates better commercial decisions. Learn more about Pearstop and how better procurement data can give your business its edge. procurement data, facilities management, construction procurement, infrastructure data, invoice line analysis, spend visibility, supplier benchmarking, data quality, fuel surcharges, procurement software, Pearstop, The Data Edge podcast.
- Transcript
S1 · E15
June 25 · 9 min
100k£ secretly billed on packaging
We started Pearstop because we were fed up. This episode talks about another very relevant and difficult hurdle in the world of procurement and data - secret costs and hidden billings. Listen to the Data Edge every Thursday to stay informed on the latest data updates and more!
- Transcript
June 18 · 23 min
Dead Whales and Data
Back to the Data Edge - a podcast for those interested in understanding the Data and AI world. In today's episode we talk about waste management, data and the world of procurement with an interesting and unexpected anecdote that will grip our listeners. Tune in for an educative and insightful conversation between Stephanie, CEO at Pearstop and Mike from SLS!
- Transcript
June 11 · 14 min
Use Data & Save 100k on Procurement
The conversation delves into the challenges of reading and utilizing spend data, highlighting the importance of data structuring and classification for effective analysis. It emphasizes the shift in data classification and the benefits of structured data for procurement and finance teams. Whether you're a procurement manager, finance lead, or operations director dealing with messy invoice data, unstructured spend, or poor visibility across suppliers and categories — this episode covers why spend data classification is the foundation of any serious spend analytics programme. Takeaways Spend data analysis requires structured and classified data for effective utilization. The shift in data classification has enabled scalable and accurate data processing. Clean, classified spend data gives procurement and finance teams the visibility to negotiate better, cut waste, and make faster decisions. Chapters 00:00 Introduction to Spend Data Analysis 09:33 Data Structuring and Classification 15:41 Conclusion and Call to Action Keywords: spend data analysis, procurement data classification, spend analytics, UNSPSC classification, invoice data, procurement visibility, spend management, facilities management procurement, construction procurement, finance data quality
- Transcript
S1 · E13
June 3 · 14 min
AI in FM: today's steps for tomorrow's decisions
The conversation explores the use of data in facilities management and the challenges of collecting, managing, and utilizing data effectively. It delves into the role of AI in data analysis and decision-making, as well as the importance of clean and structured data for creating measurable value. The conversation also highlights the need to empower people through technology and attract the younger generation to the facilities management industry. Takeaways Data collection and management are crucial for effective decision-making and creating measurable value. Empowering people through technology and attracting the younger generation to the facilities management industry are key focus areas. Chapters 00:00 The Role of Data in Facilities Management 11:31 Types of Data and Business Impact 17:04 Empowering People and Attracting the Younger Generation
- Transcript
S1 · E11
April 30 · 18 min
AI for Asset Management & Data Quality
In conversation with SPIE Netherlands about how they improved the quality of their asset data. Our guest is the amazing Martijn van Balkom – passionate, smart, and guaranteed to teach you something new today! This episode provides insights into the challenges and opportunities surrounding asset data management. It highlights the importance of data quality, the impact of AI technology, and the role of asset data in advising and optimising maintenance strategies.
- Transcript
S1 · E10
April 23 · 18 min
Data Quality in Manufacturing and Procurement
Unlocking Data Quality in Manufacturing and Procurement with Jonas Hauswurz"In this episode, Stephanie Wiechers chats with Jonas Hauswurz, founder of Neomir, about the real issues behind data quality problems in manufacturing and procurement. Discover practical use cases, how to identify critical data leaks, and the importance of focusing on operational teams rather than just executives.Key Topics: The distinction between data problems and data quality issues How Neomir's software detects bad data at scale and automates resolutions The significance of rules in data validation, e.g., ensuring cars always have four wheels Practical applications in manufacturing, procurement, and asset management The challenge of tying data quality to measurable ROI and cost savings The different approaches to data quality: identification vs. pre-filling and machine learning Why operational teams often detect data issues before management does Strategies for engaging ground-level staff to improve data quality