---
title: "AI’s “1998 Moment”: Salesforce AI CEO Clara Shih on Impact, Trust, and Time-to-Value"
author: "Unique Research"
sourcePublication: "Unique Research (FFCAP English library)"
originalPublishedAt: "2024-08-02T17:30:00+08:00"
canonical: "https://ffcap.cn/en/research/20240802a0796u00"
source: "https://ffcap.cn/en/research/20240802a0796u00"
language: "en"
cnSourceUrl: "https://view.inews.qq.com/a/20240802A0796U00"
---

# AI’s “1998 Moment”: Salesforce AI CEO Clara Shih on Impact, Trust, and Time-to-Value

_Unique Research · Insight · August 2, 2024 · Clara Shih, CEO, Salesforce AI_

## Key points

1.  **Trust:** 74% of the public worry about unethical AI use; 63% about bias — enterprises must build on trust and transparency.
2.  **Data foundation:** 86% of tech decision-makers say output quality tracks input data quality, freshness, and security.
3.  **Use cases forming:** 92%+ of sales/service/marketing/commerce teams considering or already using AI (content, comms, process).
4.  **Skills & policy lag:** 56% of workers believe genAI will transform their roles; only 21% say their company has a clear AI use policy.

## FOMO time

Salesforce _State of Data and Analytics_: **77%** of business leaders feel genAI FOMO — marketing 88%, sales 78%, service 73% (Nov survey; little reason to think July 2024 cooled). Shih: “Most executives worry they’ll miss AI’s upside — and they’re right. This looks like **internet 1998**: huge opportunity and huge hype. How do we filter signal and find high-impact enterprise use cases?”

## Show us the money — and the value

After 18 months of over-promise, responsible vendors (Salesforce among them) try to manage expectations; many non-tech execs still treat genAI as a magic key. Zuckerberg’s “multi-year” framing hit Meta’s stock; Pichai faced Wall Street pressure to keep investing for later returns. Shih on impatience: companies don’t want siloed, hard-to-customize buys — they need turnkey demos for the board _and_ configurability per business unit. First-year myth: “just plug ChatGPT into the org.” Reality: data-leak risk, failed private-LLM rushes, and the need for **per-user context** (same forecast prompt, different answers for two sellers). Context data is what makes an LLM production-ready.

Bottom line (Shih): like late-90s “is the internet overhyped?” — yes in the cycle, and the winners will be transformative after several experiment cycles and years. That’s why she calls it AI’s **1998 moment**.

---

Editorial English edition published on UniqueCapital: https://ffcap.cn/en/research/20240802a0796u00
On-site reading page: https://ffcap.cn/en/research/20240802a0796u00