Tim Kreilkamp — engagement dossier (v9)

For: Luke Olson — go-forward reference for the Kreilkamp engagement.
Compiled: v9 — 2026-05-28. Rewritten against the 2026-05-27 sit-down (full 2h21m meeting recording + Luke's corrections). Replaces v8 (2026-05-27, pre-meeting prep).
Source: The meeting recording + synthesis (obb-kreilkamp-debrief-2026-05-27.pages.dev) + public sources + Dan's earlier intel. Internal-only — never shared with Tim or any non-OBB party.

What changed from v8 → v9. v8 was a pre-meeting prep doc — opening scripts, pivot ladders, "what to say in the room." The meeting happened, so this is now the engagement reference: corrected facts, what was decided, what's still open, and what we now actually know about Tim's people, pipeline, and businesses. The pre-meeting tactical scaffolding has been retired. Where the meeting corrected a prior assumption, it's flagged CORRECTED.

1. Corrections the meeting forced (read first)

CORRECTED The driver ATS is Tenstreet, not iRecruit. When Luke said "Which is iRecruit?" Tim cut in: "If we're talking about drivers, we don't use iRecruit for drivers." Jeff: "It's Tenstreet." (00:12:09) iRecruit IS used at Kreilkamp Holdings — for non-driver roles at the other companies (brokerages, warehouse, feed mill). For the driver pilot, iRecruit is irrelevant. The entire "iRecruit write-back" integration plan from v8 is dead. Tenstreet is the integration target. Tenstreet has its own API and already does basic AI fit-scoring.

CORRECTED The scope is 150-200 hires/year, not "40 drivers." The "hire 40 tomorrow" line was a real signal, but it's the current gap (40 empty trucks today), not the job. To hold the line against attrition, Kreilkamp must hire 150-200 drivers/year, every year (00:34:34). The pilot is a sustained attrition-replacement engine, not a one-time gap-fill. This is the single most important reframe — it makes the "10 qualified leads every 40 days, forever" cadence the actual product.

CORRECTED The "$150K" in the room is driver pay, not a pricing anchor. Tim said it twice (00:51:58, 01:36:32), both times about driver compensation — "I wish we could pay them $150K, economics don't allow it." Do NOT treat it as a ceiling for OBB's price. Any internal note that used it that way is wrong.

CORRECTED The pricing ladder is gone. v8 walked in with a $65K → $50K → $45K milestone ladder. In the room, Luke voluntarily backed off $50K to Tim and Jill directly (01:49:18): "there may not be enough there for a fifty-thousand-dollar pilot… I don't think it should cost you fifty thousand dollars." Then Jill steered him lower (the 401k analogy, below). The win-benchmark number is now an open loop Luke owns — see §6.

CORRECTED Who was actually in the room. Tim, Jeff Swift (recruiter), Jill (CFO), and Josh (operations). No "Vince," no "Joe" — both were whisper transcription mishears. Jeff left around 01:46; the entire pricing/partnership/succession conversation (01:47–02:21) was Luke + Tim + Jill — principal and dollars-gate, both present.

CORRECTED Fleet numbers + the Iowa "acquisition." Operational reality from the room: ~285 trucks, ~800 trailers, ~275 available, 140 currently producing revenue, 180 is the benchmark, ~40 empty today. (The "300+ trucks / 1,000+ trailers / 18M miles" figures are the website's marketing numbers.) And the Iowa "Gray Transportation acquisition" v8 cited was not a clean acquisition — per Tim it was effectively a brokerage where they took the drivers and customers. Luke named this AI-research error to Tim's face in the room; Tim was gracious. Worth a self-audit pass on any remaining "rosier than reality" claim.

2. The pipeline as it actually works (Jeff's funnel)

This is the real funnel, from Jeff's walkthrough (00:10–00:55). It replaces every guessed funnel detail in v8.

