I ka baara kura ye ka AI Agentw don kurun kɔnɔ: AI Native Companies bɛ baara kɛ cogo min na tiɲɛ na

In This Article

Abonnés kanulenw,n ye kalo damadɔ tɛmɛnenw kɛ ka ɲɛmɔgɔw ɲininka AI-native sosiyetew la. Ne dalen b’a la sisan ko:AI cidenw donni ani u ɲɛnabɔli YE baara ye, i ka baara mana kɛ fɛn o fɛn ye.Nin jikuru juguba in na, n b’a fɛ k’a jira cogo min na AI-native tɔn saba —Linear, Ramp, ani Factory — ye nin sariyakolo in bila baara la. Nin ye kuma dɔw ye minnu bɔra u kelen-kelen bɛɛ la:Nan Yu (Head of Product, Linear): “I bɛna AI ekipu mɔgɔ dɔw sɔrɔ i bɛ se ka baara di minnu ma ani ka kuma u fɛ i n’a fɔ i bɛ kuma mɔgɔw fɛ cogo min na.”Geoff Charles (CPO, Ramp): “N’i tɛ baara kɛ ni Claude Code ye, i jɔyɔrɔ mana kɛ fɛn o fɛn ye, n’a sɔrɔla i tɛ baara kɛ ka ɲɛ.”Eno Reyes (CTO, Factory): “An ye codified product management, frontend UI, data analysis, and more into reusable skills that any employee can invoke.”A kalan ka taa a fɛ walasa ka kɔnɔna lajɛ AI-native company kelen-kelen bɛɛ bɛ baara kɛ cogo min na.Linear: Making AI agents first-class teammatesI bɛ se ka @mention an agent to create and assign an issue to another agentLinear’s approach to agents is shaped by their product. Nan (Linear’s Head of Product) dalen b’a la ko:Agents ka kan ka kɛ baarakɛlaw fɔlɔ ye. I ka kan ka se k’u fara porozew kan, k’u bila koɲɛw la, k’u fɔ kumasenw kɔnɔ.Nka, Nan dalen b’a la fana ko hadamaden bɛ to ka jatebɔ kɛ tuma bɛɛ a kɔlɔlɔw la. Linear bɛ fɛnw dilan cogo min na ni cidenw ye senna-tɛgɛrɛ bɛɛ la, o filɛ nin ye:Gɛlɛya faamuya. Cidenw bɛ kiliyanw ka baro bɛɛ kalan ani k’a lajɛ ka surunya ka bɔ Intercom, Zendesk ani Gong la. U bɛ koɲɛw dabɔ u yɛrɛma, k’u de-dupe ka ɲɛsin kɔkankow ma, k’u bila ekipu ɲuman na. A furakɛcogo dɔn. Ikomi cidenw bɛ se ka kiliyanw ka barow sɔrɔ, u bɛ se k’i dɛmɛ ka segin-ka-bɔnye kɛ spec dɔ kan ni kunnafonidilanw dɛmɛni hakilinaw samali ye ka bɔ sira caman kan. Aw ye fɛɛrɛ tigɛ. Agentw bɛ se k’i ka spec kari ka kɛ tikiti bɛton ye k’u bila ekipu ɲumanw na otomatiki la. Linear la, sisan, cidenw bɛ tikiti fanba dilan. Hadamaden ka baara ye k’u ka baara lajɛ ani k’u kɔnɔkow ladilan waati kɔnɔ.Execute. Bugw ni fɛn misɛnninw bɛ Di ka ɲɛsin cidenw ma i n’a fɔ Codex ani Cursor. Walasa ka fɛn gɛlɛnw sɔrɔ, ɛntɛrinɛti injiniyɛriw bɛ Claude Code daminɛ ani ka Linear MCP kɛ ka bɔko dafalen sama.Ka bɔ Nan fɛ: A bɛ kɛ i n’a fɔ cidenw bɛ se ka min ɲɛnabɔ, o bonya bɛ ka bonya kalo saba o kalo saba. Modeli kura ni harnesses bɛ ka dancɛ gɛlɛya ka bɔ fixes nɔgɔmanw na ka taa poroze gɛlɛnw na ka taa a fɛ.I b’a fɛ ka jɔ ni cidenw ye i n’a fɔ Linear? Nin ye fɛɛrɛ 4 ye minnu bɛ se ka kɛ, Nan ye minnu tila i ka kulu bɛ se ka min kɛ bi:Developer bɛɛ ka kan ka default kɛ agent coding tool ɲɛmɔgɔ dɔ la. Nin ye fɛɛrɛ fɔlɔ ye min ka nɔgɔn. Baarakɛminɛn ofisiyali (Cursor, Claude Code, walima Codex) di ani k’a ɲɛnabɔ walasa i ka se ka baarakɛcogo ye.Supplement with an async cloud coding agent. Async kɔkanna agents bɛ se ka fɛn misɛnnin caman caman cili kelen-kelen ani bug fixes. Cursor ni Devin bɛ ni saraka ɲumanw ye yan.Aw ka sinsin a kan ko dilanbagaw ni PMw ka baara kɛ k’a ɲɛsin kodebase ma. Agent minnu bɛ i n’a fɔ Claude, olu bɛ sira dɔ da wuli min ka dɔgɔn PMw ni dilanbagaw ye walasa u ka se ka fɛn caman Changer directement codebase kɔnɔ. Bɛɛ ka kan k’a cɛsiri ka kɛ jɔlikɛla ye.PMw ni sugukɛlaw ka kan ka default kɛ AI interface dɔ la. O baara ninnu ka kan k’u ka baara 80-100% kɛ kumaɲɔgɔnya-yɔrɔ dɔ fɛ — a kɛra Claude ye, ChatGPT, Notion AI, walima o ɲɔgɔnna dɔ. Nan bɛ siniɲɛsigi ye hadamadenw bɛna jɛ ka baara kɛ ni cidenw ye yɔrɔ min na spec nivo la — ka min ɲɛfɔ ka kan ka jɔ ani mun na — o kɔfɛ ka fɛnw tɛmɛn cidenw ma walasa u ka fɛn bɛɛ ɲɛnabɔ duguma.Ramp: Measuring AI proficiency in 4 levelsRamp is rapidly pushing their employees to be AI-native in 4 levelsNi Linear b’a jira cogo min na ka cidenw kɛ i ka kulu kɔnɔ mɔgɔ fɔlɔw ye, o tuma na Ramp b’a jira cogo min na i bɛ se k’i ka sosiyete bɛɛ sɔrɔ k’u ta.San 2025, Ramp ye fɛn 500+ ci, ka se $1B sɔrɔta ma, k’o bɛɛ kɛ ni PM 25 ye.U y’o bɛɛ kɛ k’a ɲini baara kelen-kelen bɛɛ fɛ (eng, product, design, sales, marketing, legal, finance) ka don kurun kɔnɔ ani ka baara kɛ ni cidenw ye.Geoff (CPO Ramp) ye framework dɔ tila ɲɔgɔn na walasa ka UI seko jateminɛ mɔgɔ kelen-kelen bɛɛ ye baarakɛla min n b’a ye ko a bɛ se ka kɛ kosɛbɛ:L0: Tuma dɔw la, a bɛ baara kɛ ni ChatGPT ye. A ka c’a la, o mɔgɔw tɛna kɛ sosiyete la waati jan kɔnɔ. N’i ​​tɛ i yɛrɛ daminɛbaga ye ni bonya hakilina ye AI baarakɛminɛnw kan, Geoff ko a bɛna gɛlɛya kosɛbɛ k’i dege ka tɛmɛn fɔlɔ kan.L1: A bɛ baara kɛ ni GPTw, porozɛw ani kɔnɔna AI baarakɛminɛnw ye ani k’u sɛmɛntiya. Nin mɔgɔw bɛ ka AI kɔlɔsi nka u ma baara lakika kɛ otomatiki ye fɔlɔ.L2: Build an app that automates part of theirbaara. O mɔgɔw bɛ se ka kode kɛ walima ka hakilina nafamaw di mɔgɔ wɛrɛw ka baara kan ni AI baarakɛminɛnw ye.L3: Sisitɛmu jɔbagaw. O mɔgɔw bɛ ka AI infrastructures ni sekow jɔ minnu bɛ mɔgɔ tɔw bɛɛ teliya ekipu kɔnɔ.Société ka laɲini ye ka bɛɛ gɛlɛya ka wuli lada la. L0w yɛrɛ bɛ sugandi ka bɔ. L1w bɛ kɛ L2w ye. L2w bɛ kɛ L3w ye. Wa L3w bɛ nɔ bila jɛkulu tɔ la. Geoff fana ye fɛɛrɛ 5 fɔ, baarakɛda o baarakɛda bɛ se ka minnu ta walasa ka kɛ AI-native ye:Remove all friction. Aw bɛ se ka AI baarakɛminɛnw sɔrɔ minnu bɛ fɔ kosɛbɛ, ni dan tɛ tokenw walima baarakɛnafolo la ani ka AI sekow kɔnɔna marayɔrɔ dɔ dabɔ mɔgɔ bɛɛ bɛ se ka min sama. Ni setup ka gɛlɛn, mɔgɔ fanba tɛna adopte.Make adoption visible. Foroba Slack kanw dabɔ, mɔgɔw