What are the key takeaways from this Practical AI episode?

Key insights include: **Workflow Mechanism Selection:** Assign the simplest mechanism that works to each individual task, not each project. LLMs should explain, route, or translate — not act as databases, decision-makers, or self-checkers. Stanford's surgical co-management tool restricted the LLM to three outputs — yes, maybe, or no — eliminating hallucination risk entirely while screening next-day elective surgeries.; **Guardrails vs. Prompt Rules:** A rule written inside a prompt is not deterministically enforced. Guardrails require actual code-level machinery — gates, checks, confidence thresholds, or human escalation paths. Design every workflow step with a recovery mechanism: roll back outputs, retract writes, or require pre-approval for irreversible actions before they execute.; **Data Quality as a Ceiling:** Fix data quality before building any AI workflow, because every downstream task inherits its limitations. The relevant question is not whether data is generally good, but whether it is accurate, complete, consistent, and fresh enough for the specific task. Stale data produces silent errors, not flagged failures.

What did Heidi Porter discuss on Practical AI?

Three speakers at the Midwest AI Summit share frameworks for building durable AI workflows in production environments. Heidi Porter covers data grounding and guardrails, Kevin Solorio addresses adoption with blue collar workers, and Chris Shinkle with Andrew Dunlap present a constraint-based model for determining how much autonomy AI coding agents should have. Key topics include: **Workflow Mechanism Selection:** Assign the simplest mechanism that works to each individual task, not each project. LLMs should explain, route, or translate — not act as databases, decision-makers, or self-checkers. Stanford's surgical co-management tool restricted the LLM to three outputs — yes, maybe, or no — eliminating hallucination risk entirely while screening next-day elective surgeries.; **Guardrails vs. Prompt Rules:** A rule written inside a prompt is not deterministically enforced. Guardrails require actual code-level machinery — gates, checks, confidence thresholds, or human escalation paths. Design every workflow step with a recovery mechanism: roll back outputs, retract writes, or require pre-approval for irreversible actions before they execute..

How long is this episode of Practical AI?

This episode is 54 minutes long. SignalCast provides an AI-generated summary so you can get the key insights in about 3 minutes.

Practical AI

Abundant Intelligence at the Midwest AI Summit

October 29, 2026

54 min episode · 2 min read

Heidi Porter

Episode

54 min

Read time

2 min

Topics

Health & Wellness, Investing, Fundraising & VC

AI-Generated Summary

Published Oct 29, 2026

Key Takeaways

✓Workflow Mechanism Selection: Assign the simplest mechanism that works to each individual task, not each project. LLMs should explain, route, or translate — not act as databases, decision-makers, or self-checkers. Stanford's surgical co-management tool restricted the LLM to three outputs — yes, maybe, or no — eliminating hallucination risk entirely while screening next-day elective surgeries.
✓Guardrails vs. Prompt Rules: A rule written inside a prompt is not deterministically enforced. Guardrails require actual code-level machinery — gates, checks, confidence thresholds, or human escalation paths. Design every workflow step with a recovery mechanism: roll back outputs, retract writes, or require pre-approval for irreversible actions before they execute.
✓Data Quality as a Ceiling: Fix data quality before building any AI workflow, because every downstream task inherits its limitations. The relevant question is not whether data is generally good, but whether it is accurate, complete, consistent, and fresh enough for the specific task. Stale data produces silent errors, not flagged failures.
✓Agent Adoption with Frontline Workers: Build AI tools with workers, not for them, by starting with a deliberately imperfect prototype and inviting correction. Workers who identify errors early and see changes implemented become product champions through the IKEA effect. Establish a data flywheel — offline evaluation datasets updated continuously from online trace sampling — to capture edge cases post-deployment.
✓Constraint-Based Agent Autonomy: Six constraints — token budget, organizational readiness, goal definition, domain knowledge, verifiability, and code oversight — determine which of four operating modes fits a team: individual, shared, collaborative, or fully autonomous. SEP's Forum credit union mobile banking project, with no token budget limits and clear requirements, achieved a six-times reduction in development effort using a multi-agent workflow.

What It Covers

Three speakers at the Midwest AI Summit share frameworks for building durable AI workflows in production environments. Heidi Porter covers data grounding and guardrails, Kevin Solorio addresses adoption with blue collar workers, and Chris Shinkle with Andrew Dunlap present a constraint-based model for determining how much autonomy AI coding agents should have.