Timestamps: 00:00 - The difference between data issues and data quality flaws 00:25 - Why data quality is often misunderstood in organizations 00:55 - How Neomir's software identifies and automates fixing bad data 01:20 - Practical examples in manufacturing and asset management 03:12 - Creating rules for data validation with AI assistance 04:29 - Common data errors in manufacturing, such as bill of materials inaccuracies 05:15 - Propagation of data checks through entire supply chains 07:28 - The role of data quality in cost savings and strategic decision-making 08:23 - Indirect effects of data quality on financial performance 09:58 - Approaches to measuring ROI for data quality initiatives 11:03 - Advantages of transparency and rough ROI estimates over precise calculations 13:23 - Engaging operational teams for better data insights 15:45 - How management often underestimates data issues until front-line staff reveal them 16:24 - The importance of targeting conversations at data specialists and operational staff 18:05 - Closing thoughts and how to connect with Jonas for manufacturing and procurement data challenges Resources & Links: Stephanie Wiechers - LinkedIn Jonas Hauswurz - LinkedIn
- Transcript
S1 · E10
April 9 · 16 min
AI & Data Standards
𝗕𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗮 𝗥𝗲𝗹𝗶𝗮𝗯𝗹𝗲 𝗗𝗮𝘁𝗮 𝗟𝗮𝘆𝗲𝗿: 𝗜𝗻𝘀𝗶𝗴𝗵𝘁𝘀 𝗳𝗿𝗼𝗺 "𝗧𝗵𝗲 𝗗𝗮𝘁𝗮 𝗘𝗱𝗴𝗲" 𝗣𝗼𝗱𝗰𝗮𝘀𝘁 In this episode of The Data Edge, Erwin de Werd and guest Stephanie Wiechers explore the critical aspects of data quality, standardization, and data movement for organizations aiming to leverage AI and advanced analytics effectively. They discuss practical challenges and strategic considerations for companies of all sizes seeking to build trustworthy, scalable data infrastructure.  𝗠𝗮𝗶𝗻 𝗧𝗼𝗽𝗶𝗰𝘀: ✔ The increasing importance of data quality and reliability in AI applications ✔ Challenges in creating and trusting dashboards due to data flaws ✔ How data movement between systems influences decision-making and analytics ✔ The role of standardization in cross-entity data sharing and efficiency ✔ Trends and best practices for adopting data standards and improving data governance ✔ The impact of AI tools like Copilot on data analysis and development ✔ Strategies for smaller businesses to align with industry standards despite resource constraints  𝗧𝗶𝗺𝗲𝘀𝘁𝗮𝗺𝗽𝘀: 00:00 - Introduction and overview of data quality challenges in AI development 00:30 - The surge in democratized data analysis and its responsibilities 01:34 - Risks of trusting dashboards with potential data flaws 03:07 - The importance of data reliability for decision-making 04:13 - Moving data across systems to enable advanced analytics 05:18 - The significance of data standardization in different industries 06:34 - How data lakes and recent platforms support data integration 07:45 - The role of data quality as a foundation for dashboards and AI models 08:26 - Standardization trends and industry-specific norms 09:13 - Cost considerations and strategic choices in implementing standards 10:27 - Challenges and strategies for smaller companies adopting standards 11:48 - Practical steps for transitioning from non-standard to standardized data 12:18 - Industry standards like UNSPSC and industry-specific frameworks 13:25 - The strategic value of standardization for cost savings and operational efficiency 14:09 - Use cases in procurement and spend analysis 15:13 - The growing importance of data quality and standardization in analytics 16:02 - Final thoughts and future topics  𝗥𝗲𝘀𝗼𝘂𝗿𝗰𝗲𝘀 & 𝗟𝗶𝗻𝗸𝘀: • UNSPSC (United Nations Standard Products and Services Code) – Industry-standard classification for products and services
- Transcript
S1 · E9
April 2 · 17 min
Succesfactors for AI
𝗨𝗻𝗹𝗼𝗰𝗸𝗶𝗻𝗴 𝘁𝗵𝗲 𝗣𝗼𝘄𝗲𝗿 𝗮𝗻𝗱 𝗣𝗶𝘁𝗳𝗮𝗹𝗹𝘀 𝗼𝗳 𝗔𝗜 𝗮𝗻𝗱 𝗗𝗮𝘁𝗮 𝗠𝗮𝗻𝗮𝗴𝗲𝗺𝗲𝗻𝘁 In this episode of The Data Edge, Erwin de Werd and Stephanie Wiechers explore how AI can transform data management from a headache into a strategic advantage — if used wisely. They discuss the pitfalls of overhyped AI solutions, the importance of building robust systems, and practical steps to improve data quality. 𝗞𝗲𝘆 𝗧𝗼𝗽𝗶𝗰𝘀: The proliferation of AI "skills" and why over 90% are ineffective How automation, when done properly, enhances data quality and operational efficiency The challenge of discerning quality in AI tools and avoiding superficial solutions Practical examples of AI in lead generation (Dream 100 strategy) and content creation How to build trust in AI-driven data solutions amidst industry hype The importance of authentic, human-centered communication in AI content The distinction between front-end conversation and back-end automation in data management Planning for a future where AI and data quality ensure better decision-making 𝗧𝗶𝗺𝗲𝘀𝘁𝗮𝗺𝗽𝘀: 00:00 - Introduction: Transforming data management with AI 00:30 - Why most AI skills are ineffective and what they entail 01:25 - Explanation of skills as standard operating procedures (SOPs) 02:24 - The explosion of AI skills on platforms like Instagram and their usability 