โŒ

Normal view

There are new articles available, click to refresh the page.
Before yesterdayMain stream

AI ํ™•์‚ฐ์— โ€˜์‚ฌ์šฉ์ž๋‹นโ€™ ์š”๊ธˆ ๋ชจ๋ธ์ด ์‚ฌ๋ผ์ง„๋‹คยทยทยทCIO์—๊ฒŒ ํ•„์š”ํ•œ ํ˜‘์ƒ ์ „๋žต์€?

27 November 2025 at 01:03

์ฃผ์š” ์—”ํ„ฐํ”„๋ผ์ด์ฆˆ ์†Œํ”„ํŠธ์›จ์–ด ๋ฒค๋”๊ฐ€ ์‚ฌ์šฉ์ž๋‹น ์š”๊ธˆ์ œ ์ฑ…์ • ๋ฐฉ์‹์„ ํฌ๊ธฐํ•˜๊ณ  ์‚ฌ์šฉ๋Ÿ‰ ๋˜๋Š” ์—์ด์ „ํŠธ ์ƒํ˜ธ์ž‘์šฉ ๊ธฐ๋ฐ˜ ์š”๊ธˆ์ œ๋กœ ์˜ฎ๊ฒจ๊ฐ€๋ฉด์„œ, CIO๋Š” ๋ผ์ด์„ ์Šค ํ˜‘์ƒ ์ „๋ฐ˜๊ณผ AI ํ™œ์šฉ๊ด€๋ฆฌ ๋ฐฉ์‹์„ ์‚ฌ์‹ค์ƒ ๋‹ค์‹œ ์„ค๊ณ„ํ•ด์•ผ ํ•˜๋Š” ์ƒํ™ฉ์— ๋†“์˜€๋‹ค.

์ด ๊ฐ™์€ ์š”๊ธˆ์ œ ์ „ํ™˜ ํ๋ฆ„์€ ๋ถ€์ธํ•  ์ˆ˜ ์—†๋‹ค. IDC๋Š” ์ตœ๊ทผ ๋ณด๊ณ ์„œ์—์„œ โ€œAI ์—์ด์ „ํŠธ๊ฐ€ ๋ฐ˜๋ณต์ ์ธ ์ˆ˜์ž‘์—…์„ ๋””์ง€ํ„ธ ๋…ธ๋™์œผ๋กœ ๋น ๋ฅด๊ฒŒ ๋Œ€์ฒดํ•˜๋ฉด์„œ, 2028๋…„๊นŒ์ง€ ์ˆœ์ˆ˜ ์‚ฌ์šฉ์ž๋‹น ๊ฐ€๊ฒฉ ์ฑ…์ • ๋ชจ๋ธ์€ ์‚ฌ๋ผ์งˆ ๊ฒƒ์ด๋ฉฐ, ์ „์ฒด ๋ฒค๋”์˜ 70%๊ฐ€ ์ƒˆ๋กœ์šด ๋ชจ๋ธ๋กœ ๊ฐ€์น˜ ์ œ์•ˆ์„ ์žฌ์ •๋น„ํ•˜๊ฒŒ ๋  ๊ฒƒโ€์ด๋ผ๊ณ  ๋ถ„์„ํ–ˆ๋‹ค.

๋น„์šฉ ๋ณ€๋™์— ๋Œ€๋น„ํ•ด์•ผ ํ•  ์‹œ์ 

์„ธ์ผ์ฆˆํฌ์Šค์™€ ์›Œํฌ๋ฐ์ด ๊ฒฝ์˜์ง„์€ ์ตœ๊ทผ ํˆฌ์ž์ž ๋Œ€์ƒ ์ฝœ์—์„œ ์š”๊ธˆ์ œ ๋ณ€๊ฒฝ ๊ฐ€๋Šฅ์„ฑ์„ ์ง์ ‘ ์–ธ๊ธ‰ํ–ˆ๋‹ค. ์›Œํฌ๋ฐ์ด CEO ์นผ ์—์…ด๋ฐ”ํ๋Š” ์‹ค์  ๋ฐœํ‘œ ์ž๋ฆฌ์—์„œ โ€œ์šฐ๋ฆฌ๋Š” ๊ธฐ์กด ๊ณ ๊ฐ ๊ธฐ๋ฐ˜์„ ์ค‘์‹ฌ์— ๋‘๊ณ  ์žˆ์œผ๋ฉฐ, ๋‹จ์ˆœํžˆ ์‚ฌ์šฉ์ž ์ˆ˜๊ฐ€ ์•„๋‹ˆ๋ผ ์‚ฌ์šฉ์ž๋‹น ๋งค์ถœ์— ์ง‘์ค‘ํ•˜๊ณ  ์žˆ๋‹คโ€๋ผ๊ณ  ๋งํ–ˆ๋‹ค.

์–ดํผ์—ฃ์ง€(UpperEdge)์˜ ์ž๋ฌธ ์‹ค๋ฌด ์ฑ…์ž„์ž์ธ ์• ๋ค ๋งจ์Šคํ•„๋“œ๋Š” โ€œ๊ธฐ์—…์€ ๋‚ด๋…„์— ์ƒ๋‹นํ•œ ๋น„์šฉ ๋ณ€๋™์— ๋Œ€๋น„ํ•ด์•ผ ํ•œ๋‹คโ€๋ผ๊ณ  ์ง€์ ํ–ˆ๋‹ค. ๊ทธ๋Š” โ€œ์‚ฌ์ „ ๋ณดํ˜ธ ์žฅ์น˜๊ฐ€ ์—†๋‹ค๋ฉด ์˜ˆ๊ธฐ์น˜ ์•Š์€ ์ค‘๋„ ๊ฐ€๊ฒฉ ์ธ์ƒ์ด ๋ฐœ์ƒํ•  ์ˆ˜ ์žˆ๊ณ , AI ์ œํ’ˆ์ด ์‚ฌ์šฉ๋Ÿ‰ ๊ธฐ๋ฐ˜ ๋ผ์ด์„ ์Šค๋กœ ์ „ํ™˜๋˜๋ฉด์„œ ์˜ˆ์ƒ๋ณด๋‹ค ๋งŽ์€ ๋น„์šฉ์„ ๋ถ€๋‹ดํ•  ๊ฐ€๋Šฅ์„ฑ๋„ ํฌ๋‹ค. ์„ธ์ผ์ฆˆํฌ์Šค๊ฐ€ ์—์ด์ „ํŠธํฌ์Šค(Agentforce)๋ฅผ ํ†ตํ•ด ์‚ฌ์šฉ์ž ์ˆ˜ ๊ธฐ๋ฐ˜ ๋ชจ๋ธ์—์„œ ์‚ฌ์šฉ๋Ÿ‰ ๊ธฐ๋ฐ˜ ๋ชจ๋ธ๋กœ ์ด๋™ํ•˜๋Š” ์‚ฌ๋ก€๊ฐ€ ๋Œ€ํ‘œ์ โ€์ด๋ผ๊ณ  ์„ค๋ช…ํ–ˆ๋‹ค.