bɛ se ka fɛn minnu jɔ, u bɛ se k’olu tila ɲɔgɔn na yɔrɔ minnu na. Bolo bɛɛ la, a jira ko mɔgɔ minnu tɛ ɛntɛrinɛti ye, olu bɛ fɛn kabakomaw kɛ, i n’a fɔ wariko nafolosɔrɔsiraw jɔli walima feereli min bɛ kɛ ka sitiw dabɔli otomatiki la.Aw ka dɛmɛ di bololabaarakɛlaw ma. Biro waatiw ladon, mɔgɔ bɛɛ bɛ se ka fara minnu kan walasa ka AI sekow ni baarakɛcogo jɔ. Aw ye AI dɔnnikɛlaw sugandi minnu ka baara bɛɛ ye ka kibaru duman fɔ, ka mɔgɔw sigi sen kan, ani k’u dɛmɛ u ka se “aha waati” ma. Track baarakɛcogo ani ka don a ko la. Ramp bɛ tokenw tacogo lajɛ AI baarakɛminɛnw bɛɛ kɔnɔ baarakɛla kelen-kelen bɛɛ la. Kunnafoni ɲɛmɔgɔya bɛ nin kunnafoniw tila ɲɔgɔn na walasa ka jatebɔ dabɔlen da ani ka sen don a la ni mɔgɔ dɔ ka baara dɔgɔyara.A kɛ baarakɛlaw ta ye. PM ka ɲininkaliw kɔnɔ sisan, kalan kɛrɛnkɛrɛnnen dɔ bɛ kɛ, i ka kan ka baarakɛminɛn dɔ jɔ min kɔnɔ, o kɔfɛ, i k’a ɲɛfɔ mun na i y’a jɔ ani a bɛ baara kɛ cogo min na. Geoff y’a ka ɲɛmɔgɔya hakilila lajɛ jɔyɔrɔ bɛɛ la Ramp la layini kelen kɔnɔ:I ka baara ye k’i ka baara kɛ otomatiki ye.Ka bɔ Geoff fɛ: “Ni n y’a fɔ n ka kulu ye siɲɛ 10 ko CTA ka kan ka kɛ foli sanfɛ, o labɛnni t’a fɔ ko a siɲɛ 11nan na. It’s encoding that feedback into an automated design crit process or AI skill so that it never happen again.”Factory: AI native kabini tile fɔlɔNi Linear ni Ramp b’a jira cogo min na sosiyetew bɛ AI cidenw ta, Factory b’a jira min bɛ kɛ n’i ye so jɔ u lamini na kabini tile fɔlɔ. Factory ye mɔgɔ 55 ka AI porogaramuw dilanko tɔn ye min nafa ye $300M ye min labɛnna AI lamini na kabini duguma. Nin ye fɛn ye min b’u kɛ danfara ye:Ka fɛn dilanni injiniyɛriw taFactory tɛ PMw ni ɛntɛrinɛti injiniyɛriw ta u danma. O nɔ na, ​​u bɛ fɛn dilanni injiniyɛriw ta minnu bɛ AI cidenw ɲɛnabɔ ani ka baara kɛ ni u ye. Don min bɛ kɛ cogoya la, o bɛ i n’a fɔ:Aw ye taamasiyɛnw sɛgɛsɛgɛ ka bɔ ciden boliw la walasa k’a dɔn sigida ye desizɔn juguw ta yɔrɔ min na. Labɛnw sɛbɛn tɛ i n’a fɔ kode, nka i n’a fɔ mara (misali la, seko dɔ ladamuni, lint sariya kura, walima otomatiki nɔgɔlen)Aw ye PR minnu dɔrɔn lajɛ, cidenw bɛ minnu taamasiyɛn ko faratiba bɛ u la (cidenw bɛ u janto tɔw la).I ka hakilina kuraw sugandi ani ka sɔsɔli kɛ fɔlɔfɔlɔko kan ni baarakɛɲɔgɔnw ni AI ye.Nin baara in man nɔgɔn, wa a bɛ dɔnniya jugu de wajibiya, nka leverage ye enormous.Make your codebase agent-readyAgents mago bɛ kodebase la u bɛ se ka baara kɛ min kɔnɔ tiɲɛ na walasa u ka se ka baara kɛ. Factory bɛ kodebasew jate kɔgɔli hakɛ duuru kɔnɔ, wa Level 3 ("Standardized") ye yɔrɔ ye min na ekipu fanba ka kan ka laɲini fɔlɔ.Factory’s agent readiness frameworkEncode expertise into reusable skillsNi i ka kodebase labɛnnen don ciden ma, fɛɛrɛ nata ye ka dɔnniya di cidenw ma walasa u ka se ka latigɛ ɲumanw kɛ sekow fɛ (a