03:20 - The common problem of people not doing the work when using AI tools 03:50 - Strategic laziness: automating repetitive tasks with quality checks 04:32 - Pitfalls of trusting AI outputs without proper validation 04:57 - Challenges in training AI models to produce accurate, high-quality content 05:44 - Limitations of custom GPTs in professional tasks like LinkedIn content 06:22 - The importance of investing effort upfront to create effective automation systems 06:47 - Why cost savings lead to underinvestment in AI automation 07:34 - Challenges of relying on incomplete or careless prompts 07:45 - The habit of short-input prompts and the impact on output quality 08:13 - Building outreach strategies with AI: the Dream 100 example 08:51 - Automating research and outreach to generate leads efficiently 09:35 - Using AI to identify influencers and industry events for strategic networking 10:58 - The need for consistency and authenticity in AI-generated content 12:04 - How good copywriters leverage AI as a starting point, not a replacement 12:51 - Authenticity remains crucial despite the efficiency gains from AI 13:17 - Connecting AI automation in data management with operational layers of business 14:09 - The importance of backend automation for data quality and integrity 15:14 - Trust issues in procurement and other industries regarding AI promises 16:26 - The hype versus reality of AI solutions, and the upcoming industry shakeout 17:08 - Final thoughts: Deepening the conversation in future episodes
- Transcript
S1 · E9
March 26 · 18 min
AI & Human Collaboration
🎙️ 𝘁𝗵𝗲 𝗱𝗮𝘁𝗮 𝗲𝗱𝗴𝗲 — 𝗱𝗮𝘁𝗮 𝗾𝘂𝗮𝗹𝗶𝘁𝘆 𝗶𝗻 𝗮𝗶 𝗽𝗿𝗼𝗷𝗲𝗰𝘁𝘀 In this episode, Erwin and Stephanie delve into the complexities of data quality in AI projects, emphasizing that messy data often leads to costly mistakes. They explore how human-AI collaboration and understanding the limitations of models like LLMs are crucial for success. 🔑 𝗞𝗘𝗬 𝗧𝗢𝗣𝗜𝗖𝗦 The common misconception that first data categorization is 100% accurate — and why errors are part of the process The reality of achieving high data quality and near automation (up to 95%) in data processing Expectations vs. reality: Why clients sometimes expect AI to be a 'magic bullet' and how to set realistic goals The importance of contextual knowledge and communication to improve model accuracy Methodologies for training AI models as 'new employees', including leveraging human expertise and internal knowledge A real-world construction project: data categorization challenges, including language issues (tablets as lozenges) Differentiating LLMs like ChatGPT from specialized machine learning models The role of human-AI cooperation in improving data quality and operational efficiency Creating a knowledge center for clients through ongoing data training and model refinement The value of building IP within organizations by developing tailored data solutions and models ⏱️ 𝗧𝗜𝗠𝗘𝗦𝗧𝗔𝗠𝗣𝗦 00:00 Introduction: The impact of messy data on industry costs 00:30 Setting the stage: From data quality to correction hiccups 01:14 Why initial categorization often isn't perfect — and it's normal 02:02 The misconception of AI producing perfect results immediately 02:50 Achieving high data quality and near automation possibilities 03:17 Managing client expectations around AI and data processing 04:05 Importance of communication about processes and contextual insights 05:14 When models don't perform as expected: Training methodologies 05:45 Example project in construction: Data categorization challenges 06:47 Using dashboards to identify and fix misclassified data 08:11 Language nuances affecting classification (e.g., tablets as lozenges) 08:58 Differences between LLMs like ChatGPT and task-specific ML models 10:16 The core distinction: General language models vs. specialized models 12:11 Why consistency and rule-based training are vital 13:24 Human-AI collaboration enhancing data accuracy 14:02 Implementing biases and industry knowledge to improve models 15:19 Building an organization's IP through data and model development 16:21 Potential for transparency: Sharing system rules with clients 17:05 Recap: Differentiating AI types and combining human expertise 18:18 Closing: Key takeaways on data, AI, and IP in projects
- Transcript
S1 · E8
March 19 · 18 min
Insights from the field
Turning Data Quality into Strategic Power: Insights from Procurement and Data Management In this episode, Erwin de Werd and Stephanie Wiechers dive into the complexities of data quality in procurement and data management. They share actionable insights from a detailed conversation with a seasoned procurement professional, highlighting how data quality impacts savings, efficiency, and strategic decision-making.Main Topics: The role of data quality in procurement and its impact on savings Category management in both predictable and project-based industries The importance of data accuracy at different organizational scales How large enterprises handle demand unpredictability and data complexities Navigating risk aversion and innovation adoption in data-driven projects Small pilots and iterative approaches for embracing AI and new technologies In this episode: Stephanie shares her interview with a former head of procurement from a billion-dollar enterprise The discussion on category management illustrates how scale and industry type influence data practices The conversation highlights the critical role of data quality in unlocking procurement savings Insights on managing organizational change and technology adoption speed, including AI • Practical advice: start small with pilot projects to manage risks and learn quickly
