Aug. 4, 2026

Why Most AI Projects Don’t Show ROI | Sid Bharath

Show Notes

Most AI projects fail long before the model gets a fair chance to prove ROI.

In this episode of An Hour of Innovation podcast, Vit Lyoshin speaks with Sid Bharath, founder of Refound, about why companies struggle to turn AI projects into real business value. Sid explains why big “AI makeover” projects often disappoint, why better models alone do not fix broken workflows, and how teams can identify the smallest useful piece of work to automate before scaling agents across the business. The conversation covers workflow audits, AI-native organizations, human-in-the-loop work, AI agents in sales and operations, and the skills people need as repetitive work moves to AI.

They discussed:

* Why large AI transformation projects often fail to show ROI

* How older models, unclear workflows, and missing context create bad AI outcomes

* What leaders misunderstand about becoming “AI-native”

* How to audit a workflow before deciding what to automate

* Why sales, CRM updates, proposals, contracts, and admin work are strong AI-agent use cases

* What parts of work should stay human, including strategy, creativity, relationships, and review

* How AI changes decision-making and day-to-day work inside a company

* Why agents work best when employees learn by seeing them operate on real tasks

* What startup founders can automate first when time and resources are limited

* Which business processes may shrink or disappear as agents become normal

* Why Sid believes AI agents will become part of the company org chart

Sid Bharath is the founder of Refound, where he helps non-technical companies rethink their operations for the AI age. His work focuses on finding business bottlenecks, mapping real workflows, and building AI agents that employees can manage alongside their existing roles. His perspective is especially useful for founders, operators, product leaders, engineers, and business teams trying to move from AI experiments to practical workflow automation.

Timestamps

00:00 Introduction

01:20 Why AI Projects Miss ROI

03:10 What It Means to Be AI-Native

04:13 Why AI Implementations Fail

09:47 How Refound Finds Automation Opportunities

13:11 What to Automate and What to Keep Human

19:55 How AI Changes Everyday Decision-Making

24:16 Training Teams to Work With AI Agents

27:06 Where AI Agents Don’t Belong

28: 59 A Startup Playbook for AI Operations

33:00 The Future of AI-Driven Business