jɔyɔrɔba ye sɛbɛnni taamasiyɛnw dɔrɔn de ye). Factory bɛ baara kɛ ni sekow ye walasa ka dɔnnikɛlaw ni sosiyete ka dɔnniyaw kodɔn ka kɛ fɛn ye, ciden walima baarakɛla bɛɛ bɛ se ka baara kɛ ni min ye. Factory bɛ baara kɛ ni seko minnu ye a kɔnɔna na, olu lisɛli filɛ nin ye ani u bɛ u ka markdown filew ladɛrɛsi walasa i ka se ka kopi kɛ ani k’u sɛmɛntiya:Product management skill. Fɛn dilannenw sariyakolow, dolow 11 ka ko kɛlenw hukumu (a juru tara Airbnb ka Brian Chesky fɛ), PRD misali, jatebɔ rubriki, ani kanko ɲɛminɛni bɛɛ bɛ kɛ markdown file kelen kɔnɔ.Frontend UI integration skill. A bɛ Droid bilasira fɛnw jɔcogo kan ni Factory ka dilancogo ye, sirako bɛnkanw, ani sɛgɛsɛgɛli sariyaw.AI data analyst skill. Sɛgɛsɛgɛli sɛgɛsɛgɛli kɛ, ka yecogo dɔw jɔ, ani ka jatebɔ rapɔɔriw dilan ni Python ekosisɛti dafalen ye.Kɔnɔna baarakɛminɛnw seko. Aw ye admin panneaux, support consoles, ani operational dashboards jɔ ni doncogo ɲɛnabɔcogo ɲumanw ye ani audit logging baked in.Vibe coding skill. Rapidly prototype new web apps from scratch with modern frameworks.N’i bɛ se k’i ka mɔgɔ ɲumanw bɛ fɛn minnu dɔn, olu kodɔn ka kɛ sekow ye, i mago t’a la ka kɛrɛnkɛrɛnnenw ta baara bɛɛ la.6 steps to put all this into practiceTo recap:Onboarding and managing agents is becoming the core job for every function.Here are six things you can put into practice right now:FromLinear:Aw bɛ default (dafalen) kɛ developpeur bɛɛ la ka kɛ agentic coding tool ye i n’a fɔ Cursor, Claude Code, walima Codex. Aw ye PMw ni dilanbagaw bila ka don codebase kɔnɔ. A to u ka PRw ni kurun kode di ni cidenw ye. Yɛlɛma misɛnnin kelen-kelen bɛɛ bilali dabila ɛntɛrinɛti injiniyɛri dɔ fɛ.Ka bɔ Ramp la:AI ka seko suman i ka kulu kɔnɔ. Ramp’s 4 levels framework bɛ daɲɛ jɛlen di i ma mɔgɔw bɛ yɔrɔ min na ani u ka kan ka taa yɔrɔ min na. AI baarakɛcogo lajɛ ani k’a kɛ baarakɛcogo jigiyaw dɔ ye. I tɛ min suman, i tɛ se k’o ɲɛ ka taa a fɛ, wa dusudonw nafa ka bon. Ka bɔ Factory:Score i ka codebase walasa ka agent labɛn. Baara kɛ ni Factory’s agent readiness levels ye walasa k’a faamu ni i ka codebase labɛnnen don.I ka ekipu ka dɔnniya kodi ka kɛ seko ye min bɛ se ka baara kɛ kokura. I ka mɔgɔ ɲumanw bɛ fɛn minnu dɔn, i k’olu kodɔn seko ni dɔnko dosiyew kɔnɔ ani k’a nɔgɔya hadamadenw ni cidenw bɛɛ bolo ka baara kɛ n’u ye.Ka tɛmɛ o bɛɛ kan, kurun kɔnɔ cidenw i n’a fɔ i bɛ hadamaden don cogo min na. Aw ye contexte di u ma, k’u siri aw ka operational stack la, ka hadamaden dɔ to a ka jaabi di u ka sɔrɔw la.Aw ka hakilina di n ma commentaires kɔnɔ!

You May Also Like

13 Trending Songs on TikTok in March 2026 (+ How to Use Them)

Our Best AI PM Workflows for Prototyping, Strategy, and Personal OS (2026)

China’s Cybersecurity Agency Issues Security Warning About OpenClaw