Winning with AI
Practical Applications for
Mid-Market Manufacturing
Agenda
Introduction
Case Studies
Data Readiness
Next Steps
Your Speakers
Charles Pensig, Co-Founder
- Wharton Center for Analytics and AI, Director
- Silicon Valley $2-300M hyper growth tech startups, 2 acquisitions
- Expertise in Automation, Insights, and AI for Commercial success
Derek Chang, PhD, Co-Founder
- Stanford Ph.D. Electrical Engineering
- Multi-$Bn Government Research projects contributor
- Silicon Valley for $2-300M hyper growth tech startups, 2 acquisitions
- Expertise in AI and Optimization for Operational and Commercial success
Creating competitive advantage through data and AI, by answering the 5 foundational questions:
And trusted by…
What we heard: "What would you most like to understand about data and AI?"
Practical applications for growth and efficiency, and how to measure success are at the forefront.
What we heard: "What challenges in your business could be solved by AI?"
Commercial, Operations, and Finance use cases are at the top of the list.
What we heard: "What are your top concerns in adopting data and AI technology?"
Whether the team and systems are ready, and whether you will get ROI from investing, are top concerns.
Your questions and pains are shared by many
In a BCG + World Economic Forum Survey of Manufacturing Executives poll of 1,800 executives, 89% had plans to implement AI, 75% of those with plans got started, and 24% of those that started met their goals.
Executives Surveyed
Plan to Implement AI
Have Started
Met AI Goals
This conversation aims to answer 3 questions:
Agenda
Introduction
Case Studies
Data Readiness
Next Steps
Case Studies
What value can I get from data and AI?
You're sitting on a gold mine, the question is where to start digging.
Case 1: Organized data foundation saves contract manfucturer 1,000 hours per year
"We were drowning in reports—hundreds of them outdated, redundant, and inconsistent across departments. It was chaos."
- Operating Partner at PE Sponsor
Departmens
Data Siloes
Reports
IT Analysts
Case 1 Challenge: "Every department had its own version of the truth. No one knew which numbers to trust."
Data Fragmentation
Different functional groups relied on separate data sources, leading to discrepancies.
Manual Reporting & Human Error
Critical reports were manually compiled, increasing the risk of errors and consuming valuable employee time.
Bottlenecks in Automated Reporting
Only 2 people were responsible for generating reports, creating a significant bottleneck.
Lack of Scalability
Reports were often specific to individuals or teams, making it difficult to scale insights across the organization.
Case 1 Solution: "They didn't just build a data warehouse, they built a foundation for us to grow on."
Clean, Centralize, and Automate Key Data
- Know what you can trust and create "Golden Data"
- Create something, anything, people can trust
- Test and iterate to first moment of trust
Build PoC Power BI Reports
- Build a simple interface
- Start building muscles
- Usability feedback
Train IT Team and Dashboard Users
- Change management for builders
- Change management for users
- Weekly office hours
Case 1 Impact: "I used to wait days for reports, and now I get answers in seconds. Every morning, I log in and have everything I need. No emails, no delays—just data, exactly how I want it."
Hours Saved per Year
Across its IT team, executives, and fund sponsors.
Reduction in Wait Time
The fund used to send up to 5-10 requests per week, each taking days-weeks to fulfill. Power BI = seconds.
Key Man Risk
Whereas reports used to depend on key members of IT, all data requests are now self serve.
Cultural Shift
Employees who relied on manual reporting started proactively requesting Power BI implementations for their workflows.
Sustained Data Governance
Post-project, the company’s IT team continued developing the warehouse while adhering to best practices.
Case 2: Insights dashboard and automation earn global healthcare manufacturer +18% revenue within 12 months
"It was terrible. You couldn't quickly look at last year, this year, month to month—what’s going on, where are we hitting, where are we missing?"
- Global VP of Sales
Case 2 Challenge: "You don’t realize how much time you’re wasting until you have something better. We were constantly checking, validating, and rechecking numbers."
Data Fragmentation
Account executives and finance teams spent excessive time validating disparate numbers.
Bottlenecks
1-2 financial analysts responsible for reporting/
Human Error
JC Speer described the previous process as "terrible"—relying on massive Excel sheets that slowed down even basic queries. Small errors could cascade into larger miscalculations, impacting decision-making.
Limited Scalability
There was no seamless way to provide live updates for leadership or external partners. Reports were all "one-offs".
Case 2 Solution: "Now, instead of wrestling with numbers, we’re having conversations about growth strategy and tactics."
A Single Source of Truth
Centralized data source with automated pipelines from Excel spreadsheets, ERP, and CRM.
