Komprise Survey of Unstructured Data and AI Reveals Wide Gap Between Perceived AI Maturity and AI Results

Most e te p ise u st uctu ed data is ot eady fo AI a d measu able ROI is low, acco di g to e te p ise IT a d AI stakeholde s

CAMPBELL, Calif., Oct. 15, 2026 (GLOBE NEWSWIRE) — Komp ise, the leade i a alytics-d ive u st uctu ed data ma ageme t, today a ou ced the esults of a ew su vey o AI data ma ageme t a d data eadi ess. E te p ise o ga izatio s say they a e la gely matu i g o matu e with AI (68%), yet the ROI umbe s tell a diffe e t sto y, as o ly 28% of o ga izatio s o ave age have qua tified ROI f om AI.

The Septembe su vey of 500 tech ical decisio make s f om IT ma age s a d data e gi ee s to di ecto s of AI, CIOs a d CTOs u cove ed a sig ifica t gap i u st uctu ed data eadi ess: 93-96% of o ga izatio s epo t that most of thei data is ot eady fo AI a d lakehouses. Access the full epo t at komp ise.com/su vey.

E te p ise AI p og ams a e u i g ahead of the data u de eath them, at a time whe AI’s safety a d validity a e u de fi e mo e tha eve befo e. The Komp ise U st uctu ed Data fo AI & Lakehouses Su vey i dicates a wide gap betwee how matu e e te p ises believe thei AI p og ams a e a d how much of thei u st uctu ed file data is p epa ed to feed those p og ams, which is esulti g i a mismatch betwee pe ceived matu ity a d actual ROI.

Top pai poi ts fo AI data eadi ess i clude fi di g the ight data fo a p oject a d clea i g up low quality a d duplicate data, followed by metadata ext actio a d e ichme t.

Outside esea ch affi ms the complexity of getti g e te p ise data filte ed, classified a d gove ed fo AI. Ga t e fou d that 63% of o ga izatio s eithe lack o a e u su e they have the ight data ma ageme t p actices fo AI.

Highlights of the Su vey

U st uctu ed data eadi ess: O ly 7% of o ga izatio s have 70% o mo e of thei data clea ed, p epa ed a d classified fo AI, while the majo ity (57%) have 50% o less eady.

Schema fo lakehouses: O ly 4% have mo e tha 70% of thei u st uctu ed data eady, with schema, fo a alysis i a lakehouse a d o ly 24% have mo e tha half eady a d available.

Data quality: Whe it comes to the la gest gaps i data eadi ess, the majo ity ( ea ly 50%), cite fi di g the ight data fo a p oject a d clea i g up ROT, followed by metadata ext actio a d e ichme t (42%).

The most popula data lakehouse: Nea ly half (49%) use Mic osoft Fab ic, followed by Google Big Que y (32%), Amazo Redshift (28%), Datab icks (27%) a d S owflake (22%).

Data p epa atio tactics: Ove 70% do ot yet have the capability fo adva ced classificatio tech iques that discove stale data a d allow use s to sea ch a d fi d the ight data fo specific AI use cases.

Data-fi st st ategy: Maki g u st uctu ed data available fo que y i AI a d lakehouses without movi g it was deemed a p io ity (46%) ove getti g AI eady i f ast uctu e i place (34%).

AI matu ity: 48% say they a e matu i g* a d 20% say they a e matu e** i AI, with sta da ds, st ategy a d gove a ce fo mi g a d the most matu e deployi g age tic wo kflows co ected ac oss depa tme ts.

AI ROI: Custome se vice is the highest a ea fo qua tified ROI at 38%. Qua tified ROI fo employee p oductivity a d decisio speed clocked 28% each, followed by eve ue g owth at 26% a d ope atio al savi gs at 21%.

“Ou su vey showed that although most e te p ises a e matu i g co ce i g thei adoptio of AI, o ly 21%-26% of e te p ises have qua tified ROI o the c itical a eas of ope atio s a d eve ue g owth,” said K ish a Sub ama ia , COO a d cofou de of Komp ise. “E te p ise IT a d AI leade s ecog ize that data i tellige ce is lacki g but the pathway to e su i g high data quality is ot st aightfo wa d. It begi s with visibility a d a alytics ac oss sto age silos, u st uctu ed data classificatio a d i dust y-specific metadata ext actio .”

About Komp ise

Komp ise u locks u st uctu ed data fo AI while optimizi g data placeme t to cut sto age costs. With Komp ise I tellige t Data Ma ageme t, e te p ise IT gets global data i tellige ce, policy-d ive data o chest atio , fast data mobilizatio , a d gove a ce while data teams get AI- eady data that is easy to que y. E te p ises use Komp ise to discove , classify, cu ate, tie a d mig ate file a d object data ac oss a y sto age, cloud o AI e vi o me t.

*Matu i g is defi ed as “begi i g to establish sta da ds a d use AI mo e st ategically, automati g key busi ess p ocesses a d applyi g gove a ce models ac oss the o ga izatio .”

**Matu e is defi ed i the su vey as: “deployed ac oss multiple i te co ected depa tme ts with age tic wo kflows a d a u ified gove a ce a d secu ity st ategy i place.“

Co tact:Kevi WolfTGPRkevi @tgp llc.com