Who approved it?

AI proposed: add a 'sustainability' section to every RFP response

The sales AI (Copilot) analyzed the last 18 RFP responses and proposed adding a 1-page 'sustainability' section to every RFP response, covering the company's ISO 14001 certification (in progress), the 12% reduction in water usage since 2022, and the 100% recyclable packaging. The reasoning: 8 of the last 12 RFPs included a sustainability question, and the current responses are inconsistent. The proposal is awaiting the VP of Sales review.

AI proposed: add a chatbot to the distributor portal

The customer service AI (Gemini) analyzed the last 6 months of distributor support tickets and proposed adding a chatbot to the distributor portal to handle tier-1 questions (order status, pricing, delivery). The reasoning: 64% of tier-1 questions are repetitive and can be answered from the documentation; the chatbot would save 8 hours per week of customer service time. The proposal is awaiting the customer success manager review.

AI proposed: lower the raw material safety stock for Product A

The procurement AI (Northwind internal automation) analyzed the last 12 months of raw material A usage and proposed lowering the safety stock from 30 days to 21 days. The reasoning: usage has been stable (CV of 8%), and the safety stock ties up $180K in working capital. The proposal is awaiting the procurement manager review.

AI proposed: end-of-month financial close in 3 days (from 5)

The finance AI (Claude) proposed shortening the end-of-month financial close from 5 business days to 3 business days. The reasoning: 4 of the 5 close tasks (AR aging, AP aging, bank reconciliation, accruals) can be automated with the new ERP; only the variance analysis and CFO review remain manual. The proposal is awaiting the controller review.

AI proposed: update the discount policy to 7% standard

The sales AI (Copilot) analyzed the last 6 months of discount data and proposed raising the standard discount from 5% to 7%. The reasoning: the median discount given is 6.2%, suggesting the 5% policy is consistently overridden; raising the policy to 7% would reduce the review-queue load and align the policy with reality. The proposal is awaiting the VP of Sales review.

AI proposed: switch the IT help desk to a tier-0 self-service model

The IT AI (Claude) analyzed the last 90 days of help desk tickets and proposed a tier-0 self-service model for password resets, account unlocks, and software install requests. The reasoning: 41% of tickets are tier-0 and can be resolved in under 5 minutes; the change would save an estimated 6 hours per week of help desk time. The proposal is awaiting the head of IT review.

AI proposed: standardize the customer onboarding email sequence

The customer success AI (Copilot) proposed standardizing the customer onboarding email sequence to 5 emails (welcome, portal access, first order, training, 30-day check-in) sent on days 0, 1, 5, 7, 30. The reasoning: the current sequence is ad-hoc per CSM, and the standardization would improve the 30-day completion rate from 71% to an estimated 88%. The proposal is awaiting the customer success manager review.

AI proposed: rename Product A to 'Northwind Pro Clean'

The marketing AI (Copilot) analyzed the last 12 months of customer feedback and proposed renaming Product A from 'Industrial Degreaser Heavy Duty' to 'Northwind Pro Clean'. The reasoning: the new name is shorter, more memorable, and 73% of customer survey responses called the name 'too long' or 'confusing'. The proposal is awaiting the VP of Sales review.

AI proposed: hire 1 sales engineer dedicated to the Southeast territory

The sales AI (Copilot) proposed hiring 1 sales engineer dedicated to the Southeast territory (TN, KY, GA) ahead of the regional sales manager. The reasoning: 4 of the 6 inbound distributor requests in the Southeast have been technical and would have benefited from a sales engineer's involvement; the cost of a sales engineer is $130K/year, vs. the risk of losing 3 of the 6 deals ($360K ARR). The proposal is awaiting the VP of Sales review.

AI proposed: add a 4th shift on line 1

The operations AI (Northwind internal automation) analyzed the last 90 days of line-1 throughput and proposed adding a 4th shift (weekend, 16 hours) to increase throughput by 22%. The reasoning: line-1 is at 87% utilization on weekdays, but weekend capacity is 0%. The proposal is awaiting the VP of Operations review.

AI proposed: hire 2 more process engineers for the ISO certification

The engineering AI (Claude) analyzed the ISO 9001 certification timeline and proposed hiring 2 more process engineers to support the documentation and process-mapping work. The reasoning: the current engineering team is at 92% utilization, and the ISO work requires 0.5 FTE of dedicated engineering time for 6 months. The proposal is awaiting the VP of Engineering review.