Realtime Sales Dashboards
Instead of waiting for spreadsheets to load, account executives could now log in, click, and instantly see sales trends, missing targets, and growth opportunities.
Customer QBR Dashboard
Global VP Sales emphasized how the new system even changed client interactions: "one of our largest customers, was so impressed with our speed and data accuracy that they realized they needed a better system themselves.".
Realtime Commission Calculator
Realtime feedback showing salespeople exactly where their effort was going and what it was turning into.
Case 2 Impact: "We’re seeing orders increase just from being able to have these data-driven discussions. The insights are driving real revenue."
Tangible Revenue Impact
- $300k immediate return from a key distributor after AE identified purchasing inconsistencies in seconds—a task that would have taken days before.
- 18% year over year revenue growth across US and International regions.
Decisions instead of Debate
- 25% reduction in manual data requests, saving finance and sales teams weeks of work annually.
- Seconds instead of days to generate QBRs, enabling more strategic discussions rather than time-consuming data gathering.
- Live customer reporting, allowing account executives to react to conversations on the fly.
A Culture Shift Towards Data-Driven Decisions
- Sales team focused on strategy instead of debating numbers.
- Leadership discussions became more proactive, using real-time insights rather than waiting for outdated reports.
Case 3: AI product segmentation and demand forecasts drives 5% revenue and 2% gross margin boost
Global medical-simulation manufacturer with 80+ years of history, faced major data quality & operations issues. CEO Ken Miller noticed messy data and outdated techniques were interfering with the accuracy of foundational company-wide initiatives.
Case 3 Challenge: Messy data and outdated techniques were interfering with the accuracy of foundational company-wide initiatives.
No Single Source of Truth
Half of the product master data was missing, leading to fragmented and incomplete information.
Untrustworthy Analysis
Reports generated unintuitive and inconsistent results, eroding confidence in strategic insights.
Time Wasted on Validation
Teams spent valuable time re-interrogating numbers instead of focusing on critical decision-making.
Case 3 Solution: Automated product-naming system enables CEO to fully analyze full product line.
Clean Data Foundation
Cleaning 8,000 SKUs unlocked previously unavailable analysis, and formed a foundation for all future analyses.
Key Revenue Drivers
Product segmentation revealed 120 SKUs were responsible for 50% of the company's revenue.
Strategic Portfolio & Forecasting
Reliable data enabled the CEO to rationalize the portfolio and forecast demand with AI
Case 3 Impact: the algorithm’s prediction exceeded human predictions by 3.5x. Orders spiked, and the company was able to capture all the unprecedented additional revenue.
50% Portfolio Reduction
CEO immediately cut 4,000 SKUs, enabling focus on high-revenue and margin products.
2% Margin Improvement
The company's margin profile improved by 200 basis points in 12 months.
5% Revenue Growth
AI demand forecasting caught a spike, earning a net new 5% revenue within 12 months.
Agenda
Introduction
Case Studies
Data Readiness
Next Steps
Data Readiness
How do I create opportunities for my organization?
The 5 questions offer a guide to data maturity
And they give way to general solutions
To find specific solutions, we recommend scanning each department for data readiness
And after identifying opportunities, weighing them against feasibility for triaging
Agenda
Introduction
Case Studies
Data Readiness
Next Steps
Next Steps
What can I do today?
Where can I create measurable impact within a quarter?
Please reach out with any questions, big, small, or tiny:
Appendix:
Pillars of AI Readiness
AI Readiness stands on 4 pillars
It starts with leadership
Establish a clear vision and strategy
Identify potential opportunities with large returns
- Automation of routine tasks to increase efficiency
- Insights to enhance decision making
- Prediction to get ahead and anticipate needs
Ensure strategic alignment
- How does it enhance our competitive advantage?
- How does it help us grow top line?
- How does it help operate efficiently?
Prioritize
Weigh ROI against feasibility
Build a robust data foundation
Cultivate culture and empowerment
Focus on people, not tech
- Training & empowerment
- Hire for AI and data fluency, not just technical skills
Expectation of continuous learning and improvement
- Provide a safe space to experiment and try new ideas
Track performance and measure ROI
- It’s an iterative process, not a one-time project
Appendix: The Future of AI
What is AI?
Artificial Intelligence (AI) is computer systems that perform human intelligence tasks such as problem-solving and decision-making.
AI is able to processes information, recognize patterns, and make decisions, by learning from data.
How do machines learn from data?
How do machines learn from data?
How do we get closer to human intelligence?
Rule-Based Heuristics
Analytical AI
Generative AI
How do Large Language Models work?
A Large Language models capture the statistical relationships between words. They have learned to guess next words exceptionally well.
Here's what you can do with Language AI now…
And before Language AI, there is a mountain of value in Analytical AI
