Build vs Buy · Decision Instrument

Integration Pipeline Total Cost of Ownership

Beta

Models custom, AI-assisted integration development against a commercial iPaaS (Zapier, Workato, Dell Boomi, MuleSoft) over three years. Staffing is explicit, vendor pricing is discountable, and AI is a bounded speed factor — it compresses implementation, not the credential wrangling and upstream API quirks that consume most integration effort.

Buy · 3yr

$306K

NPV $266K

Build · 3yr

$116K

NPV $104K

Delta

$190K

62% cheaper to build

Break-even

Month 5

Cumulative crossover

Under these assumptions, owning the integration code is cheaper by $190K over three years, breaking even at month 5. The result rests on a 1.0× AI speedup reaching 55% of build hours — move either and re-read.

Cumulative spend over 36 months

Build front-loads engineering labor, then flattens into maintenance. Buy is close to linear.

Sensitivity · what actually moves the answer

← Favors buildingCurrent scenarioFavors buying →

Integration count$275K range

Monthly transaction volume$125K range

Developer rate$97K range

AI coding speedup$18K range

AI reach$17K range

Vendor discount$16K range

Each bar shows how far today's delta would move if that one input, alone, swung to the low or high end of its plausible range. Longer bar means more leverage over the decision.