๋งจ์Šคํ•„๋“œ๋Š” ์ผ๋ถ€ ๊ธฐ์—…์ด AI ๋Œ€ํ™” ์‚ฌ์šฉ๋Ÿ‰์„ ์˜ˆ์ƒ๋ณด๋‹ค ํ›จ์”ฌ ๋น ๋ฅด๊ฒŒ ์†Œ์ง„ํ•˜๊ณ  ์žˆ๋‹ค๊ณ  ์–ธ๊ธ‰ํ–ˆ๋‹ค. ๊ทธ๋Š” โ€œ๋ฌด์—‡์ด โ€˜๋Œ€ํ™”โ€™๋กœ ์ •์˜๋˜๋Š”์ง€, ์–ด๋–ค ๋ฐฉ์‹์œผ๋กœ ์‚ฌ์šฉ๋Ÿ‰์ด ์‚ฐ์ •๋˜๋Š”์ง€ ๋ช…ํ™•ํžˆ ๊ทœ์ •ํ•˜์ง€ ์•Š์€ ์ฑ„ ์ฃผ๋ฌธ์„œ๋ฅผ ์ž‘์„ฑํ•ด ๋ฌธ์ œ๋ฅผ ๊ฒช๋Š” ๊ธฐ์—…๋„ ์ƒ๋‹น์ˆ˜๋ฅผ ์ฐจ์ง€ํ•œ๋‹ค. ์ด๋Ÿฐ ๊ฒฝ์šฐ ๊ธฐ์—…์€ ์ž์‹ ๋„ ๋ชจ๋ฅด๊ฒŒ ์‚ฌ์šฉ๋Ÿ‰ ํ•œ๋„๋ฅผ ์ดˆ๊ณผํ•˜๊ฒŒ ๋˜๊ณ , ๊ทธ ๊ฒฐ๊ณผ ์ƒ๋‹นํ•œ ์ถ”๊ฐ€ ๋น„์šฉ์„ ์ง€๋ถˆํ•˜๊ฒŒ ๋œ๋‹ค. ์‚ฌ์ „์— ์ ์ ˆํžˆ ํ˜‘์ƒํ–ˆ๋‹ค๋ฉด ํ”ผํ•  ์ˆ˜ ์žˆ์—ˆ์„ ์ƒํ™ฉโ€์ด๋ผ๊ณ  ์ง€์ ํ–ˆ๋‹ค.

์œ„ํ—˜์„ ๊ณ ๊ฐ์—๊ฒŒ ์ „๊ฐ€ํ•˜๋Š” ๊ตฌ์กฐ

๊ทธ๋ ˆ์ดํ•˜์šด๋“œ ๋ฆฌ์„œ์น˜์˜ ์ˆ˜์„ ์• ๋„๋ฆฌ์ŠคํŠธ ์‚ฐ์น˜ํŠธ ๋น„๋ฅด ๊ณ ๊ธฐ์•„๋„ ๊ฐ™์€ ์˜๊ฒฌ์„ ๋‚ด๋†จ๋‹ค. ๊ณ ๊ธฐ์•„๋Š” โ€œ๊ธฐ์—…์šฉ ์†Œํ”„ํŠธ์›จ์–ด ์š”๊ธˆ์ œ๊ฐ€ ๊ตฌ์กฐ์  ์žฌํŽธ ๋‹จ๊ณ„์— ๋“ค์–ด์„ฐ๋‹ค. SaaS ๊ณ„์•ฝ์˜ ๊ธฐ๋ฐ˜์ด์—ˆ๋˜ ์‚ฌ์šฉ์ž๋‹น ์š”๊ธˆ์ œ๋Š” AI๋กœ ์ธํ•œ ์ธ๋ ฅ ์ถ•์†Œ ํ๋ฆ„ ์•ž์—์„œ ์ ์ฐจ ํž˜์„ ์žƒ๊ณ  ์žˆ๋‹ค. ๋น„์ฆˆ๋‹ˆ์Šค ๊ฐ€์น˜๊ฐ€ ์ธ๋ ฅ ๊ทœ๋ชจ์—์„œ ์ž๋™ํ™” ์ค‘์‹ฌ์œผ๋กœ ์ด๋™ํ•˜๊ณ  ์žˆ๋‹คโ€๋ผ๊ณ  ๋ถ„์„ํ–ˆ๋‹ค. ์ด์–ด ๊ทธ๋Š” โ€œ์ด ๋ณ€ํ™”๋Š” ์ž‘์€ ์กฐ์ • ์ˆ˜์ค€์ด ์•„๋‹ˆ๋ผ ๊ณ„์•ฝ ๊ตฌ์กฐ๋ฅผ ๊ทผ๋ณธ์ ์œผ๋กœ ๋’คํ”๋“œ๋Š” ์ „๋žต์  ์ „ํ™˜์— ๊ฐ€๊น๋‹ค. ๋ฒค๋”๊ฐ€ ๋‹จ์ˆœํžˆ ์˜ค๋ž˜๋œ ๊ฐ€๊ฒฉ ๋ชจ๋ธ์„ ์—†์• ๋Š” ์ฐจ์›์ด ์•„๋‹ˆ๋‹คโ€๋ผ๊ณ  ์ง€์ ํ–ˆ๋‹ค.

