A shekara ta 2018, Google ta yi abin ban mamaki: sun kirkiro sunan aiki mai suna. Cassie Kozyrkov ta zama Chief Decision Scientist na kamfanin — ba Chief Data Officer, ba VP of Analytics, amma wanda aikinsa shi ne taimakawa kamfanin yanke shawara mai kyau.
Me ya sa su yi haka? Saboda Google ta gano abin da kamfanoni da yawa har yanzu ba su gane shi: kuwa da bayanai ba shi da amfani da bayanai yadda ya kamata. Suna da petabytes na bayanai, sojojin masana'antar bayanai, da kayan aikin ML na duniya. Amma suna ganin irin wannan salon: nazarin ban mamaki wanda ba a yi amfani da shi, dashboards wanda ba a canja hali saboda, namun daji na AI wanda ya samar da haske amma ba takaice.
"Decision Intelligence shine ilimin canja bayanai zuwa aikin da ya dace a kowane mataki, a kowane yanayi."
Haba, ba sabon suna ga ra'ayoyi tsohuwa ba ne. Decision Intelligence yake wakiltar canji mai zurfi a yadda kamfanoni ke tunani game da alaka tsakanin bayanai, nazarin, da aiki. Idan kuka karanta gidanmu na Collaborative Decision Making, kuka gani bangaren mutane na wannan equation — shekaru 240 na bincike daga Condorcet zuwa Google Project Aristotle wanda ya nuna cewa ra'ayoyin mutane da yawa, daidai da taro, suna fi ra'ayin mutum daya.
Decision Intelligence yake daukar wannan gini da yake tafiyar da: me ya faru lokacin da mu kara AI, causal modeling, da mafita na tsari?
The $3.1 Trillion Problem: Haske ba tare da Aiki
Haka kuwa lamba wanda ya kamata ya damuna kowa: 65% na kamfanoni har yanzu suna amfani da bayanai don tabbatar da shawarar da suka yi, maimaimakon barin bayanai ya jagoranci shawara (Gartner, 2024). Suna da dashboards na Business Intelligence. Suna da ƙungiyoyin Data Science. Amma bayananai ba su canja hali.
The Analytics-Action Gap
- BI ya ce: "Siye-siye ya ragu da 12% a Q3."
- Data Science ya ce: "Siye-siye zai ragu da 8% a Q4."
- Ko daya daga cikinsu bai ce: Aikin gani, ko sakamako zai zama, ko yadda za a san idan ya aiki.
McKinsey ya kiyasta cewa wannan analytics-action gap na kamfanoni na kasuwanci $3.1 trillion a kowace shekara a bayanai mara amfani.
Wannan shine matsala wadda Decision Intelligence ta warke. Ba ta hanyar kara dashboards ko namun daji na ML — amma ta sake tsara tsarin daga bayanai zuwa aiki zuwa kimantawar sakamako.
Tarihin Mai Girma: Daga Decision Engineering zuwa Decision Intelligence
Tushen tunani na Decision Intelligence ya koma shekara ta 1950 — zuwa lokacin da ya samar da artificial intelligence, operations research, da aikin Herbert Simon na Nobel Prize-winning akan bounded rationality. Amma ilimin zamani ya fito daga hanyoyi biyu:
The Academic Track
Dr. Lorien Pratt (Rutgers PhD, tsohon binciken DARPA) ya fara amfani da kalmar "Decision Engineering" a shekara ta 2010, daga baya aka canza zuwa "Decision Intelligence" a shekara ta 2012. Aikinta ya haɗa machine learning, causal reasoning, da shawarar yanke shawara na ƙungiya a cikin ilimin injiniya.
"Kalmar 'Decision Engineering' ba ta siyayya. Mun canza dukkan kayan talla da matsayinmu."
The Industry Track
Cassie Kozyrkov (Duke PhD, statistician) ya gina aikin Decision Intelligence na Google daga shekara ta 2018 zuwa 2023. Ta horar da dubban Googlers a cikin hanyoyin DI, tsakanin Research/ML da aikin kamfanin. Google ya ce "Decision Intelligence Engineering."
"Kimiyyar data plus kimiyar al'umma da gudanarwa."
Haduwarsu ta faru saboda hanyoyi biyu sun kai ga irin wannan katangare: sophisticate na fasaha ba tare da tasirin shawara. Aikin Pratt na ilimi ya nuna me ya sa haka; aikin Kozyrkov na masana'antu ya nuna yadda ake warke shi a kowane mataki.
