Business case
The honest ROI of an AI chatbot — with the formula
Most ROI calculations for AI chatbots are fiction. Not malicious — just the reflex of stacking every nice-sounding number. Here's an honest formula, four real cost lines and a worked example based on data from real deployments. Including the scenario where a chatbot isn't worth it.
Why most calculations are bullshit
A typical sales-deck calculation: “1,000 questions/month × 5 minutes each × €25/hour × 60% automation = €1,250 savings per month.” Sounds logical. The problem: that formula ignores setup time, iteration time, escalation management and the fact that you can't halve your team because the bot handles 60%.
What you don't do
“Bot saves 60% — so we drop 60% of the team.” That's not ROI, that's a wrong assumption you'll correct two weeks later.
The four cost lines that actually count
- Chatbot licence (fixed, monthly)
- Setup time: gathering sources, writing prompts, running tests — one-off 8-16 hours
- Iteration time: ~4 hours in month 1, ~30 min/month after that
- Escalation management: time your team spends on well-handed-off escalations
The biggest surprise is always point 4. A good escalation costs your team less time than answering a cold email — because the advisor gets context. That's value, but it's not “zero minutes.”
An honest example — webshop with 200 questions/month
Take a webshop, 200 customer questions per month, average handling time 5 minutes, support hourly cost €30 (incl. employer costs). The bot handles 60%, escalates 10% with context (saving 2 of those 5 minutes), 30% remains direct human work.
€300
agent time saved per month
€10
escalation time saved per month
€140
Chatwize licence per month (example)
Month 1 net: 8h setup × €30 = €240 one-off. Monthly: €310 benefit – €140 licence = €170 net. Payback: about 6 weeks.
This is the boring number
€170/month from pure time savings. The interesting number comes from the 35-65% out-of-hours — not in “saved agent time” but in “revenue not lost.” That's the harder ROI story to quantify, and it consistently turns out to be the bigger one.
The three hidden costs no one prices in
- Maintaining bad sources costs time. Budget one extra day in month 1 to clean up your product pages and FAQ — you should have done it anyway, but the bot makes it urgent.
- Edge cases need ownership. Who decides how the bot handles a new product line? Agree it now, otherwise it goes orphan.
- Iteration discipline is the most expensive hidden cost. A bot without monthly review degrades over time. Schedule it, or your ROI evaporates within six months.
When a chatbot isn't worth it
Honest answer: under 50 questions per month the time-saving formula gets thin. Under 30 questions per month you may as well skip it — or deploy with a different goal (lead capture, 24/7 presence for rare-but-valuable questions). But then you're not counting agent-hour ROI; you're counting missed-conversion ROI.
The bad ROI story: “We save 60% on support.” The good ROI story: “We serve 40% more questions, at moments we were otherwise closed.” The second one is true more often — and the second one pays back more.
Ready to make this happen for your team?
Book a short demo and we'll show how Chatwize fits your customer questions, channels and processes.