Key Questions Answered

•Workflow Mechanism Selection: Assign the simplest mechanism that works to each individual task, not each project. LLMs should explain, route, or translate — not act as databases, decision-makers, or self-checkers. Stanford's surgical co-management tool restricted the LLM to three outputs — yes, maybe, or no — eliminating hallucination risk entirely while screening next-day elective surgeries.
•Guardrails vs. Prompt Rules: A rule written inside a prompt is not deterministically enforced. Guardrails require actual code-level machinery — gates, checks, confidence thresholds, or human escalation paths. Design every workflow step with a recovery mechanism: roll back outputs, retract writes, or require pre-approval for irreversible actions before they execute.
•Data Quality as a Ceiling: Fix data quality before building any AI workflow, because every downstream task inherits its limitations. The relevant question is not whether data is generally good, but whether it is accurate, complete, consistent, and fresh enough for the specific task. Stale data produces silent errors, not flagged failures.
•Agent Adoption with Frontline Workers: Build AI tools with workers, not for them, by starting with a deliberately imperfect prototype and inviting correction. Workers who identify errors early and see changes implemented become product champions through the IKEA effect. Establish a data flywheel — offline evaluation datasets updated continuously from online trace sampling — to capture edge cases post-deployment.
•Constraint-Based Agent Autonomy: Six constraints — token budget, organizational readiness, goal definition, domain knowledge, verifiability, and code oversight — determine which of four operating modes fits a team: individual, shared, collaborative, or fully autonomous. SEP's Forum credit union mobile banking project, with no token budget limits and clear requirements, achieved a six-times reduction in development effort using a multi-agent workflow.

Notable Moment

Stanford Health Care's pre-AI triage process missed half of patients eligible for surgical co-management. After deploying an LLM-based screening tool, analysis of false negatives revealed the errors were largely in existing human processes, not the model — prompting broader workflow improvements beyond the original AI scope.

Know someone who'd find this useful?

Episode Transcript

Narrator: Welcome to the Practical AI Podcast, where we break down the real world applications of artificial intelligence and how it's shaping the way we live, work, and create. Our goal is to help make AI technology practical, productive, and accessible to everyone. Whether you're a developer, business leader, or just curious about the tech behind the buzz, you're in the right place. Be sure to connect with us on LinkedIn, X, or Blue Sky to stay up to date with episode drops, behind the scenes content, and AI insights. You can learn more at practicalai.fm. Now, onto the show. Daniel: I had the privilege of being the emcee at this year's second annual Midwest AI Summit, and it was amazing to see everyone there at the summit just engage in a really useful practical way around subjects like AI agents and decision models and AI security. And I arranged with the conference organizers for us to be able to share a few snippets from speakers at the Midwest AI Summit. So today's episode is gonna be a very special one where you'll get to hear a selection of a few snippets from speakers at the Midwest AI Summit. We'll have Heidi Porter, who is a principal at Moyer Group AI, and then Chris Shinkle, Andrew Dunlap, who are with SEP, and then Kevin Solorio, who is a staff software engineer at Focus. I hope you enjoy the content. I know I certainly did. And, yeah, would love to see you at next year's Midwest AI Summit. We'll be sure to update our audience here when the details of that are shared from the conference organizers. So enjoy this selection of a few talks from the Midwest AI Summit. We have with us Heidi Porter. She's a tech technology professional with decades of experience in networking, data analytics, machine learning, AI, and on premise cloud deployments at large companies. Her company, Moyer Group AI, provides education, strategy, and consulting on AI tools and where it belongs in your tasks and workflows. And it's those workflows that she's gonna talk about how to make, durable AI augmented workflows. Help me welcome Heidi Porter. Heidi: Alright. Thank you. Got the clicker here. Good morning. I'm Heidi Porter. Quick show of hands. Who has had an AI pilot that did not make it past the demo stage or stalled at the start of production. So there's a good amount, and congratulations to all the rest of you who just, you know, hit it on the first time. Right? So a lot of times these workflows fail because... It's not because the model was the problem. It's because the LM was asked to do too much. And Don and Jeff just gave great talks about when to say no to AI and when to say yes to AI. And so sometimes people ask it to hold the data, be the system of record, it's not a database, make the decision, and check its own …

Get the full transcript (9,459 words) + summary by email — free

One-time email with the complete transcript and AI summary of this episode. No account needed.

One email, no spam. We’ll also show you what SignalCast does.

Browse all Practical AI transcripts →