๊ณ ๊ธฐ์•„๋Š” ์†Œํ”„ํŠธ์›จ์–ด ์—…์ฒด๊ฐ€ ๊ฑฐ์˜ ๋ชจ๋“  ์œ„ํ—˜์„ ๊ณ ๊ฐ์—๊ฒŒ ๋– ๋„˜๊ธฐ๊ณ  ์žˆ๋‹ค๋Š” ์ ์ด ํ•ต์‹ฌ์ด๋ผ๊ณ  ๊ฐ•์กฐํ–ˆ๋‹ค.

๊ทธ๋Š” โ€œ๋ฒค๋”๋Š” AI ์—ฐ์‚ฐ ๋น„์šฉ ๋ณ€๋™์„ฑ์„ ๊ณ ๊ฐ์—๊ฒŒ ์ „๊ฐ€ํ•˜๋Š” ํ•œํŽธ, ๊ณ ๊ฐ ์ธก ์ƒ์‚ฐ์„ฑ ํ–ฅ์ƒ๋ถ„์„ ๋งˆ์ง„์œผ๋กœ ํก์ˆ˜ํ•˜๊ณ  ์žˆ๋‹ค. ์ด ๊ฐ™์€ ์œ„ํ—˜ ์ „๊ฐ€ ๊ตฌ์กฐ๋Š” ์ด์ œ ์‚ฌ์šฉ๋Ÿ‰ ๊ธฐ๋ฐ˜ ๊ฐ€๊ฒฉ ๋ชจ๋ธ์— ๊นŠ์ด ๋‚ด์žฌ๋ผ ์žˆ๋‹คโ€๋ผ๊ณ  ์„ค๋ช…ํ–ˆ๋‹ค. ๋˜ํ•œ โ€œ์˜ˆ์ธก ๊ฐ€๋Šฅํ–ˆ๋˜ ๊ธฐ์กด ๋ผ์ด์„ ์Šค ๋ชจ๋ธ์€ ํฌ๋ ˆ๋”ง, ์ธํ„ฐ๋ž™์…˜, ์ด๋ฒคํŠธ ๋“ฑ ๋ชจํ˜ธํ•œ ๋‹จ์œ„๋กœ ๋Œ€์ฒด๋˜๊ณ  ์žˆ์œผ๋ฉฐ, ์ด๋Š” ๊ฐ€์น˜ ๊ตํ™˜์„ ํ๋ฆฌ๋„๋ก ์„ค๊ณ„๋œ ๊ฒฝ์šฐ๊ฐ€ ๋งŽ๋‹คโ€๋ผ๊ณ  ์ง€์ ํ–ˆ๋‹ค. ๋งŽ์€ ์ƒํ™ฉ์—์„œ ์ด๋Ÿฌํ•œ ๋‹จ์œ„๊ฐ€ ํˆฌ๋ช…์„ฑ์ด๋‚˜ ์‚ฌ์ „ ์ •์˜๋œ ์ƒํ•œ ์—†์ด ์ธก์ •๋˜๊ณ  ์žˆ๋‹ค๋Š” ์„ค๋ช…์ด๋‹ค.

์ด์–ด ๊ทธ๋Š” โ€œ๊ฒฐ๊ณผ์ ์œผ๋กœ ๋ฒค๋”๋Š” ์•ˆ์ •์ ์ธ ๋งˆ์ง„์„ ํ™•๋ณดํ•˜๋Š” ๋ฐ˜๋ฉด, ๊ณ ๊ฐ์€ ์ดˆ๊ณผ ์‚ฌ์šฉ ์œ„ํ—˜์„ ๋– ์•ˆ๊ฒŒ ๋˜๋Š” ๊ตฌ์กฐ์  ๋น„๋Œ€์นญ์ด ๊ณ ์ฐฉ๋œ๋‹ค. ํšจ๊ณผ์ ์ธ ํ˜‘์ƒ์„ ์œ„ํ•ด์„œ๋Š” ์กฐ๋‹ฌ ๋ถ€์„œ๊ฐ€ AI ๋ฉ”์ปค๋‹ˆ์ฆ˜, ์‹œ์Šคํ…œ ํ…”๋ ˆ๋ฉ”ํŠธ๋ฆฌ, ์ง€์ถœ์„ ์œ ๋ฐœํ•˜๋Š” ํ–‰๋™ ์‹ ํ˜ธ๋ฅผ ์ดํ•ดํ•˜๋Š” ์—ญ๋Ÿ‰์„ ๊ฐ–์ถฐ์•ผ ํ•œ๋‹ค. ๊ทธ๋ ‡์ง€ ๋ชปํ•˜๋ฉด ๊ณ„์•ฝ์ด ํ†ต์ œ ๋ฒ”์œ„๋ฅผ ๋ฒ—์–ด๋‚˜ ์˜ˆ์‚ฐ์ด ๋น ๋ฅด๊ฒŒ ๋ฌด๋„ˆ์งˆ ์ˆ˜ ์žˆ๋‹คโ€๋ผ๊ณ  ๊ฒฝ๊ณ ํ–ˆ๋‹ค.

CIO์—๊ฒŒ ํ•„์š”ํ•œ ์ „๋žต

์ด๋Ÿฐ ๋ณ€ํ™”๋Š” ๊ธฐ์—…์ด AI ์‚ฌ์šฉ๋Ÿ‰์„ ์ถœ์žฅ ๊ฒฝ๋น„ ์˜ˆ์‚ฐ์ฒ˜๋Ÿผ ์—„๊ฒฉํ•˜๊ฒŒ ๋ฐฐ๋ถ„ํ•˜๊ณ  ๊ด€๋ฆฌํ•˜๋„๋ก ์š”๊ตฌํ•œ๋‹ค. ํ˜„์žฌ AI ๋„์ž… ๋ฐฉ์‹๊ณผ ๋น„๊ตํ•˜๋ฉด ํ–ฅํ›„ ๋ฒค๋” ํ˜‘์ƒ ์ „๋žต๋„ ํฌ๊ฒŒ ๋‹ฌ๋ผ์งˆ ์ˆ˜ ์žˆ๋‹ค.

