Tim Kreilkamp — sales meeting dossier

🔴 RED FLAG on Repete interface — must say in the room:

We have NOT verified that Repete's mill-automation system exposes an integration surface (API, webhook, file export, anything) that we can wire AI agents to from the outside. Repete is shipping a Generative AI Co-Pilot themselves in 2026, which suggests they're thinking about the interface layer — but their integration posture toward third-party AI vendors is unknown to us. We're willing to do that research and find out, but we cannot promise interoperability today. If Repete's system is closed, our value at FGF shifts to the office-side (orders, billing, consulting notes, patron portal) where Repete isn't operating anyway.

Inferred: Back-office ERP likely AGRIS (Cultura Technologies) or Cultura Feed-Mill Manager. Truck/scale tickets integrated to Repete; load-out billing automated downstream. Customer service remains phone-heavy for the small-farm long-tail.

Automation opportunities (specifics, not generic)

Production is the solved problem here. Leverage is office-side and customer-side — which sits outside Repete's domain regardless of how their interface shakes out:

1. Phone-order intake transcription. Most feed-mill orders still come by phone from farmers. AI-transcribed → structured-order intake with confirmation text-back to the farmer. Single highest-leverage feed-mill automation in 2026 per industry press.
2. Ingredient/inventory forecasting. Dairy consulting + custom mixes = high SKU complexity. AI demand forecasting reduces stockouts/overordering.
3. Patron-portal upgrades — conversational ordering, automated COA/batch-cert delivery, proactive low-bin alerts (BinSentry is the named-vendor incumbent here).
4. Dairy-consulting note capture. Field consultants do farm visits; notes get lost. AI-transcribed visit notes → CRM → automated follow-up.
5. Billing chase-down. Farmers pay on harvest/milk-check cadences, not net-30. AI-drafted harvest-cycle-aware AR sequences preserve relationships.
6. Patron onboarding — automated credit-application processing (currently a downloadable PDF).

Resistance / friction notes (farmers + phone staff don't trust AI lightly)

• Farmer tech distrust is real + load-bearing. A 65-year-old dairyman who's bought feed for 30 years will hang up on a chatbot. Every farmer-facing AI tool needs (1) a clear "you're talking to AI" identifier when AI is in the loop, (2) a human escape hatch on call one — a number that reaches a Kreilkamp staff person who knows them by name, (3) a "mill staff coffee-counter ready" briefing so staff can explain in 30 seconds what AI is doing.
• The mill staff hearing about it second-hand at the coffee counter is the failure mode. Talk to the people who answer the phones BEFORE you deploy anything that touches their workflow.
• Tim's Repete decision was not "let's try AI" — it was a $30M capex bet on industrial-grade automation with a proven vendor. His bar for new automation is "shows a clean ROI in production volume, labor, or compliance." Pair soft wins (chat) with hard operating-metric wins.
• John R. Kreilkamp may share authority on FGF spending — let Tim raise this; don't surface his father's full name unprompted.

---

Hotel + real estate — INTENTIONALLY OUT OF SCOPE

Per Luke's call: defer entirely. No frames, no vendor names, no pitch. If Tim raises hotel ops, the move is to listen, ask follow-ups, and schedule a follow-up conversation when there's actual context. Do not ad-lib hospitality expertise.

---

Cross-cutting signals — Tim's decision style

Acquisition pattern: buy-and-hold, sell when fit breaks. - 1935 founding → still family-owned - 2006 Farmers' Implement → held 20 years → sold Feb 2026 only because Johnson Tractor offered better customer service network - 2006 Farmers' Grain & Feed → still owned, expanded - March 2025 Gray Transportation → opportunistic acquisition when seller wound down

Not a financial-engineering family. They operate.

For a Relic engagement, this profiles to: - A scoped pilot inside ONE business (Farmers' Grain & Feed office-side is highest-conviction first move). Define a 60-day window. Define what "working" looks like in advance. Run it. Decide together. - Outcomes framed in customer-continuity language — not tech-feature language. - Don't pitch all his businesses at once. Pick one, win it, expand.

---

Industry examples of AI work done well (and where it's failed)

Worth referencing: - Standard Logistics × Optimal Dynamics — 70%+ of accepted loads auto-dispatched after AI evaluation. (Source: CCJ.) The kind of case study Tim would respect. - LoadAi (Optym) — driver-aware load matching factoring pay model, hours, freight requirements. - LoadStop — markets driver-retention impact via preference-aware planner. - BinSentry — IoT bin-level inventory for feed mills. - Repete's own Generative AI Co-Pilot for feed mills, shipping in 2026 — operators "chat with their data" to identify bottlenecks. Don't lead with this — see Repete-interface red flag above. But if Tim mentions it, we should be ready to talk about it.