Lead vendors: CDL Life, Trucker's Report, Drive My Way + multi-carrier leads. Spend mix ~25% / 50% / 25%. Two of four do prescreening text-drips on Jeff's behalf; two of four run call-center "matchmakers" for no-response re-engagement (Jeff does NOT want OBB to be another outsourced matchmaker — "I don't think we're going to use that").
The numbers (April): 1,200–1,400 leads/mo → 1,924 outreach attempts → 192 phone interviews scheduled (~10% of attempts) → ~200 phone screens executed → 75 moved forward → ~12–14 hires. ~1% lead-to-hire. 592 "no response" in April (~31%) — the leak is reach, not lead volume. They could buy more leads but can't reach the ones they have.
Tenstreet = the ATS. Receives vendor + direct leads, hosts the full application (incl. FMCSA fields), does AI fit-scoring (experience, accidents, violations — green-check flags). Has Pulse (in-app messaging on driver tablets — NOT SMS). Tenstreet has no SMS for Kreilkamp ("we don't have it through there," 00:58:57).
Calendly = self-scheduling. Sends reminder texts/emails, writes booking back into Tenstreet. This is where the current text touches actually come from — not Tenstreet.
Downstream: 30-min phone screen → MVR + Clearinghouse + PSP → Safety dept gates yes/no → offer → background + drug test (72-hr window, no re-do) → 2-day orientation in WI. Experienced drivers (3+ mo) test out and can earn revenue the Friday after orientation; inexperienced go to a 4–6 week finishing program ("cost difference bigger than this building"). Experienced hires are the golden ticket — Tim + Josh both stressed it.
VoIP recruiting line: 1-800-999-7112 (separate from the general 1-800 painted on trucks). No call instrumentation today — nobody knows how many calls hit which number ("IT could tell us"). Foundational metric for sizing the missed-call rescue.
Recruiting team: Jeff + Amy + Nicole, with a third recruiter being added (confirm hired vs pending — affects pilot KPI baseline).

3. The pilot as now scoped (post-discovery)

Discovery in the room sharpened the pilot to two connected pieces, both validated by Jeff and broadened by Tim:

Missed-call text-rescue — every missed call, not just after-hours. The moment a call is missed (or a voicemail ends), an automatic SMS goes out with a mobile-friendly web form. Tim spontaneously broadened this from after-hours-only to every missed call (00:46): "we miss them in business hours too." The ROI headline is Tenstreet's own data: 30–40% of missed calls go to the next carrier on the driver's list — defensible because it's their number, not ours.
Pre-screen compression with calendared callback (Paradox-style). Text-first capture of a small set of decision fields, then a scheduled callback. Jeff validated the design constraints himself: fewer than 5 questions, under ~50 characters each, and do NOT use "years of experience" as a gate — experienced drivers self-select out of long forms. Jeff ran Paradox at a prior job doing exactly this and is positively disposed.

Integration target = Tenstreet API (write SMS/voice-intake outcomes back onto the contact record) + the VoIP carrier API (missed-call detection to trigger the text-rescue). NOT iRecruit. Data should ultimately live in Tenstreet — Jeff: "Tenstreet ideally, that's the only place."

The two-demographic insight (validated): experienced drivers distrust AI and dislike texting; younger drivers prefer it. The design has to serve both — text-first for the mobile-native majority, with a fast human handoff that never makes an experienced driver feel handled by a bot.

Held firmly out of scope (Tim's own stance): full phone-screen automation. Tim: "it has to be done by a person — I get AI broker calls and I hang up on them." This is the OBB "AI behind staff, never in front of your people" principle, now reinforced by Tim himself. AI does intake + routing + drafting + scheduling; the recruiter does every substantive conversation.

4. Tenstreet integration reality

Full technical findings: kreilkamp-tenstreet-research report. The load-bearing points:

Verdict: doable-but-gated. Tenstreet has a real working integration API — but it's XML-over-HTTP / SOAP, not REST/OAuth. Auth = Tenstreet-issued API key + Company ID + a pre-approved "source" string. Access is request-gated (email integrations@tenstreet.com) + a paid activation (~$500 reported, ~$100–200/mo — third-party numbers, do not bake into margin).
It does the two things we need: POST-on-subject-update to a listener URL of ours (the webhook-equivalent for new leads), and inbound write-back of our pre-screen results as custom questions / tags / display fields keyed to the Subject ID. Field IDs come from Tenstreet support.
The SMS-seam risk is LOWER for Kreilkamp than the generic research warned. The research flagged Tenstreet's native SMS product as a dual-texting collision risk — but Kreilkamp doesn't use Tenstreet SMS (Jeff: "we don't have it through there"). Their texts come from Calendly + lead-vendor drips. So our Twilio layer doesn't collide with a Tenstreet texting thread. Still confirm opt-out/STOP handling so we don't cross the Calendly touches.
Timeline: ~3–5 weeks to a working loop — the slip risk is Tenstreet's approval queue and field-ID iteration, not code.
Don't over-promise Tenstreet's fit-score export (marketed, not confirmed API-exportable).

5. The people (who actually matters)

Replaces v8's public-source name guesses (Danielle Berndt / Bob Christie) with the people from the room and the names that carry weight.