AI proposed: add a 'customer NPS follow-up' workflow (automated)

The customer success AI (Copilot) proposed adding an automated 'customer NPS follow-up' workflow: every detractor (NPS 0-6) gets a personal follow-up email within 48 hours, every promoter (NPS 9-10) gets a thank-you email and an invitation to the customer reference program. The reasoning: 23% of detractors in 2025 cited 'no follow-up' as a top reason for the low score; the workflow would close that gap. The proposal is awaiting the customer success manager review.

AI proposed: add a 'preferred customer' tier to the discount policy

The sales AI (Copilot) analyzed the last 12 months of distributor data and proposed adding a 'preferred customer' tier to the discount policy: customers with 24+ months tenure and $200K+ annual revenue get a 3% additional discount on top of the standard 5%. The reasoning: the proposed tier would affect 12 of the 142 active distributors and would lock in $680K of revenue at risk of churn. The proposal is awaiting the VP of Sales review.

AI proposed: move all 1:1s to async written (not video)

The HR AI (Copilot) proposed moving all 1:1 meetings to async written (a shared document, updated weekly) instead of synchronous video. The reasoning: 4 of the last 5 employee-survey results named 'too many meetings' as the top pain point, and 38% of 1:1 time is reportedly spent on status updates that could be async. The proposal is awaiting the head of People review.

AI proposed: end the 'monthly expense report' for small purchases

The finance AI (Claude) proposed ending the 'monthly expense report' for purchases under $50 (in place since 2018). The reasoning: 78% of expenses under $50 are receipts for meals or coffee, the average expense report takes 22 minutes, and the new Ramp system can capture these automatically with the corporate card. The proposal is awaiting the controller review.

AI proposed: retire the legacy 'Customer Service' email address

The IT AI (Claude) proposed retiring the legacy customerservice@northwind.com email address (in use since 2008) and migrating to the new distributor portal. The reasoning: 78% of customer service email volume is now handled in the portal, and the legacy address receives less than 10 emails per week. The proposal is awaiting the head of IT review.

AI proposed: move the daily standup to 8:30 AM (from 9:00 AM)

The operations AI (Northwind internal automation) proposed moving the daily line-1 standup from 9:00 AM to 8:30 AM. The reasoning: 6 of the 8 incidents in the last 90 days occurred in the first 2 hours of the shift, and a 30-minute-earlier standup would give the team more time to respond. The proposal is awaiting the plant manager review.

AI proposed: add the 'engineering manager' career path

The engineering AI (Claude) proposed adding the 'engineering manager' career path to the engineering ladder. The path is a 4-level ladder (Engineering Manager I, II, Senior, Director) with compensation bands aligned to the engineering ladder and the people-management ladder. The reasoning: 6 of the 8 engineering managers currently report to a VP and have no clear growth path above the manager level. The proposal is awaiting the VP of Engineering review.

AI proposed: add a 'first-time customer' discount of 3%

The sales AI (Copilot) proposed adding a 'first-time customer' discount of 3% (on top of the standard 5%) for the first order from a new distributor. The reasoning: 8 of the last 12 new distributors cited 'first-order pricing' as a deciding factor; the proposed discount would cost an estimated $24K/year but is expected to win 4-6 additional new distributors ($180K-$240K ARR). The proposal is awaiting the VP of Sales review.

AI proposed: end the 'twice-yearly physical inventory count'

The operations AI (Northwind internal automation) proposed ending the 'twice-yearly physical inventory count' (in place since 2003) and replacing it with a continuous cycle-count model. The reasoning: the variance is consistently below 0.5%, the cycle-count model would save an estimated 200 person-hours per year of inventory team time, and the accuracy is expected to improve. The proposal is awaiting the plant manager review.

AI proposed: end the 'punch card' time clock for line staff

The operations AI (Northwind internal automation) proposed ending the 'punch card' time clock for line staff (in place since 1995) and moving to badge-tap (the existing security badge system). The reasoning: the punch-card system is 30 years old, requires manual data entry into payroll (4 hours per week of HR time), and the badge system already tracks entry/exit with 99.8% accuracy. The proposal is awaiting the plant manager review.

