The technology evaluation is done, the quote is in, the operational logic holds - and then the project stalls in the budget round. Warehouse automation projects rarely fail on technology. They fail because no one framed the case so Operations, IT, and Finance can approve it at the same time. This article shows who sits at the table, what question each one is really asking, and the order in which to dissolve the three most common blockers.
A warehouse automation project rarely has fewer than six people with a veto at the table. The site manager wants relief, the CFO wants numbers, IT wants to know how deep the change cuts. Each of them can stop the project, and each asks a different question. A business case that answers only one of those questions convinces one person and leaves the other five in doubt. That is where most projects stall - not on the robot, but on the approval.
For the project sponsor - usually the site or operations manager carrying the topic internally - the real work is therefore not vendor selection. It is internal alignment: the right argument to the right person, in the right order.
Why automation projects fail on approval, not on technology
The data backs this up. In the EHI study "AI and Automation in Retail Logistics 2026", 69.4% of retailers name high investment cost as the biggest barrier to automation projects - for robotic systems the figure rises to 77.8%. The number-one obstacle is not technical feasibility. It is how the project gets funded and approved.
Why one argument is never enough lies in the structure of the decision. What convinces the site manager - tangible relief in daily operations - looks incomplete to the CFO, because it does not touch his metrics. What reassures the CFO - liquidity, balance-sheet impact, unit cost - never surfaces in the site manager's daily concerns. IT, in turn, judges the same project by a third measure: person-days. One document has to serve all three at once, or the approval splinters into competing interests.
That is the bad news for anyone waiting on the best possible quote. The good news: the blockers are predictable. Once you know which question comes from whom, you can build the case so the objections are answered before they are voiced.
Who sits at the table - and what each needs to hear
Six roles show up in almost every approval round. The table below pairs each with the question behind its official question - and the argument that actually lands.
| Role | Question behind the question | What lands |
|---|---|---|
| Site / Operations Manager | How do I cut operational risk without creating new risk? | Higher throughput, less labor dependency, predictable peaks, lower all-in cost per pick, no asset ownership on site |
| Branch / 3PL Manager | Automate without a commitment longer than my contracts carry | No stranded assets, cost follows volume, short minimum term |
| CFO / Finance | What does it cost, what does inaction cost, how does it hit EBITDA? | All-in cost per pick vs. status quo, risk variabilization, address EBITDA openly |
| IT Lead | How deep does this cut, how many person-days? | Effort in person-days not months, standard interface, defined project end |
| Managing Director / COO | Strategic move or tactical fix? | Scalability, strategic optionality, OpEx allows later course corrections |
| Procurement | Market price, benchmarkable contract? | Minimum volumes, volume tiers, SLAs and exit clauses defined early |
Three of these roles decide the pace and the outcome. The site manager is usually the initiator and the champion. He carries the project internally as long as the case delivers his core promises: more throughput, less dependence on hard-to-fill roles, predictable peaks, a lower cost per pick, and no new equipment liability on site. Take one of those away and he goes quiet - and without a champion, a project rarely reaches the managing director.
The CFO does not reflexively reject operating costs. A blanket "OpEx is always better" is exactly what does not work on him. What works is an argument that meets the EBITDA question head-on: capital expenditure spreads across the depreciation period, while an operating-cost model loads the current period in full. That only balances out when the per-pick price sits below the CapEx equivalent - and that number belongs on the table, not a preference for one model.
The IT blocker is almost never a content problem. It is an assumption problem. Anyone who hears "warehouse automation" and pictures integrating an automated small-parts store budgets a six-to-twelve-month IT project. For an AMR (autonomous mobile robot) retrofit that is the wrong reference: two to four weeks over a standard interface is realistic. How that effort actually breaks down is covered in the article on the IT effort of warehouse automation.
The one-page business case
A business case that has to clear several layers of hierarchy has a built-in problem: the longer it gets, the less likely every decision-maker reads it. A 30-page document is a good working file - as a decision paper for a management team ruling on four requests in an hour, it is too long. Approval lands on one page; the detailed math goes in the appendix.
Five sections carry that page.
Problem statement
Two or three sentences, quantified rather than qualitative. Not "we need to be more efficient", but order lines per day, labor cost, unfilled positions, peak volume with its temp-labor share, error and returns cost. A good problem statement serves every role at once: the managing director sees the strategic risk, the CFO the quantifiable cost, Operations its own reality.
Solution approach
Two or three sentences on technology and business model - and why this combination. The key point: the technology decision and the procurement decision are independent. The same AMR solution can be bought, leased, taken as Robotics-as-a-Service, or run on pay-per-pick. Which model appears here follows the approval path, not the technology.
Financial assessment
A table that sets the status quo against the automated options over three horizons - typically three, five, and seven years. It shows investment, annual and cumulative cost, the payback point, and the balance-sheet treatment per model. The most important figure is not in the "savings" header but in the line beneath it: what the status quo costs over the same period. That single number speaks to the CFO more directly than any efficiency claim.