๋งˆ์ผ“ํ”„๋ฃจ๋ธAI(Market-Proven AI)์˜ CEO ์• ๋Ÿฐ ํผํ‚จ์Šค๋Š” ์ฃผ์š” ์†Œํ”„ํŠธ์›จ์–ด ๋ฒค๋”๋ฅผ ๊ต์ฒดํ•  ๋•Œ ๋ฐœ์ƒํ•˜๋Š” ์ „ํ™˜ ๋น„์šฉ์ด ๋งค์šฐ ํฌ๊ธฐ ๋•Œ๋ฌธ์—, ๊ธฐ์—…์ด โ€œ๋‹ค๋ฅธ ๊ณณ์œผ๋กœ ์˜ฎ๊ธฐ๊ฒ ๋‹คโ€๋ผ๊ณ  ์••๋ฐ•ํ•˜๋Š” ์ „๋žต์€ ์‹คํšจ์„ฑ์ด ๋–จ์–ด์ง„๋‹ค๊ณ  ์ง€์ ํ–ˆ๋‹ค. ๋”์šฑ์ด ๋ชจ๋“  ๋ฒค๋”๊ฐ€ ์‚ฌ์šฉ์ž ์ˆ˜ ๊ธฐ๋ฐ˜ ์š”๊ธˆ์ œ๋ฅผ ํฌ๊ธฐํ•˜๊ฒŒ ๋˜๋ฉด, ์• ์ดˆ์— ์˜ฎ๊ฒจ๊ฐˆ ์„ ํƒ์ง€๋„ ์‚ฌ๋ผ์ง„๋‹ค.

๋ถ„์„๊ฐ€๋“ค์€ ๋ฒค๋”๊ฐ€ ์ƒˆ๋กœ์šด ์š”๊ธˆ์ œ๋กœ์˜ ์ „ํ™˜์„ ์š”๊ตฌํ•  ๋•Œ, ํ˜‘์ƒ์—์„œ ๋ชจ๋“  ์šฉ์–ด๋ฅผ ๋ช…ํ™•ํ•˜๊ฒŒ ์ •์˜ํ•˜๋Š” ๊ฒƒ์ด ๊ฐ€์žฅ ๋จผ์ € ํ•ด์•ผ ํ•  ์ผ์ด๋ผ๊ณ  ์ž…์„ ๋ชจ์•˜๋‹ค.

ํผํ‚จ์Šค๋Š” โ€œ์‚ฌ์šฉ๋Ÿ‰ ๊ธฐ๋ฐ˜ ๋ชจ๋ธ์€ ์„œ๋กœ ๋‹ค๋ฅด๊ฒŒ ํ•ด์„๋  ์ˆ˜ ์žˆ๋‹ค. ์‚ฌ์šฉ์ž๊ฐ€ ๋กœ๊ทธ์ธํ•œ ์‹œ๊ฐ„๋งŒ์„ ์˜๋ฏธํ•˜๋Š”์ง€, ์•„๋‹ˆ๋ฉด ํ‡ด๊ทผ ํ›„ ์ž๋™์œผ๋กœ ์ด๋ค„์ง€๋Š” ๋ฐฑ์—…๊นŒ์ง€ ํฌํ•จํ•˜๋Š”์ง€ ๋ถ„๋ช…ํžˆ ๋”ฐ์ ธ์•ผ ํ•œ๋‹คโ€๋ฆฌ๊ณ  ๋งํ–ˆ๋‹ค. ๊ทธ๋Š” โ€œ์–ด๋ ค์šด ์งˆ๋ฌธ์„ ํ•ด์•ผ ํ•œ๋‹ค. โ€˜์‚ฌ์šฉ๋Ÿ‰์ด ๋ฌด์—‡์„ ์˜๋ฏธํ•˜๋Š”์ง€ ๊ตฌ์ฒด์ ์œผ๋กœ ์„ค๋ช…ํ•˜๋ผโ€™๊ณ  ์š”๊ตฌํ•ด์•ผ ํ•œ๋‹คโ€๋ผ๊ณ  ๊ฐ•์กฐํ–ˆ๋‹ค.

๋˜ ๋‹ค๋ฅธ ์ „๋žต์€ ๊ณ„์•ฝ ์‚ฌ์šฉ๋Ÿ‰์˜ ์ƒํ•œ์„ ์„ ์—„๊ฒฉํ•˜๊ฒŒ ๋ช…์‹œํ•˜๊ณ , ์ด๋ฅผ ์ดˆ๊ณผํ•˜๊ธฐ ์ „์— ๋ฐ˜๋“œ์‹œ ๋ฒค๋”๊ฐ€ ์„œ๋ฉด์œผ๋กœ ์Šน์ธ์„ ์š”์ฒญํ•˜๋„๋ก ์ฑ…์ž„์„ ๋ถ€์—ฌํ•˜๋Š” ๋ฐฉ๋ฒ•์ด๋‹ค. ๋ฒค๋”๊ฐ€ ์Šน์ธ ์—†์ด ์ดˆ๊ณผ ์‚ฌ์šฉ์„ ๋ฐœ์ƒ์‹œํ‚จ๋‹ค๋ฉด, ํ•ด๋‹น ๋น„์šฉ์€ ๋ฒค๋”๊ฐ€ ๋ถ€๋‹ดํ•˜๋„๋ก ํ•˜๋Š” ๊ตฌ์กฐ๋‹ค.