In the room

Tim Kreilkamp — President. Fourth generation. Decision authority + the relationship gate. "I'm an operator, not a salesman." Plain-spoken, profane, folksy ("apple pie in the sky," "knife without a handle"), prone to macro tangents (white-collar AI displacement, immigration + trucking rates, succession). Re-asks the same question several ways. The "chutzpah" in the business is his. See §7 for full profile.
Jeff Swift — Recruiter / pilot end-user. The single most important operational ally. Knows his numbers cold, technically literate (runs Claude Code on a Mac mini at home off old Ethereum mining rigs; has an agent that answers his voicemails + emails), positively disposed to conversational-AI (ran Paradox before). Enthusiastic and unthreatened — joked "why aren't you doing this, why am I here?" Recruits for the other Kreilkamp companies too, not just trucking. His calendar is what the automation touches. Win Jeff = win the build.
Jill — CFO ("queen of money"). The dollars-gate. Quiet but every utterance is high-signal. She explicitly steered Luke to come in low (the 401k analogy, §6) and raised the NDA/protection concern. Their vendor culture is hourly/flat-fee — outcome pricing is novel to her, so it must be translated into hourly-equivalents she can compare.
Josh — Operations. Knows fleet ops numbers fluently (the 285-trucks/140-producing/40-empty data is his). Aligned with Tim that experienced drivers are the priority. One of two people who deal with "Matt."

Named but not in the room

Amy & Nicole — Jeff's recruiter teammates. A third recruiter is being added.
Matt Hauncey — Power BI consultant from Green Bay ("works for a church"); Tim meets him weekly, had a 4pm with him that day.
Monica — controller; named in the succession context.
John Shiz — younger manager named in the succession context.
Danielle — uses Copilot alongside Tim.
HR person retiring after 35 years — still does the 401k census manually (a named Phase-2 automation candidate Tim floated).
Tim's son — works at the feed mill (FGF). Part of the fifth-generation succession picture.

Note: former VP Logistics Jim Jansen (in older public sources) passed away April 2023 — do not reference.

6. Pricing — where it landed + the open loop

OPEN LOOP Luke owes Tim + Jill a win-benchmark number. He asked them to name it (02:09:54): "I would be helped if you know the number that makes sure everybody is happy at the win benchmark." Neither named a figure — they deflected. Luke owns coming back with it.

What was actually said in the room

Luke walked off $50K directly to Tim + Jill (01:49:18). He chose the durable-relationship play over the one-time win, reading the room with the principal and CFO both watching — a strong move, not a weak one.
Jill's 401k-advisor analogy (02:04:19) is the clearest pricing signal of the meeting: a 401k advisor who charges significantly less than the usual 1% built a huge book (her mom, her neighbor now invest with him) and makes far more than he would have at 1% with a few clients. Translation: come in low, build the relationship, expand through trust. Tim was present and silent — and given their decision shape ("we usually talk through it, never had a split vote"), Tim's silence is tacit endorsement.
Vendor culture is hourly/flat-fee, not outcome-based (Jill: lawyer hourly, insurance premium). Luke's outcome-tied pitch is novel to them — it has to be expressed in terms they can compare to an hourly bill.

The shape to bring back

Per Jill's steer + Luke's read: low, properly-priced pilot setup + a durable, outcome-tied monthly residual that only grows as Kreilkamp wins. Luke's own words: "I would so much rather get the pricing on the durable down the road every month, tied to the actual deliverable… the test needs to be properly priced so that it wins, and it wins at a price everyone's excited about." The cadence product — 10 qualified warm leads every 40 days, sustained — is what the residual is tied to. (Detailed scope + pricing structure lives in the obb-kreilkamp-scope-pricing deliverable and the partnership-vs-vendor one-pager obb-incentive-alignment; this section is the record of what was said + the open loop.)

ACTION NDA: Tim asked Luke to sign one (02:10:51) — he's been burned before ("we beta things, pay for them, then they make it part of their master"). Luke committed in the room. Small contract action item.

7. Tim the operator (validated + refined)

The meeting confirmed most of v8's operator read and added sharper signals:

Conservative on commitment — won't sign what he can't defend. The clearest tell: his ERC abstention. "I did not do it — nothing, no, not one government penny. I couldn't sign on the bottom line… we can hang our heads very high." (01:38) He left money on the table to avoid risk exposure. This is the temperament he will apply to the pilot. Price and scope it so there's nothing for him to feel exposed on.
Buy-and-hold, relationship-first. "You're the home team, in our backyard, a friend of Dan Weber's, connected to a feed mill." He maps partners into his own network. He rewards direct relationships (why Dan isn't at the table).
Patron-facing AI is a hard no — his line, not just ours. "I hang up on AI brokers." Reinforces the OBB principle; never propose anything that puts AI in front of his drivers, shippers, or customers.
The "wake up at 2am" obsession: a predictive-at-fault-accident agent off Samsara/Motive data. He loved Luke's "actuarial, not predictive" framing. But he raised the killer objection himself: "we have to be careful collecting too much information because that's discoverable… unless we figure out how to legally store that." (01:11) Luke deferred well ("knife without a handle"). OPEN Luke should name it affirmatively in any proposal — "the accident-prediction model is on the roadmap; it needs a data-discoverability posture before we touch it" — so Tim feels deferred-to, not deferred-from.
Decision process: "we talk through it, we don't split, I don't think we've ever had a split vote." Consensus shop. Tim's silence on a question = endorsement.
Demonstrated automation buyer: Repete-automated the ~$30M FGF feed mill (opened Nov 2021), Samsara fleet-wide, Power BI across all companies. Not in "convince me AI works" mode.

STRUCTURAL RISK — succession. Tim disclosed (02:16): fourth generation, fifth "online," succession-planning underway (John Shiz, Monica, his son at the mill), "probably between zero and five years" to handoff. Luke flagged the risk that the next generation may "play not to lose" rather than "play to win" — and Tim agreed the chutzpah is in him, not guaranteed downstream. Implication: the long-tail recurring revenue could evaporate at a generational handoff. Front-load durable value capture; don't price the pilot assuming a decade of residual.

8. The fleet of businesses (Tim's portfolio, as disclosed)

Tim opened up his holdco in the post-Jeff conversation. At least 5–7 distinct businesses:

Kreilkamp Trucking (Allenton WI; the pilot's home) — founded 1935, family-owned dry van + reefer + flatbed.
Brent Redmond Transportation (a second trucking company, California / cold-chain — "Brett Redman" in the transcript).
Two brokerages.
The Iowa "thing" — not a clean acquisition; effectively a brokerage where they took the drivers + customers. (This is the fact Luke's AI got wrong and he corrected in person.)
WB Warehousing & Logistics — warehousing footprint.
Farmers' Grain & Feed (FGF) — the Repete-automated feed mill; Tim's son works there. Out of pilot scope; informs Tim's automation-buyer profile.
The 401k / HR back-office — very low turnover (unlike drivers); the retiring 35-year HR person does the census by hand.
Greenfield mentioned (likely another entity — confirm).

Phase-2/3 expansion lanes Tim himself floated: warehouse inventory automation (license-plate + location scanning), 401k census automation, income-statement automation. But the model is depth, not breadth: Luke's frame — "I want all my clients to fit on this hand; I'd rather do five or six deep than expand." Tim's frame — "I want this to be ours, not to go to competitors." Both want exclusivity + depth. Phase 2 is the same trucking lane (ongoing-ops + the next automation Tim points at), not a land-grab across his portfolio.

9. Open loops Luke owns

OPEN Win-benchmark price number — deflected in room; Luke comes back with it (§6).
DONE Tenstreet API research — committed in room, completed (§4). Next step is the pre-scope ticket to integrations@tenstreet.com to confirm pricing/tier, new-lead event granularity, custom-question field IDs, update-by-Subject-ID.
OPEN VoIP recruiting-line call volume — "IT could tell us." Needs a follow-up email ask; it sizes the missed-call rescue ROI.
OPEN NDA — Luke committed; execute it.
OPEN Motive vs Samsara — Motive is on trial for the whole fleet and likely to win. Don't bake Samsara into any Phase-2 plan; assume Motive is the platform ~6 months out.
OPEN Data-discoverability posture — a one-paragraph answer ready before any data-aggregating Phase-2 conversation (Tim's own objection).
OPEN Third recruiter — hired or pending? Affects the pilot KPI baseline.
OPEN Dossier self-audit — scrub any remaining AI "rosier than reality" claim (the Iowa error is the warning).

10. Hold-the-line items (going forward)

No access to the books or anything we can't afford to make a mistake on. Tim raised financial access in the original conversation; Luke declined. Trust scales access incrementally. If re-asked at month 4 ("can you handle payroll automation?"), it's "a different engagement with different governance — we can scope it separately."
No patron-facing AI, ever. Now doubly reinforced by Tim's own "I hang up on AI brokers." Not an upgrade path; a hard line.
$150K is driver pay, not our price. Don't anchor on it.
Don't oversell capability. We have not built for trucking driver-hiring specifically; we have the underlying agent architecture (intake, API interfacing, tone-appropriate drafting, human escalation). Domain knowledge layers on with operator partnership. The Iowa correction proves Tim values candor over polish.
Bus-factor honesty. Three partners on shared infra + documented patterns; for a defined-end pilot the bus factor is manageable; if a longer engagement makes it load-bearing, that's a conversation we have openly (a Kreilkamp-specialist agent that documents process + state, or re-papering). Don't sell a continuity story we can't honor — and don't name Petey/Kevin as trucking backup (Petey's a counselor; Kevin's scope doesn't cover trucking).

