Most leaders already feel the sting of this, even if they do not put numbers to it:
- Large scale change efforts fail somewhere between half and roughly 70 percent of the time in the literature.
- Process improvement efforts specifically are often estimated to miss their goals in 60–70 percent of cases.
- When automation is executed poorly, it actively damages data usage, processes, employee morale, and customer satisfaction instead of improving them.
The pattern is consistent: tech is rarely the root problem. The operating model is.
This post is about that gap.
1. The uncomfortable baseline: change mostly does not work
You can argue about whether the failure rate is 50 percent, 60 percent, or 70 percent, but decades of research agree on one thing: a majority of change and improvement initiatives underperform or fail outright.
On the process side, multiple sources peg failure or underperformance of process improvement initiatives around the 60-70 percent mark, often attributing it to weak planning and lack of follow-through, not bad intentions.
Layer AI or automation on top of that reality and you do not magically get disciplined operations. You get:
- Faster dysfunction
- Expensive dashboards that no one trusts
- More pressure on already stretched teams
So when an AI program fails to move the needle, it is rarely because the model was not clever enough. It is because it landed in an operating environment that was not ready to support it.
2. Why workflow “fixes” do not stick
The research and practitioner write-ups all circle the same handful of issues.
Over complication
- Improvement efforts add steps, rules, and edge cases instead of removing them.
- Teams redesign the flow for every scenario instead of focusing on the 60-80 percent that actually matter.
- Diagrams look beautiful in workshops and completely unmanageable in real work.
No clear owner
- “Operations owns it” usually means nobody owns the end-to-end performance.
- Responsibility is spread across a steering committee, so decisions are slow and accountability is fuzzy.
- When things start to slip, everyone blames the tool or the framework, not the lack of ownership.
No baseline or measurement
- Teams change the process without a clear starting benchmark.
- Success is defined as “we went live” instead of “we reduced cycle time by X” or “we cut rework by Y percent”.
- A year later, nobody can say whether the change helped, hurt, or did nothing.
Change is never embedded in daily work
- New ways of working never make it into:
- SOPs
- Onboarding
- Performance expectations
- Manager routines
- People quietly slide back into old habits because the environment still rewards the old behavior, not the new one.
Poorly executed automation makes it worse
Gartner is blunt here: if automation is done poorly, it harms data usage, processes, morale, and customer satisfaction.
You see that when:
- Bots break every time a field or screen changes
- Frontline staff do double entry to “help the automation”
- Customers get stuck in loops with no clean human escape hatch
The result is not efficiency. It is fragility.
3. The hidden costs of failed workflow optimization
When a workflow project does not stick, you do not just lose the project budget. You also pay in:
Direct costs
- Internal time spent on workshops, meetings, and retraining
- External consulting or implementation spend
- Tool licenses that survive long after the initiative has died
Indirect human costs
- Change fatigue: “this is just another initiative that will fizzle out”
- Lower trust in leadership promises
- Less willingness to engage the next time you really do need people to change
Prosci’s work on process improvement and change is clear: if you do not bake change management into process work, you extend timelines, increase resistance, and reduce impact.
Operational costs
- More workarounds and shadow processes
- Conflicting versions of “the way we do things” between teams and shifts
- Misaligned metrics that push teams to optimize locally instead of globally
By the time you add AI into this mix, you are trying to amplify something that is already noisy.
4. What the successful minority does differently
The companies that actually see sustainable gains from workflow and systems work are not using magic frameworks. They are doing the basics with discipline.
Across case studies and guides from players like Lucid, Prosci, BCG, and others, a few patterns show up again and again.
They pick a few critical value streams, not “everything.”
- Instead of launching ten improvement projects, they start with one or two flows that clearly touch:
- Revenue
- Risk
- Customer experience
- They keep scope contained enough to finish and learn, then scale.
They name an owner for the entire process
- One person is accountable for the end-to-end outcome, even if many teams participate.
- That owner has the authority to adjust roles, rules, and tools when needed.
- Steering groups support, they do not replace ownership.
They define a small set of operational KPIs upfront
Before changing anything, they agree on:
- Baseline performance (cycle time, error rate, handoffs, rework)
- Target range or improvement for the next 3–6 months
- How they will collect and review those metrics
That makes it possible to say something stronger than “people seem happier” after go live.
They design for adoption, not just design
- Updated SOPs and how to guides
- Onboarding that teaches the new way as the default
- Manager coaching and huddles focused on the new flow
- Feedback channels so teams can raise issues and suggest tweaks
Prosci calls this embedding change management inside process work instead of bolting it on at the end.
They keep improvements small and continuous
- Short cycles (30-60 days) with visible wins work better than giant all at once redesigns.
- Each cycle tightens the process and the culture of improvement at the same time.
5. Where Systems & Workflow Optimization advisory actually helps
If you are an internal leader, you already know most of this in your gut. The hard part is doing it while running the business.
That is where an external Business Operations partner like Scale Crew can be useful, without being a crutch.
In practice, here is how Systems & Workflow Optimization work should help you, not replace you:
Diagnose with you, not to you
- Map real workflows across HR/finance/CX/ops/revenue, including:
- Who actually does what
- Where handoffs and delays live
- Which tools and spreadsheets are really in play
- Compare the “official” process with the shadow process people actually use.
Simplify before you automate
- Remove redundant steps, approvals, and tools first.
- Standardize just enough to reduce variation where it hurts, not everywhere.
- Only then decide where AI or automation belongs, if at all.
Clarify ownership and metrics
- Name a clear process owner for each critical flow.
- Agree a handful of KPIs that will tell you whether the new design is working.
- Build a simple rhythm (weekly or monthly) where owners review performance and adjust.
Embed the change so it survives the project
- Help your team update SOPs, onboarding, and training.
- Build lightweight dashboards so leaders and teams can see what changed.
- Leave behind runbooks and playbooks your people can actually use without a consultant in the room.
The goal is not to ship a pretty process map. It is to leave you with workflows that keep working when the project team has gone home.
6. A quick self-check you can run this week
Pick one process you have “improved” in the last 12-18 months. Maybe:
- New hire onboarding
- Case escalation in support
- Invoice approval
- Customer handoff from sales to success
Ask your leadership team and process participants three questions:
- Did complexity go up or down?
- Fewer steps, tools, and decisions, or more
- Is there a single named owner for this process now?
- Not a committee, an actual person
- Can we show that performance actually changed?
- Do we have before/after numbers on speed, quality, or experience
If you cannot answer those questions confidently, you do not have a tools problem. You have a workflow design and ownership problem.
And that is exactly the kind of problem Systems & Workflow Optimization is meant to solve.