ํผํ‚จ์Šค๋Š” ํŠนํžˆ ์—์ด์ „ํŠธ ๊ธฐ๋ฐ˜ ์ƒํ˜ธ์ž‘์šฉ, ๊ทธ์ค‘์—์„œ๋„ ์™„์ „ ์ž์œจํ˜• ์—์ด์ „ํŠธ์˜ ํ™œ๋™์ด ๊ฐ€์žฅ ๊นŒ๋‹ค๋กœ์šด ๊ฐ€๊ฒฉ ๋ณ€์ˆ˜๋ผ๊ณ  ์ง€์ ํ–ˆ๋‹ค. ๋ฒค๋”๊ฐ€ ์—์ด์ „ํŠธ์˜ ํ–‰๋™์„ ํ†ต์ œํ•˜๊ณ  ์–‘์ธก์ด ์‚ฌ์šฉ๋Ÿ‰ ๊ธฐ๋ฐ˜ ๋ชจ๋ธ์— ํ•ฉ์˜ํ•œ ๊ฒฝ์šฐ, ์‚ฌ์šฉ๋Ÿ‰์ด ๋Š˜์–ด๋‚ ์ˆ˜๋ก ์ˆ˜์ต์ด ์ฆ๊ฐ€ํ•˜๊ธฐ ๋•Œ๋ฌธ์— ๋ฒค๋”๋Š” ๋” ๋งŽ์€ ํ™œ๋™์„ ์œ ๋„ํ•  ๊ฐ€๋Šฅ์„ฑ์ด ์žˆ๋‹ค.

๋ฌด์–ด ์ธ์‚ฌ์ด์ธ  ์•ค ์ŠคํŠธ๋ž˜ํ‹ฐ์ง€์˜ ์ˆ˜์„ ์• ๋„๋ฆฌ์ŠคํŠธ ์ œ์ด์Šจ ์•ค๋”์Šจ์€ ์ƒˆ๋กœ์šด ๊ฐ€๊ฒฉ ๋ชจ๋ธ์ด ์ ์šฉ๋˜๊ธฐ ์ „์— ์œ ์˜ˆ ๊ธฐ๊ฐ„์„ ์š”๊ตฌํ•˜๋Š” ๊ฒƒ์ด CIO๊ฐ€ ์ œ์‹œํ•  ์ˆ˜ ์žˆ๋Š” ํ˜„์‹ค์ ์ธ ์ถœ๋ฐœ์ ์ด๋ผ๊ณ  ์กฐ์–ธํ–ˆ๋‹ค.

์•ค๋”์Šจ์€ โ€œ1๋…„ ์ •๋„์˜ ์œ ์˜ˆ๋ฅผ ์š”๊ตฌํ•˜๋Š” ์ „๋žต์ด ๋„๋ฆฌ ํ™œ์šฉ๋  ๊ฒƒโ€์ด๋ผ๊ณ  ๋งํ–ˆ๋‹ค. CIO๋Š” ์ด ์‹œ๊ฐ„์„ ํ†ตํ•ด ํ˜„์žฌ ์‚ฌ์šฉ๋Ÿ‰์„ ์ •ํ™•ํ•˜๊ฒŒ ํŒŒ์•…ํ•˜๊ณ , ๊ฐ€๊ฒฉ ์ „ํ™˜ ์ดํ›„๋ฅผ ๋Œ€๋น„ํ•œ ๋ชจ๋ธ์„ ๋งˆ๋ จํ•  ์ˆ˜ ์žˆ๋‹ค. ์•ค๋”์Šจ์€ ์ด๋ฅผ ์œ„ํ•ด ํ•€์˜ต์Šค(FinOps) ๊ด€๋ จ ์—ญ๋Ÿ‰ ๊ฐ•ํ™”์™€ ๋” ์ •๊ตํ•œ ๊ด€์ธกยท๊ณ„๋Ÿ‰ ๋„๊ตฌ ํˆฌ์ž๊ฐ€ ํ•„์š”ํ•˜๋‹ค๋ฉด์„œ, โ€œ์ด์ œ ์—์ด์ „ํŠธ ์‚ฌ์šฉ๋Ÿ‰์„ ๊ณ„๋Ÿ‰ํ•˜๋Š” ์†Œํ”„ํŠธ์›จ์–ด๋„ ์–ด๋А ์ •๋„ ์ˆ˜์ค€์— ๋„๋‹ฌํ•˜๊ณ  ์žˆ๋‹คโ€๋ผ๊ณ  ์กฐ์–ธํ–ˆ๋‹ค.

์•ค๋”์Šจ์€ โ€œ๋ฒค๋”์˜ ๋Œ€์‘ ๋ฐฉ์‹๋„ ์˜ˆ์ธก ๊ฐ€๋Šฅํ•˜๋‹ค. ๋ฒค๋”๋Š” โ€˜X๋‹ฌ๋Ÿฌ์— 10์–ต ๊ฐœ์˜ ํ† ํฐ์„ ์ œ๊ณตํ•˜๊ฒ ๋‹คโ€™๋Š” ์‹์œผ๋กœ ์ œ์•ˆํ•  ๊ฒƒ์ด๋‹ค. ํ† ํฐ์ด๋“  ์š”์ฒญ์ด๋“  ์ด์ œ ๋‘˜์€ ๊ฑฐ์˜ ๊ฐ™์€ ์˜๋ฏธ๋กœ ์“ฐ์ด๊ณ  ์žˆ๋‹คโ€๋ผ๊ณ  ์„ค๋ช…ํ–ˆ๋‹ค. ์ดˆ๊ธฐ์—๋Š” ๋Œ€๋Ÿ‰ ์„ ๊ตฌ๋งค ํ˜•ํƒœ์ด๊ธฐ ๋•Œ๋ฌธ์— ํ† ํฐ ๋‹จ๊ฐ€๊ฐ€ ๋‚ฎ๊ฒŒ ์ฑ…์ •๋˜์ง€๋งŒ, ๋ฌธ์ œ๋Š” ๊ธฐ์—…์ด ์ตœ์ดˆ ์ œ๊ณต๋Ÿ‰์„ ์ดˆ๊ณผํ•  ๋•Œ ๋ฐœ์ƒํ•œ๋‹ค. ๊ทธ๋Š” โ€œ์ดˆ๊ธฐ์—๋Š” ๋Œ€๋ถ€๋ถ„์˜ ๊ธฐ์—…์ด ํ•œ๋„๋ฅผ ๋„˜๊ธฐ๊ฒŒ ๋  ๊ฒƒ์ด๋‹ค. ์ด ๊ฒฝ์šฐ ๋ฒค๋”๋Š” โ€˜์ดˆ๊ณผ ์‚ฌ์šฉ ์‹œ ๋” ๋†’์€ ์š”์œจ๋กœ ๋น„์šฉ์„ ์ฒญ๊ตฌํ•˜๊ฒ ๋‹คโ€™๊ณ  ํ•  ๊ฒƒ์ด๊ณ , CIO๋Š” ๊ทธ ๋น„์šฉ์„ ๊ฒฐ๊ตญ ์‚ฌ์—… ๋ถ€์„œ์— ๋‹ค์‹œ ์ „๊ฐ€ํ•ด์•ผ ํ•œ๋‹คโ€๋ผ๊ณ  ๋งํ–ˆ๋‹ค.