Business Intelligence vs Data Science vs Decision Intelligence
Mai sauƙi ya fahimtar DI shine ta hanyar kwatance. Ga yadda ilimomin uku ke bambanta:
| Bangare | Intelligence Kasuwanci | Kimiyar Data | Hukumar Hukunci |
|---|---|---|---|
| Tambaya Ta Tsakiya | "Me ya faru?" | "Me za mu faru?" | "Me ya kamata mu yi?" |
| Nau'in Kimiya | Bayanai | Shineka | Shawarwari + Sakamako |
| Fito | Rahotanni, dashboards | Mudugai, hasashen | Hukunci + sakamako |
| Muhimmin Lokaci | Gabata/present | Zuwa | Kuntuni (gabata → aiki → zuwa → koyo) |
| Rawar Dan Adam | Tafsiri rahotanni | Tafsiri hasashen | Mallaki, ƙa'idoji, maye |
Mai mahimmanci: DI ba ta maye gurbin BI ko Data Science — ta kammala su. BI ta bayar da bayanai na tarihi. Data Science ta bayar da haske. DI ta kara shawarar yanke shawara, shawarar aiki, da mafita wadda ta kammala katangare tsakanin haske da tasiri.
Tsarin Decision Intelligence
A tsakiyar DI, yake aiki ne akan tsarin mai sauƙi amma mai ƙarfi:
Kallo
Tattara bayanai game da yanayin yanzu
Tsari
Tsara alakar da ke da sababi
Hukunci
Zaɓi aiki da sakamako da aka hasashen
Koyo
Auna sakamako, sabunta tsari
The Decision Intelligence Loop: Observe → Model → Decide → Learn → (mara dubu)
Haka yake kama da OODA loop (Observe-Orient-Decide-Act) daga tsarin soja. Amma akwai bambanci mai mahimmanci: Kimanta step. OODA an tsara shi don shawarar yanke shawara a lokacin yaki inda ba za ka iya tsaya don kimanta sakamako. DI an tsara shi don shawarar yanke shawara na ƙungiya inda za ka iya — da za ka — kimanta sakamako na tsari.
Causal Decision Diagrams: Gani Alakar Sababbi da Sakamako
Ginin Decision Intelligence shine causal reasoning — fahimtar ba dai alakar abin da ke da alaka ba, amma abin da ke sa abin da ke faruwa. Wannan shine bambanci tsakanin:
Correlation-Based Analytics
"Abokan ciniki wanda suka siya samfurin A suna da alaka da siye samfurin B."
Matsala: Idan mu ka siyar da B, ko siye A zai karu? Ba mu san.
Causal Decision Diagram
"Ragowar farashin A → karuwar siye A → karuwar siye B (amfani da abin da ke da alaka)."
Aiki: Muna sanin abin da za mu yi (farashin A) da tsarin (abin da ke da alaka).
Causal Decision Diagram (CDD) yake nuna alakar sababbi da sakamako. Yake nuna:
- Manufofi: Me muke neman samun sakamako
- Muhimman Ayyuka: Me ayyukan za mu iya yi
- Manyan Hanyoyi: Jerin tasirin da ke tsakanin muhimman ayyuka da manufofi
- Sauran Abubuwa: Abubuwa ba za mu iya kontrol ba amma dole mu kula da su
"Ya fi kyau a tsara bayanai kusa da shawarar da ake yi, maimaimakon kusa da bayananai da ke kewaye da shawarar."
Inda AI Yake: Karbuwa, Ba Mai Maye
Wannan shi ne inda Decision Intelligence ke yi wa daban da duka "AI za ta yi komputa komai" na hype da "mutane dole su yi shawara" na al'ada. Matsayin DI: AI ta karfafa shawarar mutane; mutane suna da alhakin.
Abin da AI Yake Yi Cikin DI
Haɗin Bayanai
Gudanar da kundin bayanai mara iyawa ga dan Adam. Tattara bayanai daga takardun 10,000 zuwa hasashen da suka dace.
Gano Al'ada
Nemo alaƙa da ban mamaki a cikin bayanai na dimension da dan Adam zai manta.
Gudanar da Sakamako
Tsara 'me idan' scenarios saurin gaske fiye da aikin dan Adam na hannu.
Abin da Mutane Ke Yi Da AI Ba Zai Iya
Ka'idoji & Ƙa'idoji
Hukunci me maye zai dace. Daidaita maslahar mai adawa.