AI proposed: replace the 'Q1 OKR review' with a quarterly written update

The executive AI (Claude) proposed replacing the 'Q1 OKR review' (a 4-hour all-hands with 60 attendees) with a quarterly written update (a 2-page document distributed to all employees). The reasoning: 23% of OKR survey responses said 'too much time in meetings,' and the written format has been shown to have a 3x higher read-through rate in the last pilot. The proposal is awaiting the CEO review.

AI proposed: end the 'custom formulation' minimum order

The engineering AI (Claude) proposed ending the 'custom formulation' minimum order of 500 units (in place since 2018). The reasoning: 14 of the last 18 custom-formulation requests were below 500 units; ending the minimum would unlock an estimated 8-12 new customers per year ($60K-$90K ARR). The proposal is awaiting the VP of Engineering review.

AI proposed: add a 'volunteer day' policy (2 paid days/year)

The HR AI (Copilot) proposed adding a 'volunteer day' policy (2 paid days per year, employees choose the cause). The reasoning: 12 of the last 18 employee-survey responses named 'company values alignment' as a top-3 reason to stay; the cost is estimated at $24K/year (60 employees × 2 days × $200/day); the expected retention impact is 1-2%. The proposal is awaiting the head of People review.

AI proposed: add a 'customer reference program' (10 named customers)

The sales AI (Copilot) proposed adding a 'customer reference program' — 10 named customers who have agreed to be references for prospects, with a small thank-you ($200 gift card per reference call). The reasoning: 8 of the last 12 RFPs requested customer references, and the reference program would shorten the sales cycle by an estimated 14 days. The proposal is awaiting the VP of Sales review.

AI proposed: move the monthly all-hands to Wednesday at 2 PM

The HR AI (Copilot) proposed moving the monthly all-hands from Friday at 3 PM to Wednesday at 2 PM. The reasoning: 67% of all-hands survey responses said Friday afternoon is a 'low-energy' time; Wednesday mid-afternoon has the highest engagement (per the calendar analytics). The proposal is awaiting the CEO review.

AI proposed: add the 'principal engineer' career path

The engineering AI (Claude) proposed adding the 'principal engineer' career path to the engineering ladder. The path is a 1-level distinction (Principal Engineer) above the existing Senior Engineer level, with a separate compensation band. The reasoning: 2 of the 4 senior engineers are operating at the principal level but have no recognition; the distinction would help retention. The proposal is awaiting the VP of Engineering review.

AI proposed: add a 'lunch and learn' program (monthly)

The HR AI (Copilot) proposed adding a 'lunch and learn' program (monthly, 60 minutes, employee-led topics). The reasoning: 18 of the last 24 employee-survey responses named 'learning and development' as a top-3 reason to stay; the program would cost an estimated $4K/year (food) and would not require dedicated time from the L&D team. The proposal is awaiting the head of People review.

AI proposed: add a 'return-to-office' day (Wednesday, all-hands)

The HR AI (Copilot) proposed making Wednesday a 'return-to-office' day for all office staff (the current 2-day-a-week remote policy is 'choose 2 days'). The reasoning: 4 of the last 6 in-person meetings had to be rescheduled because of remote-day conflicts; the proposed change would standardize the in-person day. The proposal is awaiting the CEO review.

AI proposed: end the 'in-person onboarding' for office staff

The HR AI (Copilot) proposed ending the 'in-person onboarding' for office staff (a 2-day in-person orientation in Madison) and moving to a fully virtual onboarding. The reasoning: 4 of the last 6 office hires are remote-first, the in-person onboarding costs an estimated $2,400 per hire (travel + 2 hotel nights), and a virtual onboarding has been piloted successfully for 2 hires in Q1. The proposal is awaiting the head of People review.

Decision: pilot a 4-day workweek for office staff in Q4 2026

On 2026-06-30, the executive team decided to pilot a 4-day workweek (32 hours, same pay) for office staff in Q4 2026 (October-December). The decision was driven by 2 staff-survey results in the last 12 months (87% of office staff said they would value a 4-day week), the tight labor market, and a controlled pilot. The pilot will be evaluated at the end of Q4.

Approved: hire 1 plant manager for line 3 (planned 2027)

On 2026-06-25, the executive team approved hiring 1 plant manager for the planned line-3 expansion in 2027. The hire will be recruited in Q4 2026, onboarded in Q1 2027, and will start running line 3 in Q2 2027. The cost is $140K/year plus benefits. The change was approved by Margaret Chen, proposed by Bartosz Lewandowski.