์ž์œจํ˜• ์—์ด์ „ํŠธ์˜ ์‚ฌ์šฉ๋Ÿ‰์„ ํ†ต์ œํ•˜๋ ค๋ฉด, ๋ชจ๋‹ˆํ„ฐ๋ง ์‹œ์Šคํ…œ์— ์ž„๊ณ„๊ฐ’์„ ์„ค์ •ํ•ด ์‚ฌ์šฉ๋Ÿ‰์ด ํŠน์ • ์ˆ˜์ค€์— ๋„๋‹ฌํ•˜๋ฉด ๊ฒฝ๋ณด๊ฐ€ ์šธ๋ฆฌ๋„๋ก ํ•  ์ˆ˜ ์žˆ๋‹ค. ์•ค๋”์Šจ์€ โ€œ์—์ด์ „ํŠธ ์ž์ฒด์— โ€˜X๊ฐœ์˜ ํ† ํฐ ์ด์ƒ ์‚ฌ์šฉํ•˜์ง€ ์•Š๋Š”๋‹คโ€™๋Š” ์ •์ฑ…์„ ์ง์ ‘ ์ ์šฉํ•˜๋Š” ๋ฐฉ์‹๋„ ๊ฐ€๋Šฅํ•˜๋‹คโ€๋ผ๊ณ  ๋งํ–ˆ๋‹ค.

ํ•˜์ง€๋งŒ ์œ„ํ—˜๋„ ์กด์žฌํ•œ๋‹ค. DDoS ๊ณต๊ฒฉ์ด ํ˜ธ์ŠคํŒ… ๋น„์šฉ์„ ๊ธ‰์ฆ์‹œํ‚ค๋“ฏ, ์‚ฌ์šฉ๋Ÿ‰ ๊ธฐ๋ฐ˜ ๊ณ„๋Ÿ‰ ์ฒด๊ณ„๋Š” ๊ฒฝ์Ÿ์‚ฌ๋‚˜ ์•…์˜์  ํ–‰์œ„์ž๊ฐ€ ๋น„์šฉ์„ ์˜๋„์ ์œผ๋กœ ๋Œ์–ด์˜ฌ๋ฆฌ๊ธฐ ์œ„ํ•ด ์•…์šฉํ•  ์ทจ์•ฝ์ ์ด ๋  ์ˆ˜ ์žˆ๋‹ค.

์•ค๋”์Šจ์€ โ€œ์˜ˆ๋ฅผ ๋“ค์–ด ๊ณ ๊ฐ ์„œ๋น„์Šค ์—์ด์ „ํŠธ๋ฅผ ๊ณต๊ฒฉํ•˜๊ธฐ ์œ„ํ•ด ๋Œ€๋Ÿ‰์˜ ๋ด‡์œผ๋กœ ์‚ฌ์šฉ๋Ÿ‰์„ ํญ์ฆ์‹œํ‚ค๋ฉด ์ƒ๋‹นํ•œ ๋น„์šฉ์ด ๋ฐœ์ƒํ•  ์ˆ˜ ์žˆ๋‹คโ€๋ผ๊ณ  ๊ฒฝ๊ณ ํ–ˆ๋‹ค. ๊ทธ๋Š” ์ด๋ฅผ ๋ฐฉ์ง€ํ•˜๊ธฐ ์œ„ํ•ด ๊ณ„์•ฝ์„œ์— ์‚ฌ๊ธฐ ๋ฐฉ์ง€ ์กฐํ•ญ์„ ํฌํ•จํ•ด์•ผ ํ•œ๋‹ค๋ฉด์„œ, โ€œ๊ณต๊ฒฉ์„ ์ž…์ฆํ•  ์ˆ˜ ์žˆ๋‹ค๋ฉด ๋น„์šฉ์„ ๋ถ€๋‹ดํ•˜์ง€ ์•Š๋„๋ก ํ•ด์•ผ ํ•œ๋‹ค. ์ด๋ฅผ ์œ„ํ•ด ์ ์ ˆํ•œ โ€˜ํšŒ๋กœ ์ฐจ๋‹จ๊ธฐโ€™ ์žฅ์น˜๋ฅผ ๊ณ„์•ฝ์— ๋‚ด์žฅํ•ด์•ผ ํ•œ๋‹คโ€๋ผ๊ณ  ์กฐ์–ธํ–ˆ๋‹ค.
dl-ciokorea@foundryco.com

New software pricing metrics will force CIOs to change negotiating tactics

26 November 2025 at 17:18

CIOs are being forced to rethink almost all elements of licensing negotiations, as well as managing how AI will be used, as major enterprise software vendors look to abandon per-seat pricing and shift to pricing based on consumption and/or agent interactions.

