Ohio's BigCo Playbook for AI: Eight Strategies Turning Midwest Giants into Next-Gen Innovators

AI is no longer a pilot project in Ohio; it's now cutting hours, reducing waste, and delivering new products. At the 2025 Ohio Tech Summit, seven enterprise-focused executives detailed the specific strategies that are driving this change. Here are the eight strategies that any company can implement today.

1. Use AI to Accelerate ESG

"We've already prototyped a tire at 90 percent sustainable content, and customers can buy a 70 percent version today." - Erin Spring, Goodyear

My take: ESG only sticks when it also protects margin. Goodyear used generative simulation to screen thousands of bio-based and recycled compounds. The result: a lab-validated tire with 90% sustainable materials and a commercial tire at 70% now on the road in the United States and China.

Copy this:

1. Build a living knowledge graph. Feed every spec sheet, bench result, and failure log into a domain-tuned language model so it “remembers” what worked and why.

2. Prototype virtually first. Use AI-driven finite-element or physics-informed models to score durability, weight, and manufacturability in hours instead of weeks.

3. Optimize for multiple bottom lines. Track cost, performance, and carbon footprint side by side. Let the model surface formulations that raise two of the three without tanking the third.

4. Keep experts in the loop. Materials scientists and process engineers review the short list, fine-tune parameters, and green-light only the variants that meet regulatory and supply-chain constraints.

2. Aim for High-Impact Wins, Not Shiny Objects

"In big companies, there is no faking value." - Todd James, Aurora Insights

My take: Don't waste your time if the initiative does not touch a top-three pain point. Brad Reynolds at Expedient stopped pitching generic chatbots and instead offered vector search across decades of internal files. Time to insight dropped by a factor of ten, and budgets opened fast.

Copy this: List your three ugliest workflows, break each into thirty steps, then automate the two that waste the most hours.

3. Build Trust with a Risk-First AI Framework

"We rate every AI idea on patient safety, data privacy, and brand risk." – Anagha Vyas, Cardinal Health

My take: Speed without guardrails is destined for disaster. Cardinal's matrix lets low-risk pilots launch in weeks while high-risk ideas get extra controls from day one.

Copy this: Build a one-page table with the three risk columns (safety, privacy, and brand risk), score each use case, and review only the high-risk rows in a monthly council.

4. Invite Customers to Bring Their Own Models

"We give clients a secure slot to plug in their own models." – Pinal Patel, Veeva

My take: Vendor lock-in is fading. Veeva will let life-sciences customers run proprietary language models inside the Veeva platform, keeping regulatory teams happy and speeding adoption.

Copy this: Expose a secure API for client-hosted models, provide a starter prompt library, and charge by usage.

5. Empower the Front Lines with AI

“Our technicians can ask a voice bot if a batch is in spec without removing haz-mat gloves.” – Anagha Vyas, Cardinal Health

My take: User experience makes or breaks adoption on the plant floor. Cardinal Health layered voice search on top of 300-page SOPs, cutting batch loss and overtime in radioactive-pharma production.

Copy this, make hard jobs easier, even in extreme conditions:

1. Shadow the frontline. Spend one shift watching where hands are busy, gloves are on, or temperatures are sub-zero, whether it is haz-mat batching, data entry in a freezer room, or quality checks on a roaring shop floor.

2. List the five moments of maximum friction. Glove removal, keypad typing in the cold, walking to a terminal, flipping through binders, anything that stalls the task.

3. Drop the answer box at the point of pain. Embed a voice or touch-free retrieval chatbot that surfaces the exact SOP, spec, or value the worker needs, no scrolling required.

4. Maintain a human fail-safe. Route uncertain queries to an expert until the system proves at least ninety-five percent accurate in real use.

When AI removes the micro-frictions like cold hands, bulky gloves, or noisy lines, productivity jumps and adoption sticks.

6. Turn Disconnected Data into a Unified Brain

"Imagine cloning Mary, the thirty-year veteran who knows everything." – Robert Krause, Hyland

My take: A single lake is nice in theory, but expensive and time-consuming in practice. Hyland built a federated vector index across twenty repositories, letting claims adjusters query as if everything lived in one brain. Fraud losses fell almost ten percent with zero data relocation.

Copy this: Stand up an open-source vector store on a small slice, prove search relevance first, schema cleanup second, then expand.

7. Validate and Scale with Real-World Results, Not Projections

"Developer co-pilots already give us a thirty percent productivity bump." – Doug Vargo, CGI

My take: Leaders buy clear before-and-after deltas, not hype. CGI tracked hours spent on unit tests across two sprints, introduced code co-pilots, and re-measured. The bar chart showing reclaimed hours was the driving factor for enterprise adoption.

Copy this: Baseline one repeatable task or process for one team, add AI, remeasure, and broadcast the before and after graphic.

8. Personalize Customer Messaging at Scale

“Generative AI let us jump from covering 15 percent of flight-change emails to more than 50 percent, with twice-as-fast resolution.” -Sam Porras, Slalom

My take: Customer satisfaction moves the revenue needle fastest. Slalom helped an airline plug an LLM into its schedule-change workflow, auto-drafting empathy-laden emails referencing each traveler’s itinerary and status. The system resolved issues twice as fast and lifted coverage from one in seven passengers to one in two.

Copy this:

1. Tag your top customer-service pain point (cancellations, claims, account updates).

2. Fine-tune an LLM on past high-quality responses.

3. Route AI-drafted messages to agents for one-click approval or edit.

4. Track resolution time and NPS to prove lift.

Why Ohio is the place to copy these moves

"Use case beats hype because AI is only real when it hits the P&L." – Lynn Sautter Beal, Upstart

Intel is constructing new semiconductor fabrication plants in Ohio, venture capital is flowing at record levels, and OSU is pledging AI fluency for every graduate. Add Goodyear's sustainable polymer hub in Akron and a statewide value of pairing speed with responsibility, and you have an engine for practical, large-scale innovation.

Grab a tactic and prove the ROI for yourself.

Pick any of these eight moves, run a small-team pilot this quarter, watch AI shift from theory to impact, and if you already have a proven AI use case, share it in the comments so we can build a shared playbook.