Risk and mitigation
Name three risks actively instead of leaving them to the other side: throughput below expectation, integration running long, low staff acceptance. Each gets its mitigation in the same row - a performance guarantee and a pilot against the throughput risk, pre-estimated effort and parallel operation against the integration risk, early works-council involvement against the acceptance risk. A case that names its own risks is more credible than one that hides them.
Next step
A concrete action with a date and an owner - and a pilot, not the full decision. Instead of ruling on complete automation, the approval round rules on a limited-risk test: a handful of robots, one picking station, four weeks, defined success criteria. The real decision then lands on real data from your own warehouse.
The full template with a worked example and TCO tables goes beyond the scope of this article; the structure above is enough to draft your own case.
The three most common blockers - and how to dissolve them
Even a clean business case runs into resistance that has less to do with the analysis than with the company's situation. Three patterns cover most cases - and it pays for the sponsor to know them in advance, because each one "belongs" to a specific role.
"The budget isn't there" - a CapEx problem, not a cost problem
This one belongs to the CFO and the managing director. The payback is confirmed, and the project is still shelved because other investments take priority. The diagnosis: as long as the project is booked as an investment, it competes with every other CapEx request for the same pot, regardless of ROI. The fix: switching to an OpEx model takes it out of the investment budget. Logistics then walks into the round not with an investment request but with an operating-cost comparison - not "can we invest X?" but "do we keep paying the full cost of manual picking when the alternative sits below it?" Same economic reality, different approval path. If the caution runs deeper, a pilot in the low four figures lowers the threshold further - it often falls below the formal approval limit entirely.
"IT has no capacity" - an assumption problem
This one belongs to IT. Management is behind the project, the budget is there, and it still sits because IT reports its next free capacity months out. The diagnosis is the wrong reference size: picturing classic warehouse automation (AS/RS, shuttle) means deep WMS changes and material-flow control over six to twelve months. For an AMR retrofit that is oversized - typically two to four weeks over a standard interface, with no PLC connection and no change to WMS logic. The fix: a joint workshop between internal IT and the vendor's technical team before approval, settling three things - the interface, the internal effort in person-days, and a timeline with milestones. The output is one sentence for the COO: "15 person-days over four weeks, doable alongside live operations." Plus a counter-calculation that rarely gets drawn: the status quo ties up IT capacity too - workarounds, stock corrections, scanner support - and that drops with automation. Net effort is often smaller than the gross number suggests. The depth is in WMS integration for warehouse automation.
"We're waiting for the new build" - a time-horizon problem
This one belongs to the managing director. A new build is planned ("it'll be up in 18 months"), that is where they want to "do it properly", and the interim solution in the existing building is refused with "we won't invest twice." The diagnosis has two weak points. First, logistics new-builds routinely take 30 to 48 months, not 18 - site search, permits, tender, construction, ramp-up. Second, labor cost hits now, not in three years. The fix addresses both: the cost of waiting is calculable - status quo per year, saving per year, the difference over a realistic horizon - and it usually exceeds the interim solution by a wide margin. More important is the emotional objection, "investing twice." The answer: an AMR retrofit on pay-per-pick or RaaS is not an investment in balance-sheet terms - no fixed asset, no residual book value, no removal cost. When the new build starts, the service contract ends; the interim solution behaves like temporary labor, only far more productive. As a bonus, hands-on AMR experience in the existing warehouse improves the new-build plan. How to introduce a retrofit without interrupting operations is covered in warehouse automation without downtime.
Three principles for the shortest path to approval
The three blockers follow the same pattern: the business case is sound at its core, and the resistance is structural rather than substantive - wrong approval path, wrong effort assumption, wrong time horizon. Three principles dissolve that, and they also give the sponsor the order in which to set up the round.
First, fit the business model to the approval process, not the other way around. If the investment budget is the bottleneck, the answer is not a better investment request but a model that needs none. Because the technology and procurement decisions are independent, this switch costs nothing on the solution - only on the contract.
Second, use the pilot as a decision instrument, not a delay. Four weeks, cost in the low four figures, hard data from your own warehouse. It cuts perceived risk far enough that the full rollout is approved more easily at the end than the initial investment was at the start.
Third, quantify the cost of inaction. The strongest number in the case is not the saving but what standing still costs - rising labor cost, temp-labor rates at peak, error and returns cost, missed growth. The labor market gives that number extra weight: 77.8% of retailers name the skills shortage as a driver of further automation (EHI 2026). "Doing nothing" is not a free option. It is usually the most expensive one - it just needs no investment approval. What manual picking really costs, line by line, is in the true cost of warehouse labor.
Ready to check your own case?
One free session settles whether your warehouse meets the criteria for a retrofit - rack type, aisle width, daily volume - and what the business case looks like for your site: See NEO in action.