Worth naming honestly: - Trucking + feed mill + hospitality are exactly the three verticals where 2025-2026 AI pilots have died loudest. Hallucinated load quotes. Farmer-facing chatbots that drove off long-tail patrons. Hotel-guest messaging that triggered chargebacks. If you don't surface these failure stories yourself, Tim will assume you haven't read the room. Bring them up early — "here's how this goes wrong if we're sloppy, and here's how we avoid that."

---

Discovery questions — what to ask, what to listen for

Opener (low-stakes, gets him talking)

• "Dan mentioned you've been thinking about automation across your operations. Where's the conversation in your own head right now — is it a specific pain point, or more an open exploration?"
• "What's working well right now that you wouldn't want us to disrupt?"

Trucking-specific

• "What's the single most expensive 'where's my load?' conversation your dispatch team has each week?"
• "How do you find new drivers today, and where does that pipeline break?" (bridges to the academy investment)
• "What does your TMS not do for you that you wish it would?"
• "When a reefer goes out of spec, walk me through the next 24 hours."
• "Who's your hardest shipper to get paid by, and what makes it hard?"

Feed mill-specific

• "Orders coming in by phone right now — what's the volume on a typical day, and who's catching them?"
• "On the patron portal — how many of your customers actually use it, and what's it not doing for them?"
• "Your dairy nutrition consultants — how do their visit notes flow back into the mill's records?"
• "Are you happy with how Repete's automation has paid off? Where's it left you wanting more?" (surfaces the office-side white space without criticizing Repete)
• "Anyone trying to sell you BinSentry or similar IoT-bin tooling? What's stopping you from saying yes?"

Decision-style + relationship questions

• "When you made the call on Repete in 2021, what convinced you?"
• "Who else weighs in when you're making a 12-month commitment of this size?" (lets Tim raise John if he wants to; don't name him first)
• "What's a vendor relationship that's worked well for you over years? What made it work?"
• "What's a vendor experience that left a bad taste? What broke?"

Closing-shape questions

• "If we ran a 60-day pilot on one slice of one business, which slice would you want us to prove ourselves on?"
• "Who else on your team should be in the next conversation?"
• "What would a 'this is working' signal look like to you 90 days in?"

Listen for (red flags + buy signals)

• Red flag: he names a previous AI vendor with frustration → ask what specifically broke, don't repeat
• Red flag: he can't articulate where AI would help → he's intellectually curious but not problem-pulled; pilot scope needs to be even smaller
• Red flag: he defers to John on every question → align on a shared meeting with John before signing anything
• Buy signal: he name-drops Repete's GenAI co-pilot or BinSentry → he's been doing his own reading
• Buy signal: he asks about pilot structure or pricing → he's already evaluating fit
• Buy signal: he tells you a specific number ("my dispatch team takes 30 calls a day from shippers asking 'where's my load'") → he's letting you scope; capture it precisely

---

Service-shape frames — how OBB serves Tim

Frame 1: Pilot-first, scope-defined

60-day pilot inside one business line, fixed scope, fixed price, defined success metrics. Not a retainer. Not a hand-shake about "let's see how it goes." A specific deliverable with a number attached.

Highest-conviction first pilot: Farmers' Grain & Feed office-side. Specifically: - AI-assisted phone-order capture for inbound orders - Patron portal upgrades that don't break the 65-year-old dairyman experience - Dairy-consulting note capture pipeline

Why FGF first: - Production-side is solved (Repete). Office side is real white space. - Lower stakes than touching driver/dispatcher workflows. - Faster to show a result than reefer-claim-packet work that depends on rare events.

Fee shape: scoped fixed-price pilot. Number lives in Luke's head; doesn't lead the conversation.

Frame 2: Build vs. buy

• SaaS platforms (TMS, PMS, CRM, mill ERP) — OBB does NOT build from scratch. Tim's businesses already have these. We integrate.
• Custom AI agents — this is what OBB builds. Bespoke automation tuned to specific workflows.
• Websites / marketing sites — can build (Pages-deployed). Not the highest leverage; defer unless Tim has a specific need.
• API integrations — the glue between AI agents and existing systems (Samsara, Repete-if-interfaceable, AgriCharts, McLeod/TMS).

Frame 3: Anti-disruption posture for farmer- and driver-facing tools

Every patron- or driver-facing AI tool ships with: 1. A clear "you're talking to AI" identifier when AI is in the loop. 2. A human escape hatch on call one — a number that reaches a Kreilkamp staff person who knows them by name. 3. A "mill staff coffee-counter ready" briefing so the people who answer phones can explain it.

This is the OBB differentiator. Most AI shops don't think about the 30-year farmer relationship. OBB does because we operate in similar relationship-heavy contexts already.