Approved: change the expense system from Expensify to Ramp

On 2026-06-12, the executive team approved switching the expense system from Expensify ($14K/year, 60 users) to Ramp ($18K/year, unlimited users). The change was approved by Priya Ramaswamy, proposed by the finance AI, and will be implemented 2026-07-15. The change is expected to save 8 hours per week of finance team time and improve the audit trail.

Quality hold: any line that fails QC 3x in 24h goes to engineering

Any production line that fails quality control three times in a 24-hour period is automatically placed on a quality hold. The line does not resume until a process engineer signs off. The policy was tightened from 5x to 3x on 2026-06-08 after a customer incident.

Approved: add a 'quality hold' trigger to the MES

On 2026-06-08, the executive team approved adding an automatic quality hold trigger to the MES: any line that fails QC 3x in 24h goes to engineering. The change was approved by David Werner, proposed by the operations AI, and implemented 2026-06-15. The change is expected to reduce customer-impacting quality issues by 40%.

Rejected: bring the legal function in-house

On 2026-05-30, the executive team rejected a proposal to bring the legal function in-house (currently contracted to Smith & Associates, $120K/year). The proposal was rejected because the legal workload is approximately 0.3 FTE, and the in-house cost is estimated at $180K/year, with the savings ($60K) not justifying the in-house hire.

Data retention: 7 years for financial, 3 years for operational

Financial records (invoices, payments, contracts) are retained for 7 years per IRS requirements. Operational records (production logs, quality data, sensor data) are retained for 3 years. The retention policy was last reviewed and approved on 2026-05-30 by Priya Ramaswamy ahead of the ISO 9001 audit.

Approved: increase the entry-level line operator wage to $22/hour

On 2026-05-30, the executive team approved increasing the entry-level line operator wage from $19/hour to $22/hour, effective 2026-07-01. The change was approved by Margaret Chen, proposed by HR, and is expected to reduce line-operator turnover from 28% to an estimated 18% annualized.

Approved: add Northwind as a sponsor of the Wisconsin Manufacturing Summit

On 2026-05-22, the executive team approved a $25K sponsorship of the Wisconsin Manufacturing Summit (October 2026). The change was approved by Margaret Chen, proposed by the marketing team, and is expected to generate 40+ qualified leads and 4 RFP requests. The sponsorship includes a 10x10 booth, a speaking slot for the VP of Operations, and 4 attendee passes.

Approved: add a 'veteran' hiring preference to the recruiting policy

On 2026-05-15, the executive team approved adding a 'veteran' hiring preference to the recruiting policy: veterans are guaranteed an interview for any role they meet the minimum qualifications for. The change was approved by Margaret Chen, proposed by the recruiter, and will be communicated to all hiring managers on 2026-08-01.

Merge: discount policy v1 and v2 merged into the current policy

On 2026-05-14, the legacy discount policy (v1, 2024-08-12) and the proposed new discount policy (v2, drafted 2026-04-30) were merged into the current policy (5% standard, 10% with VP, 15% with CEO). The merge was approved by Margaret Chen. The v1 and v2 records are retained in the audit log; the current policy is the single trusted record.

Conflict: discount policy is 5% vs. 7% (resolved by policy supersession)

There was a 1-month conflict between the legacy discount policy (5% standard, approved 2024-08-12) and the new sales-team-actual standard (7%, which the sales team had been giving in 62% of cases). The conflict was resolved on 2026-05-14 when the executive team approved the new policy (5% standard, 10% with VP approval, 15% with CEO approval — see 'Discount policy' under trusted knowledge). The legacy policy was superseded.

Rejected: AI proposed: add a 'chatbot' to the customer support phone line

On 2026-05-22, the VP of Sales rejected a proposal from the customer service AI (Gemini) to add a chatbot to the customer support phone line. The proposal was rejected because (1) the customer demographic (B2B distributors, 40-60 year old procurement managers) has a strong preference for human support, (2) the survey result (89% of customers said 'I want to talk to a human') is unambiguous, and (3) the cost of an AI phone bot is not justified by the projected $14K/year savings.

Reason: Customer demographic preference is for human support; the savings do not justify the change.