Muhimmi & Hukunci
Amfani da ilimin ƙungiya, wayar da kan alaka, da wayar da kan hali.
Mallakar
Mallaki hukunci. Kai dan Adam a cikin loop da masu kula da shari'a ke bukata.
Netflix yake da kyakkyawar misali. injin shawarwarinsa (AI) ya yi aiki da salon kallo don mahalarta 300 miliyan. Ya yi hasashen House of Cards zai yi nasara kafin a yi wani labari. Amma mutane — masu gudanarwa na kamfanin — suka yi shawarar gaskiya ta amincewa da shirin ne bayan an yi shi da dala miliyan 100. AI ta yi aiki da kundin data; mutane suka yi aiki da alhaki.
80% na abin da ake kallo a Netflix ya fito daga injin shawarwari. Amma Netflix ta ce "mutane, ba na'ura, su ne manyan masu shawara".
Muhimmancin Amfani 2025-2030
Decision Intelligence ta koma daga nazari zuwa amfani a kamfanoni saurin fiye da yadda aka saba:
Yanayin Yanzu (Gartner, 2025)
- 33% na ƙungiyoyi sun ƙaddamar da Hukumar Hukunci
- 17% sun yi alkawarin yin jaribai a cikin watanni 6
- 19% suna la'akari da ƙaddamarwa a cikin watanni 6-12
- 25% suna binciken don 12-24 watanni
- Kawai 7% sun ce ba su da sha'awar haka
Kiyasin Kasuwa
Gartner AI Hype Cycle 2025 ya amince da Decision Intelligence a matsayin fasaha mai canji — wanda yake a 5-20% na yanzu amfani tare da haɓaka zuwa matsakaici a cikin shekaru 2-5. Kamfanoni da suka gina DI yanzu za su sami hanyoyin da suka dace da ilimin ƙungiya lokacin da zai zama abin da ake bukata.
Daga Collaborative Decision Making zuwa Decision Intelligence
Idan kuka karanta Collaborative Decision Making guide, zaku gane tushe DI take gina:
Abin da CDM Ta Kafa
- Muhimman ra'ayoyi sun doke hukuncin mutum (Condorcet, 1785)
- Aminci na tunani ya sa ra'ayoyi suka taru (Google Project Aristotle)
- Muhimman matakai na neman ilimi
- Kazalika da ra'ayoyi za a iya rage su da tsarin da aka tsara
Abin da DI Ta Kara
- Karbuwa ta AI: Gudanar da bayanai mara iyawa ga dan Adam
- Tsarin Sababi: Tsara alakar da ke da sababi don 'me idan' analysis
- Muhimman Loop: Kimiya da aka tsara don auna sakamako
- Karbuwa ta Hukunci: Hukunci na yau da kullun an gudanar da shi ta AI tare da kulawa ta dan Adam
Kalli shi a matsayin haka: CDM shi ne tushe na mutane; DI shi ne tsarin da aka gina akan tushe. Ba zai yiwuwa a yi DI mai kyau ba tare da ka'idojin shawarwarin ƙungiya. Amma zai yiwuwa a kara karfin CDM ta hanyar kara DI.
Yadda Argumentree Ke Yi Amfani da Decision Intelligence
Argumentree ta yi amfani da ka'idojin Decision Intelligence don shawarwarin ƙungiya. Idan ba a yi shawara a matsayin abin da aka yi ba, kamfanin ya yi shawara a matsayin tsarin da aka yi don koyo:
Sakamakon: kowace shawara ta zama damar koyo. Ƙungiyoyi suna gina ilimi. Membobin sabon za su iya fahimta ko dai me aka yi shawara, me ya sa aka yi shawara — da ko hujjojin da aka yi sun dace da gaskiya.
Gidan Shawarwarin
Wannan shafi ya kunno kai da muhimman abubuwa na Decision Intelligence. Domin koyo mai cikakke — gami da gine-ginen tushe, hanyoyin aiwatarwa, zane-zanen tsarin gano sababbi, da haɗin gwiwa tare da BI/DS — dubi kayan aikace namu:
Me Yake Decision Intelligence?
Gidan shawarwarin
5,000+ kalmomi suna kunno kai da DI: asali, gini, hanyoyin tsarin gano sababbi, hanyoyin haɗin AI, aiwatarwa na ƙungiya, da binciken da aka yi.
Karanta Guide Mai Cikakke