11. Glossary (corrected)

Use these so the term can be dropped without overreaching. If Tim throws one back that isn't here, "say more about that" — never bluff.

Driver-hiring stack (the pilot's actual tools)

Tenstreet = Kreilkamp's driver ATS (applicant tracking system). Receives leads, hosts the full application incl. FMCSA fields, does basic AI fit-scoring. The pilot's integration target. (Corrects v8, which named iRecruit.)
iRecruit = an ATS used at Kreilkamp Holdings for non-driver roles at the other companies. NOT part of the driver pilot.
Pulse = Tenstreet's in-app messaging feature (driver tablets). NOT SMS. Used for drip campaigns + recruiter messages inside Tenstreet.
Calendly = the self-scheduling tool that currently sends the text/email touches and writes bookings back into Tenstreet.
Lead vendors = CDL Life, Trucker's Report, Drive My Way + multi-carrier sources that feed leads into Tenstreet.
PSP = Pre-employment Screening Program (FMCSA crash + inspection history). Checked after the phone screen.
MVR = Motor Vehicle Record. Clearinghouse = FMCSA Drug & Alcohol Clearinghouse (federal D&A violation registry — checked before hire).
Fit score = Tenstreet's existing applicant scoring (experience, accidents, violations → green-check flags).

Fleet + trucking

Samsara = current fleet telematics (GPS, hours, fuel, dashcam, reefer temps). Motive = the competitor on trial to replace Samsara fleet-wide; has AI coaching + in-cab voice alerts. Assume Motive is the platform ~6 months out.
FMCSA = Federal Motor Carrier Safety Administration. HOS = Hours of Service. CDL-A = Class A commercial license (tractor-trailer). DOT physical = DOT medical certification.
TMS = Transportation Management System (dispatch/load/billing software; McLeod, TMW are the big vendors). Kreilkamp's specific TMS still unconfirmed.
Power BI = Microsoft BI dashboards Tim runs across all companies ("Tim's report," rate-per-mile, etc.). Surfaces data but doesn't act on it — the act-on-data gap is the automation opportunity. Push-dataset = the native real-time streaming pattern into Power BI.

Feed mill (reference; out of pilot scope)

Repete = FGF's feed-mill automation vendor (wired the ~$30M Allenton mill). No public API; FLX control platform runs on MS SQL Server on Tim's hardware; integration is paid services via Repete's CNX Data Integrator. Don't promise direct Repete integration.

AI / tech

Agent = a custom AI workflow that does specific business work, routes to humans when stuck, integrates with tools — not "a chatbot."
Confabulation = an AI stating something false with confidence. Luke seeded this term with Tim as the reason humans stay in the loop.
API = how software talks to software. OAuth = scoped permission to access another app's data without sharing a password. Opus 4.7 = the model Luke named as the current stack.

12. Bottom line

The meeting went better than well. The room is bought in conceptually, Tim widened the scope himself (every missed call), and Jeff is a technically-literate ally rather than a gatekeeper. Three things shifted under our feet — Tenstreet replaces iRecruit, the real target is 150–200 hires/year, and pricing collapsed in-room with both the principal and the CFO steering Luke lower. The relationship is stronger, not weaker: the NDA ask, the "home team / feed mill" relationship-anchor speech, and the fleet-of-businesses disclosure are all gifts. The 0–5 year succession horizon is the one structural risk — front-load durable value because the long tail may not survive a generational handoff. The next move is Luke's: come back with the win-benchmark number, priced low enough that everyone's excited, durable + outcome-tied so it grows only as Kreilkamp wins.

Compiled v9 2026-05-28 against the 2026-05-27 meeting recording + Luke's corrections. Replaces v8 (pre-meeting). Companion docs: obb-kreilkamp-debrief-2026-05-27 (full meeting synthesis), kreilkamp-tenstreet-research (integration findings), obb-kreilkamp-scope-pricing + obb-incentive-alignment (scope + pricing). Sources: meeting recording 2026-05-27; FreightWaves, CCJ, gmtoday, Mid-West Farm Report, Repete Corporation, Feed & Grain, Moraine Park Technical College press, Kreilkamp.com, fgfmill.com; Dan's intel 2026-05-14; Luke's call notes 2026-05-21.

Generated 2026-05-28 by OBB Holdings LLC. For internal use only — never shared with Tim or any non-OBB party.