- Transcript
S1 · E7
March 13 · 12 min
How to reach 95% Data Quality
Ensuring Data Quality in AI Projects: A Conversation with Stephanie Wiechers In this episode, Erwin de Werd and Stephanie Wiechers explore the crucial role of data quality in AI and data projects. They discuss practical approaches to maintain high accuracy, the challenges of testing AI with AI, and the importance of human oversight to achieve reliable results. Key Topics: The impact of messy data on AI output and decision-making Strategies for achieving 95% data accuracy for automation The process of data enhancement using AI and rule-based systems Testing AI models: AI-to-AI vs. human review approaches Cost and time considerations in data quality verification The ongoing progress: from 85% to over 95% accuracy The collaborative role of humans and AI in data validation Future outlook: the importance of human involvement for reliable AI Timestamps: 00:00 - Introduction: How messy data costs industries billions 00:41 - Importance of data quality in AI and reporting 01:25 - Common issues with data errors impacting insight generation 02:17 - Automating error detection and correction in databases 02:58 - Client quality expectations and the 95% accuracy benchmark 03:26 - Achieving and validating 95% accuracy in AI models 04:01 - Using AI and internal rules for data enhancement 04:41 - Challenges of testing AI with AI and the need for human validation 05:56 - The risk of relying solely on AI for quality checks 06:37 - Human review as a reliable fallback 07:03 - The four-step process for data validation 08:25 - The iterative role of human review and AI learning 09:06 - Balancing internal and outsourced validation efforts 10:17 - Outsourcing testing versus internal validation challenges 11:13 - Current progress: surpassing 85% accuracy 12:00 - Upcoming guest episode and future projects Resources & Links: PeerStop Connect with Stephanie Wiechers: LinkedIn Note: Stay tuned for our next episode featuring a special guest from the field discussing real-world data projects and best practices.
- Transcript
S1 · E6
February 25 · 16 min
The Human Factor in AI-Driven Procurement Data Management
The Human Factor in AI-Driven Procurement Data Management In this episode, Erwin de Werd and Stephanie Wiechers explore the critical interplay between human expertise and AI in ensuring data integrity and standardization within procurement processes. Discover how organizations leverage AI to enhance categorization accuracy, streamline validation, and safeguard sensitive information. Key Topics The importance of human input in AI-driven data categorization Challenges of enterprise-level procurement data standardization Combining rule-based systems with machine learning models for enhanced accuracy The role of the validation process in ensuring data quality Leveraging large language models (LLMs) for granular categorization How ongoing user feedback refines AI performance over time Data security policies and anonymization in AI training Practical steps for integrating AI with existing procurement workflows The future of collaborative man-machine approaches in enterprise data management Timestamps 00:00 - Introduction to the role of data quality in AI and enterprise decision-making 00:42 - The importance of the human factor in AI projects 01:37 - Case study: Procurement data integrity challenge in a large organization 02:51 - Standardization challenges across multiple sites and teams 03:44 - AI complexities in categorizing diverse invoice costs 04:48 - Systemizing procurement data processes through AI and human insights 05:42 - Combining rules and machine learning for improved categorization 07:00 - Utilizing large language models for granular and flexible data classification 08:54 - Automating validation and review processes within AI systems 11:04 - Achieving high accuracy through training and feedback loops 12:19 - Validation workflows involving multiple departmental reviews 13:55 - Sharing and securing enterprise data in AI applications 15:02 - The balance between data sharing and confidentiality in AI training 16:16 - Ensuring compliance with corporate data policies and security policies 17:01 - The evolving collaboration between humans and AI in procurement 17:17 - Upcoming series: Field insights from client interviews Connect with Stephanie Wiechers: LinkedIn
- Transcript

Showing 1–20 of 25 episodes