A new colleague joined our warehouse and immediately identified three hidden bottlenecks that were holding it back.
Release time:
2026-05-21
Deep Insights is an AI‑agent‑driven smart supply chain technology service provider. With over 20 years of deep expertise in the logistics sector, it leverages the DI.AI Platform as its intelligent foundation and AI Agents as its execution units, covering four key domains—warehousing, transportation, freight forwarding, and shipping—to equip every link in the supply chain with autonomous closed‑loop capabilities for perception, decision‑making, action, and self‑healing.
Agent·Lin Onboarding Observation Diary ✏️
At nine o’clock on Monday morning, Amin, as usual, turned on WMS 。
Amin is the scheduler on the second floor of the warehouse and has been in this role for six years. His daily tasks include batching orders, assigning delivery routes, and overseeing operations on site.
A new avatar has been added to the left-hand menu of WMS—Agent·Lin ( Picking efficiency). Last week, the team leader mentioned in the group chat that a new member had joined the warehouse team—hired in the same batch as the one from the transportation team.
She clicked in, and there was already a message in the chat window.
Lin
We have reviewed the data from the past seven days and identified three issues worthy of attention:
① Order picker Zhang Wei has shown an unusually high walking distance over the past five days.
② This week, the on-time rate for waves decreased by 6.2%.
③ Over the past three days, mis-pickings have been concentrated on two SKUs.
9:02, the newcomer didn’t say hello and immediately threw out three questions.
The warehousing team faced tight KPI targets this year, with metrics like on-time wave fulfillment and the mis-picking rate frequently coming under scrutiny at weekly meetings.
But when it comes to the “why,” even Amin himself sometimes can’t quite put it into words.
Without hesitation, she clicked on the first one.
01 Every time Zhang Wei picks something up, why does he end up walking over ten meters farther than everyone else?
Lin
Zhang Wei’s average over the past 5 days is 38.2 meters per pick, 63% above the team average.
Analysis reveals that task allocation spans the A, C, and D storage areas, resulting in unnecessary travel.
💡 Recommendation: Zhang Wei should designate Areas C and D as his primary work zones, with order assignments prioritized for nearby matches. It is expected that the employee’s order-picking efficiency will improve by approximately 25%.
Adopt the suggestion
Not processed for now
Dispatches are assigned based on order priority, regardless of the agent’s location—something that doesn’t seem like a big issue at first, but over time, what ends up being lost isn’t just a few steps; it’s a significant chunk of efficiency.
Amin clicked “Accept Proposal.”
Lin
Issued to WMS · 9:10:18
Withdrawal
02 Why did this week’s on-time rate for wave scheduling drop by 6.2%?
Click on the second one.
Lin
This week’s wave on-time rate 87.3% , down from last week 6.2% . Average timeout for delayed waves 8 to 12 minutes 。
Top 3 longest waves:
• W-092: Average 41.2 m / Picking (across areas A/C/D)
• W-087: Average 38.6 m / Picking (across Areas B and D)
• W-079: Average 35.1 m / Picking (far end of Zone A)
These three waves accounted for 72% of this week’s overtime.
Under normal circumstances, the wave scheduling rule is based on order priority and SLA For time‑based order consolidation, the warehouse location distribution serves only as a secondary reference. However, the orders listed above have frequently mixed SKUs from zones C and D into the wave that is primarily centered in zone A, causing a sudden surge in cross‑zone orders.
The rules are correct; it’s the order structure that has changed.
Lin
💡 Suggestion:
Enable the intelligent wave‑planning engine, incorporate geographic clustering of storage locations into the primary dimension for order consolidation, and prioritize merging orders from nearby locations.
Starting at 6:00 AM tomorrow, a one-day trial run will be conducted; whether to enable the feature permanently will be decided after data is collected.
The on-time rate for the wave is expected to rebound to above 93%.
Adopt the suggestion
Not processed for now
Amin once again clicked “Accept Proposal.”
Lin
The intelligent wave‑planning engine has been configured for a trial run tomorrow · 9:21:47
Withdrawal
03 Why are those two part numbers always picked incorrectly?
Click on the third item.
Lin
In the past three days, there were 21 mis-pick incidents, 15 of which involved… SKU-7821 and SKU-7831。
Analysis of the causes:
The two SKU names differ by only one digit.
The spacing between storage locations is 1.2 meters, adjacent to each other.
This week’s wave density is up 23% compared to last week.
Two SKUs frequently co-occur in recent orders.
💡 Suggestion:
The spacing between storage locations has been temporarily increased to over 4 meters.
PDA Add a second scan confirmation for these two SKUs.
During peak hours, split the two SKUs between different pickers.
Adopt the suggestion
Not processed for now
It’s very comprehensive, Amin thought, and clicked “Accept Suggestion.”
Having marked all three, Amin casually flipped through the “Issued Recommendations.”
The intelligent wave‑scheduling engine in Article 2 will change the order‑consolidation logic starting tomorrow morning.
Article 1 ties Zhang Wei to Zones C and D—once the smart engine is activated, the order distribution in those zones will shift. Forcibly assigning him to a specific zone beforehand could actually undermine the flexibility of the dispatch system.
Article 3: If items 7821 and 7831 are manually split into separate picking tasks assigned to different pickers, this will override the new engine’s “nearby location consolidation” logic.
Amin tapped the first message and recalled it. Then he tapped the third one and recalled it as well.
Lin
Recommendation 1 has been withdrawn · 10:30:18
Recommendation 3 has been withdrawn · 10:30:19
As noon approached, Old Liu, the restocking team leader, came over from the platform and stood behind Amin, glancing at the screen for a while.
“How’s the new guy?”
Amin didn’t turn around; he pulled up the conversation for him to see: “The key is being able to identify problems, come up with solutions, and—most importantly—have a ‘regret‑undo’ option to retract them.”
At 5:30 p.m., as Amin was about to shut down his computer, the last message popped up in the chat window.
Lin
Today, three suggestions were processed: one was adopted, and two were withdrawn. The intelligent wave‑planning engine will begin its trial run at 6:00 a.m. tomorrow.
Amin closed his laptop after finishing reading.
As she stepped out of the office, she remembered that at the dinner table last week, the warehouse manager had mentioned the new hire in the transportation team, saying that he’d helped the company save a substantial sum last month.
@Deep Insights AI-Agent–Powered Intelligent Supply Chain Technology Service Provider
With over 20 years of deep expertise in the logistics sector, leveraging the DI.AI Platform as its intelligent foundation, and… AI Agents As an execution unit, it spans the four major domains of warehousing, transportation, freight forwarding, and shipping, endowing every link in the supply chain with autonomous closed-loop capabilities for sensing, decision-making, action, and self‑healing.
Tag:
Related News
Trying and Experiencing Immediately.
Products
Solution
Resource
About UsSales Hotline
+86-400-107-7860
Address: Room 1201, Urban Construction International Building, 500 Fushan Road, Pudong New Area, Shanghai
E-mail:info@deepinsights.com.cn
Service Line:+86-400-107-7860
Follow Us
Join our community