Why Most AI Projects Miss ROI – and How to Fix It

AI ROI traps are costing tech firms billions

Tech companies that rush AI projects without a clear value map often miss out on much of the potential profit, according to EY’s analysis. The report highlights that vague goals, siloed data and under‑estimated integration costs are common reasons for stalled returns.

Clear business outcomes beat hype

The EY study stresses that defining a measurable business outcome first—such as improving support efficiency or speeding up data‑entry—creates a concrete ROI target. Companies that start with a problem statement rather than a technology showcase tend to achieve value faster and keep total costs lower. This aligns with broader industry observations that a problem‑first approach improves success rates.

Data readiness is non‑negotiable

Siloed or low‑quality data can increase project budgets and delay timelines, EY notes. The report recommends a focused data‑audit sprint—a short intensive to map data sources, clean pipelines and establish governance—before any model is built. Analysts have observed that teams that complete a data audit tend to reduce implementation time and avoid costly re‑training cycles.

Integration costs are often hidden

Many firms underestimate the effort required to embed AI into existing workflows. EY finds that integration can represent a sizable portion of the total project budget, especially when legacy systems lack APIs. The study suggests using low‑code platforms to create thin integration layers, which can shorten development cycles. Industry surveys show that organizations leveraging low‑code tools often bring AI solutions to market more quickly.

Small business automation: a quick win

For small firms, the ROI calculus is simpler. Automating repetitive tasks—like customer queries or CRM data entry—can free a notable amount of weekly time per employee, translating to measurable savings. Using the typical Israeli automation cost for a medium‑complexity build (around ₪4,500 per weekly hour), the payback period can be relatively short, often measured in weeks rather than months. This rapid return makes AI projects attractive even for modest budgets.

What it means for Israel

Israel’s vibrant tech ecosystem, backed by the Israel Innovation Authority, can leverage these insights to stay competitive. A typical Israeli support team of three people, each spending about 10 hours/week on repetitive queries, could automate roughly ⁦60%⁩ of that work—about 18 hours/week in total. At a loaded cost of roughly ₪90 per hour, that translates to a weekly saving in the low thousands of shekels and an annual saving in the tens of thousands. With a medium‑complexity automation cost of around ₪45,000, the payback can be achieved within several months, offering a compelling case for SMEs and startups alike. Companies should therefore prioritize clear outcomes, data readiness, and low‑code integration to unlock fast, measurable returns.

The path forward

The EY report concludes that breaking out of the AI ROI trap requires discipline, transparency and incremental delivery. Start with a pilot that solves a specific pain point, validate the data pipeline, and use low‑code tools to stitch AI into existing processes. By doing so, tech firms—big and small—can turn AI from a costly experiment into a profit‑driving engine.

Explore our AI‑automation ROI calculator to model your own savings, and visit the AI‑automation data hub for benchmarks across industries.

Sources & further reading

• Original source: Google News — business