The evidence of such pricing shifts is now all but undeniable. โ€œBy 2028, pure seat-based pricing will be obsolete as AI agents rapidly replace manual repetitive tasks with digital labor, forcing 70% of vendors to refactor their value proposition into new models,โ€ said a recent report from IDC.ย 

Brace for โ€˜meaningfulโ€™ price shifts

Executives from Salesforce and Workday have spoken about such likely pricing changes at recent investor calls. โ€œWe are selling back into our base and weโ€™re focused not on just seats, but actually revenue per seat,โ€ Workday CEO Carl Eschenbach told analysts during a recent earnings call.ย 

That means, said Adam Mansfield, UpperEdge advisory practice leader, โ€œenterprises should brace for meaningful price shifts next year, both through unexpected in-term increases where proper protections arenโ€™t in place and through higher-than-anticipated costs tied to AI products transitioning to consumption-based licensing. Salesforceโ€™s move from per-seat or per-user pricing toward usage-based pricing with Agentforce is a prime example.โ€

Mansfield said that some enterprises are discovering that they were consuming AI conversations and surpassing the established volume of available conversations far faster than expected. Additionally, โ€œin many cases, they never fully clarified what constitutes a โ€˜conversationโ€™ or how usage would be counted for their specific use cases, and their order forms are void of the necessary detail. The unfortunate result is that customers inadvertently exceed their volume thresholds and suddenly owe Salesforce more money and, in some cases, itโ€™s a significant amount that could have been avoided with proper up-front negotiation.โ€

Shifting the risk to customers

Sanchit Vir Gogia, the chief analyst at Greyhound Research, agreed. โ€œEnterprise software pricing is undergoing a systemic reset,โ€ he said. โ€œThe per-seat model, long the foundation of SaaS contracts, is collapsing under the force of AI-led workforce contraction, shifting business value away from headcount and toward automation. This is not a marginal evolution. It is a strategic rupture. Vendors are not just retiring outdated pricing logic.โ€

But, he noted, the critical change is that software companies are trying to shift almost all of the risks to their enterprise customers.ย 

โ€œVendors are transferring the cost volatility of AI compute to customers while monetizing customer-side productivity gains as margin. This risk displacement is now embedded in the architecture of consumption pricing,โ€ Gogia said. โ€œWhat was once a predictable licensing model is being replaced by vague units: credits, interactions, events, deliberately designed to obscure value exchange. In many cases, these units are metered without transparency or pre-defined ceilings. The result is a structural asymmetry: vendors preserve their margin floor while customers absorb the risk of overconsumption. To negotiate effectively, procurement must now develop fluency in AI mechanics, system telemetry, and the behavioral signals that trigger spend. Without this, contracts will outpace control and budgets will unravel fast.โ€

Tactics for CIOs

This change will force companies to strictly allocate AI usage and manage it, not unlike how T&E budgets are handled. That would mean a sharp change in how AI strategies are deployed, compared with todayโ€™s AI efforts.ย 

Aaron Perkins, CEO at Market-Proven AI, argued that the very high switching costs that enterprises would face if changing major software suppliers would make the threat to go elsewhere less effective. But more importantly, if all vendors abandon per-seat pricing, there wonโ€™t be anywhere else to go.

Analysts agree that the first area of negotiation when an vendor pushes a different payment model is to precisely define all terms.ย 

โ€œA usage model means multiple things to multiple people. Are we talking just when a user is logged in or are we talking about backups that happen after they log off for the evening?โ€ Perkins asked. โ€œAsk the hard questions. โ€˜Explain to me what constitutes usage.โ€™โ€

Another tactic is to negotiate strict ceilings on how much usage is being purchased and put the onus on the vendor to reach out to you in writing and ask for permission to go beyond that. If they donโ€™t get that permission, then they have to eat any overage costs.

Perkins stressed that agentic interactions, especially fully autonomous interactions, pose an especially challenging pricing situation. If the vendor controls what the agent does and both parties have agreed to a consumption model, the vendor has an incentive to generate more activity so that it can charge more money.

Jason Andersen, principal analyst for Moor Insights & Strategy, recommends that CIOs start by insisting on a delay before the new pricing model goes into effect.ย 

โ€œAsking for a one year stay of execution is a tactic people will follow,โ€ Andersen said. This gives the CIO time to track current usage and to prepare pricing models for when the change happens. That means investing more in FinOps and purchasing better observability tools, he observed, noting, โ€œwe are starting to see decent software to meter these agents.โ€

โ€œThe vendors are going to respond โ€˜I am going to give you a billion tokens for X dollars. Tokens or requests, the terms are becoming interchangeable,โ€ Andersen said. That will deliver a lower per-token cost initially because it will be a large upfront purchase. The catch is that if the enterprise goes beyond that initial number, and in the beginning, almost all enterprises will, then there will be a penalty or overuse charge. โ€œThe vendor will say โ€˜If you go over, I am going to keep taking your money but at a higher rate.โ€™ Youโ€™ll then have to pass [the cost] along to the line of business,โ€ he pointed out.

To manage consumption by autonomous agents, monitoring systems can be coded to set off an alarm when usage hits a threshold. โ€œYou can put policies on the agents themselves that โ€˜You shall not use more than X tokens,โ€™โ€ he said.

But there are risks. Just as a DDoS attack can force an enterprise to pay far more for hosting, usage metering can expose the enterprise to attacks from a rival company or a cyberthief looking to punish the enterprise.ย 

โ€œIf you have a customer service agent and they launch a series of bots to drive up usage,โ€ that can cost a lot, Andersen said. He therefore recommends adding contract provisions to prevent this: โ€œsome sort of fraud protection. If I can prove that I am being attacked, I donโ€™t have to pay for it. But you need to have the right circuit-breakers built into it,โ€ he said.

How the administration is bringing much needed change to software license management

Over the last 11 months, the General Services Administration has signed 11 enterprisewide software agreements under its OneGov strategy.

The agreements bring both standard terms and conditions as well as significant discounts for a limited period of time to agencies.

Ryan Triplette, the executive director of the Coalition for Fair Software Licensing, said the Trump administration seems to be taking cues from what has been working, or not working, in the private sector around managing software licenses.