---

OBB credibility positioning — honest, not glaze

The frame

OBB Holdings runs AI agents across multiple businesses right now, for ourselves: - A talent/alumni-development practice with multi-audience report generation and annual-calendar tooling. - A Christian high school's operations work — live with 175+ students, automated progress reports, attendance, AI-detection support for teachers, ACT curriculum + practice materials. - An ERC appeals practice with a dedicated AI agent for client intake + qualification drafting. Live, paying clients.

This is not theoretical AI consulting. It's the same shape Tim runs — multi-business operator using AI to compound. We use what we sell.

The honest acknowledgment Tim deserves

• We are three partners, not a Fortune 500 consultancy. That's a feature for an operator like Tim, not a bug — but Luke has to land it as a feature, not apologize for it.
• We have not yet built for trucking or feed mill specifically. What we have built is the underlying agent architecture — the same primitives that read documents, interface with APIs, draft tone-appropriate communications, escalate to humans when warranted. Those primitives apply across verticals; the domain knowledge gets layered on with operator partnership.
• We do not know Repete's interface posture yet. We're willing to find out as part of pilot scoping. We can't promise integration until we've done the research.
• Truckers, farmers, and the front-desk staff who take their calls are professionally skeptical of AI. OBB knows this and designs around it. The first thing we do on any farmer- or driver-facing tool is meet the people who'd be on the other end.

What this lets Luke say in the room

"OBB Holdings is three partners. We started building AI agents to run our own work across a talent-development practice, a Christian school's operations, and an ERC-appeals firm. We discovered that nobody we hired to build this stuff understood our businesses as well as we did, so we built it ourselves. Now a small set of operators hire us to do the same for them. We're picky about who we work with because depth is the product — coverage isn't. Dan introduced us because he thought you'd find that interesting. I'm here to listen first and see if there's a piece of your operation where we'd be worth a 60-day pilot. If we are, we scope it tight, run it, and decide from results. If we're not, no harm done."

What NOT to say

• Don't list every tool stack item ("we use Cloudflare, OpenAI, Anthropic..."). That's IT-vendor talk.
• Don't promise enterprise references we don't have.
• Don't position against bigger AI consultancies ("unlike McKinsey..."). Tim doesn't care about McKinsey.
• Don't name Dan as a commission partner. The 15% is silent in Tim's presence.
• Don't pitch the retainer.
• Don't ad-lib on Repete's interface options. ("We're willing to research that as part of pilot scoping" is the answer.)
• Don't surface John R. Kreilkamp's name unless Tim raises it.

The "hit by a bus" question

Tim may ask: "What happens if one of you three gets hit by a bus?" The honest answer: - We're three partners, not three solo operators — the work runs against shared infrastructure and documented patterns, not in one person's head. - For the pilot scope we're discussing, the bus factor is manageable. The pilot has a defined end date; we're not asking Tim to depend on us for years before we've earned that. - If we earn a longer engagement and the bus-factor question becomes load-bearing, that's the conversation we have together — building a small bench, formalizing succession, or sizing the engagement to OBB's actual capacity. We don't pretend the question doesn't exist.

---

Recommendation on Dan's role in the meeting

Dan should NOT attend. His introduction is the on-ramp. The working conversation is Luke ↔ Tim, alone (or with whoever Tim brings from his side). Reasons: - Tim's buy-and-hold operator profile rewards direct relationships. A third-party introducer in the room signals "this is a referral pipeline" rather than "this is a partner conversation." - The 15% commission is silent. Dan's presence in the room makes it harder to keep silent. - Tim may have unspoken context about Dan he won't share with Dan in the room. Without Dan there, Tim talks more honestly about what he actually wants.

Dan's commission stays. The partnership structure stays as agreed. Dan just doesn't sit at the table.

---

Open items to verify in the meeting

1. Which TMS does Kreilkamp Trucking run? McLeod is the industry default; not verified.
2. Customer-side shipment visibility — project44 / FourKites / portal / EDI only?
3. Farmers' Grain & Feed's office-side ERP — AGRIS / Cultura / Excel / custom?
4. Repete's third-party integration posture — the load-bearing red flag. Can we actually talk to their system?
5. Brent Redmond Transportation's tech stack and integration depth with Allenton operation.
6. Whether Tim has a stated AI/automation posture — no public quotes found.
7. Farmers' Implement sale proceeds — capital from the Johnson Tractor deal (eff. June 1, 2026) may be deployable; ask delicately, late in the conversation.
8. John R. Kreilkamp's role in approval flow — only if Tim raises it.

---

Compiled from FreightWaves, CCJ, gmtoday, Mid-West Farm Report, Discover Wisconsin, Repete Corporation, Feed & Grain, Moraine Park Technical College press, RocketReach, Dun & Bradstreet, WB Warehousing & Logistics, Brent Redmond Transportation, Kreilkamp.com, fgfmill.com.

---

Generated 2026-05-14 by OBB Holdings LLC / Relic313 dossier system (revised). For internal use.