I’m stuck on a legitimately hard problem. We’re trying to build an ROI case for moving from our current patchwork of tools to a proper automation platform. The financial case looks good on paper—if we automate X process, we save Y hours per week, which translates to Z dollars annually.
But the catch is that the processes themselves will probably change once we have better automation in place. People will find new uses for the tools, we’ll consolidate workflows that we couldn’t before, and the initial use cases won’t look the same six months in. So my ROI projection is based on current processes, but the actual outcomes will be based on evolved processes. That makes it really hard to justify the upfront investment.
I’ve tried talking to our finance team about this, but they want hard numbers. They don’t want “well, if we’re more efficient, we might do more things.” They want to see concrete hours saved and concrete dollar impact.
Meanwhile, I’m watching what happens with switching costs. If we’re locked into Camunda licensing and we switch to something with unified pricing, the financial stability alone has value, even if we’re not sure exactly how many processes we’ll automate.
How are other people actually making these conversations work? Are you forecasting conservatively and letting the upside be a bonus? Building the case on just the first three use cases and not counting anything else? I’m curious how others handle the mismatch between what you can promise and what tends to actually happen after implementation.
The secret is to separate the licensing decision from the automation opportunity. Finance gets much more comfortable when you frame it differently.
Here’s what actually works: make the licensing argument first. Show them that consolidating AI model costs and moving to predictable subscription pricing is a horizontal improvement. That’s table stakes. You’re not saving money, but you’re eliminating variable cost surprises. Finance understands that.
Then make the automation case on top of that foundation, but conservative. Pick your lowest-risk automations—the ones you’re almost certain about. Don’t try to forecast the upside. Just get the first three wins locked in.
What we do is build phase gates. Phase 1 covers licensing consolidation plus three specific workflows with proven ROI. We commit to hard numbers. Then after six months, we review what actually happened and build the case for Phase 2 based on real savings, not projections.
Finance is way more comfortable funding Phase 2 when they can see Phase 1 actually delivered. And by then, you know which new opportunities emerged, so your Phase 2 forecast is actually grounded in reality rather than speculation.
Frame it as operational leverage, not just cost reduction. The ROI isn’t just hours saved on existing processes—it’s also the capability you unlock. Once you have proper automation in place, you can do things you’re currently not doing because the friction is too high.
For finance: put ROI on the existing three processes and call that the baseline. Then separately note that you expect additional opportunities to emerge, but you’re not including those in the formal projection. That way you’re being honest about uncertainty without letting it tank the whole case.
In practice, we found the first six months were exactly what we projected. Then Q3-Q4, new use cases showed up and the actual ROI was 30% higher than baseline. Finance liked that story—we under-promised and over-delivered.
The switching cost is actually the strongest part of your argument. If you’re currently locked into Camunda, your finance team should understand optionality value. Move to a more flexible licensing model, keep the same automation capability, and your risk profile improves. That’s an easier sell than “we’ll save hours through automation.”
For the process change dynamic, finance understands that too if you frame it as compounding. Year 1: 100 hours saved. Year 2: 150 hours because processes evolved. That’s not uncertain, that’s predictable escalation.
The real trap is treating the automation investment as a one-time purchase. Position it as laying foundational capability that gets more valuable over time.
I was stuck in exactly this position. Finance wanted guarantees, but automation always creates unexpected upside.
What actually worked was showing them the cost baseline first. We were juggling multiple AI subscriptions plus licensing fees. Moving to one unified subscription for 400+ models was immediate cost clarity. That’s the part finance gets instantly—no surprises, no mid-year changes.
Then on top of that foundation, we built the automation case on specific workflows. The no-code builder meant even non-technical people could identify new automation opportunities, which we fed into a formal prioritization process. We committed to three specific wins in Year 1, measured them, and let that data drive Year 2 planning.
The kicker: because we had unified pricing, adding new automations didn’t trigger license negotiation or model pricing debates. People experimented more, and we found twice as many automation opportunities as we’d originally forecast. Finance saw that as upside, not miscalculation, because we’d been conservative in the original case.
The tipping point was when our CFO realized the platform was paying for itself just from reduced licensing chaos, and every workflow improvement was pure upside.