Ryan Triplette is the executive director of the Coalition for Fair Software Licensing.

โ€œThey seem to be saying, โ€˜letโ€™s see if we can import that in to the federal agencies,โ€™ and โ€˜letโ€™s see if we can address that to mitigate some of the issues that have been occurring in some of the systemic problems that have been occurring here,โ€™โ€ said Triplette on Ask the CIO. โ€œNow itโ€™s significant, and itโ€™s a challenge, but itโ€™s something that we think is important that you understand any precedent that is set in one place, in this instance, in the public agencies, will have a ripple of impact over into the commercial sector.โ€

The coalition, which cloud service providers created in 2022 to advocate for less-restrictive rules for buying software, outlined nine principles that it would like to see applied to all software licenses, including terms should be clear and intelligible, customers should be free to run their on-premise software on the cloud of their choice and licenses should cover reasonably expected software uses.

Triplette said while there still is a lot to understand about these new OneGov agreements, GSA seems to recognize there is an opportunity to address some long standing challenges with how the government buys and manages its software.

โ€œYou had the Department of Government Efficiency (DOGE) efforts and you had the federal chief information officer calling for an assessment of the top five software vendors from all the federal agencies. And you also have the executive order that established OneGov and having them seeking to establish these enterprisewide licensees, I think they recognize that thereโ€™s an opportunity here to effect change and to borrow practices from what they have seen has worked in the commercial sector,โ€ she said. โ€œNow thereโ€™s so many moving parts of issues that need to be addressed within the federal governmentโ€™s IT and systems, generally. But just tackling issues that we have seen within software and just tackling the recommendations that have been made by the Government Accountability Office over the past several years is important.โ€

Building on the success of the MEGABYTE Act

GAO has highlighted concerns about vendors applying restrictive licensing practices. In November 2024, GAO found vendor processes that limit, impede or prevent agenciesโ€™ efforts to use software in cloud computing. Meanwhile of the six agencies auditors analyzed, none had โ€œfully established guidance that specifically addressed the two key industry activities for effectively managing the risk of impacts of restrictive practices.โ€

Triplette said the data call by the federal CIO in April and the OneGov efforts are solid initial steps to change how agencies buy and manage software.

The Office of Management and Budget and GSA have tried several times over the past two decades to improve the management of software. Congress also joined the effort passing theย Making Electronic Government (MEGABYTE) Act in 2016.

Triplette said despite these efforts the lack of data has been a constant problem.

โ€œThe federal government has found that even when thereโ€™s a modicum of understanding of what their software asset management uses, they seem to find a cost performance improvement within the departments. So thatโ€™s been one issue. You have the differing needs of the various agencies and departments. This has led them in previous efforts to either opt out of enterprisewide licenses or to modify them with their own terms. So even when thereโ€™s been these efforts, you find, like, a year or two or three years later, itโ€™s all a wash,โ€ she said. โ€œQuite frankly, you have a lack of a central mandate and appropriations line. Thatโ€™s probably the most fundamental thing and why it also differs so fundamentally from other governments that have some of these more centralized services. For instance, the UK government has a central mandate, it works quite well.โ€

Triplette said what has changed is what she called a โ€œsheer force of willโ€ by OMB and GSA.

โ€œThey are recognizing the significant amount of waste thatโ€™s been occurring and that there has been lock-in with some software vendors and other issues that need to be tackled,โ€ she said. โ€œI think youโ€™ve seen where the administration has really leaned into that. Now, what is going to be interesting is because it has been so centralized, like the OneGov effort, itโ€™s still also an opt-in process. So thatโ€™s why I keep on saying, itโ€™ll to be determined how effective it will be.โ€

SAMOSA gaining momentum

In addition to the administrationโ€™s efforts, Triplette said sheโ€™s hopeful Congress finally passes the Strengthening Agency Management and Oversight of Software Assets (SAMOSA) Act. The Senate ran out of time to act on SAMOSA last session, after the House passed it in December.

The latest version of SAMOSA mirrors the Senate bill the committee passed in May 2023. It also is similar to the House version introduced in March by Reps. Nancy Mace (R-S.C.), the late Gerry Connolly (D-Va.), and several other lawmakers.

The coalition is a strong supporter of SAMOSA.

Triplette said one of the most important provisions in the bill would require agencies to have a dedicated executive overseeing software license asset management.

โ€œThere is an importance and a need to have greater expertise within the federal workforce, around software licensing, and especially arguably, vendor-specific software licensing terms,โ€ she said. โ€œI think this is one area that the administration could take a cue from the commercial sector. When theyโ€™re engaged in commercial licensing, they tend to work with consultants that are experts in the vendor licensing rules, they understand the policy and they understand the ins and outs. They often have somebody in house that โ€ฆ may not be solely specific to one vendor, but they may do only two or three and so you really have that depth of expertise, that you can understand some great cost savings.โ€

Triplette added that while finding these types of experts isnโ€™t easy, the return on the investment of either hiring or training someone is well worth it.

She said some estimate that the government could save $50 million a year by improving how it manages its software licenses.ย  This is on top of what the MEGABYTE Act already produced. In 2020, the Senate Homeland Security and Governmental Affairs Committee found that 13 agencies saved or avoided spending more than $450 million between fiscal 2017 and 2019 because of the MEGABYTE Act.

โ€œThe MEGABYTE Act was an excellent first step, but this, like everything, [is] part of an iterative process. I think itโ€™s something that needs to have the requirement that it has to be done and mandated,โ€ Triplette said. โ€œThis is something that has become new as youโ€™ve had the full federal movement to the cloud, and the discussion of licensing terms between on-premise and the cloud, and the intersection between all of this transformation. That is something that wasnโ€™t around during the MEGABYTE Act. I think thatโ€™s where itโ€™s a little bit of a different situation.โ€

The post How the administration is bringing much needed change to software license management first appeared on Federal News Network.

ยฉ Federal News Network

fnr-icon-full
โŒ
โŒ