From Live AMA to Sage — how Saywise's three pivots led to the Career Management Agent
Three Pivots, One Path
Oct 1, 2026 · @HJ
From Live AMA to Sage — how Saywise's three pivots led to the Career Management Agent
Wisdom locked in people's heads became conversation. Conversation became content. Content became evidence. Now Sage reads that evidence — and starts moving toward a person's next opportunity.
Hanjong Lee · GTM & Evidence Builder
Changing course is not the same as losing your way
Startup histories usually get told as a straight line. Find a problem, build a product, gather users, grow.
Real voyages are messier. The first hypothesis wobbles the moment it meets users. A well-built feature can still fail to give people a reason to come back. When that happens, founders don't just fix the product — they rewrite the question.
Who is this really for?
What problem do they actually want solved?
How does what we build change their life?
Saywise changed course three times. More precisely: four product stages, three pivots.
Stage
Core product
Problem it set out to solve
Smallest unit of value
1
Live AMA
People's wisdom stays trapped in their heads
One good answer
2
AI Content Producer
Turning your own thinking into content is hard
One post or video
3
Resume 2.0
A résumé can't show how you think or what you've actually built
One piece of evidence
4
Sage Agent
Job seekers do every search and application by hand
Applications, interviews, offers
On the surface, four different products. Underneath, one question Saywise never let go of:
How do we discover the wisdom and potential that live inside people — but haven't yet been expressed — and connect them to the world?
The pivots weren't a retreat from that question. They expanded the answer: wisdom → expression → evidence → action.
Stage one: Live AMA — drawing wisdom out through conversation
In early 2025, Saywise launched as a Live AMA platform: experts took questions over live video and shared what they'd learned. CK, who leads Saywise, called it “Clubhouse meets LinkedIn.”
The premise was clear. The world is full of valuable experience, but most of it never leaves people's heads. Writing requires choosing a topic and polishing sentences. Answer a good question, though, and people talk about what they've lived through with surprising ease.
The flow was simple.
Someone with experience opens an AMA.
Participants ask questions.
The expert answers from experience.
AI breaks the conversation into bite-sized Q&A.
Those answers become knowledge others can discover.
The most valuable thing Saywise found wasn't the video. It was this: questions draw experience out of people.
Ask “Tell me about your expertise,” and you get abstractions. Ask “What was the hardest call you made on that project?” and you get context, trade-offs, failure, and learning.
But Live AMA had a structural problem. Good hosts and good askers had to show up at the same time. Before each event, you had to fill the room; after it, you had to give people a reason to return. Until the community reached scale, the value any single user got was limited.
The wisdom was there. The machinery for drawing it out was heavy.
Pivot one: from human-to-human to human-to-AI conversation
The problem was never a shortage of wisdom. Quite the opposite — there was plenty. Getting it out just took too many people, too much time, too much scheduling.
So instead of waiting for people, Saywise let AI ask the questions. It moved from a community platform that needed many people at once to a self-serve product anyone could start at any time.
Stage two: AI Content Producer — turning conversation into content
In this stage, Saywise aimed to be a personal content team: AI interviewer, video producer, and ghostwriter in one. The pitch at the time was “Jarvis for Social Media.” You talked with the AI; Saywise found the core ideas and turned them into a range of outputs.
LinkedIn posts and short video clips
Video scripts and teleprompter copy
Image ideas
Insights pulled from long conversations
The hypothesis was grounded. Plenty of professionals have experience worth sharing but can't produce content consistently. Picking a topic is hard, structuring a thought is harder, and speaking naturally on camera is hardest of all. An AI that asked the right questions and organized your words could lift much of that weight.
In practice, though, a new gap appeared.
December 2025: a principle found in Conversation
I used Saywise's Conversation feature in two distinct modes.
Mode 1: talking before my thinking was settled. I spoke with the AI before I'd reached a conclusion. It was great for surfacing new angles and testing counterarguments. But because I was thinking out loud, I repeated myself and the message scattered — hard to use as video.
Mode 2: talking after the writing was done. Once I'd worked the idea out in writing, I had the same conversation again. This time the message was sharp and the delivery natural. I knew what to emphasize, so the result was far more usable.
That experiment surfaced a principle:
AI is not an Insight Generator. It is an Insight Organizer.
Insight has to start with the person. AI's job isn't to invent thoughts that weren't there — it's to find the thoughts a person actually has, give them structure, and carry them into the next piece of work. So I proposed this flow:
The user explores freely in conversation with the AI.
The AI summarizes the main themes and insights.
The user personally chooses the point they want to make.
The AI shapes that chosen insight into a post and a video script.
The user restates the refined idea in their own voice.
The finished piece goes out into the world.
The heart of it was the user's choice. A sentence the AI picks automatically doesn't become the user's thought. AI proposes candidates; the user decides, “This is what I'm really trying to say.” That principle carries straight through to Sage today.
Pivot two: from volume of content to evidence of expertise
The AI content market got crowded fast. Countless services could generate posts, video, images, and scripts. Content got easier to make, but more content didn't reliably turn into revenue, funding, partnerships, or hires.
Saywise moved to a deeper question. Not how to make more content, but what that content proves about a person.
A post was no longer just something to rack up views. It became evidence of expertise — proof of what someone had experienced, how they'd judged, and what they'd actually built.
Stage three: Resume 2.0 — turning content into a living professional profile
Resume 2.0 set out to restore the context a traditional résumé erases. A résumé compresses a person into company names, titles, dates, and a few lines of results. Something important gets lost along the way.
What problem did you spot?
Why did you make the call you made?
What did you actually build?
What did you change after it failed?
Who can verify the outcome?
How does that experience connect to the next problem?
Saywise tried to show all of this through projects, stories, Q&A, video, recommendations, and examples of how people used AI. CK described Resume 2.0 as a “living profile for every other knowledge worker” — what GitHub is for engineers and Dribbble is for designers.
In the content stage, the smallest unit of value was a single post or video. In Resume 2.0, it became a single piece of evidence behind a person's ability.
“AI GTM expert” is just a claim. Connect it to the problem being solved, the organizational and market context, your role, the calls you made, the work you produced, and measurable results — and it becomes evidence. Then mark what's been externally verified, what hasn't yet, and what can't be disclosed. That is where what I call the Evidence Graph begins.
What 100 entries revealed
During the Resume 2.0 stage, I left more than 100 entries on Saywise. I wrote about AI and hiring, technical sales and materials R&D, startups and investing, global market entry, film and human dignity.
At first they looked like unrelated topics. But as they piled up, a recurring way of thinking emerged.
Asking questions that see possibilities before others do
Translating complex technology into human language
Seeing Korea and the U.S. as markets to be bridged
Turning ideas into real, shipped work
Not hiding failure or the revisions that followed
Refusing to judge people by title and productivity alone
“GTM & Evidence Builder” wasn't a label I picked because it sounded good. It was a name I found only after reading the entries side by side.
I didn't complete my identity before I started writing. I discovered who I was by writing.
That was the most human promise of Resume 2.0.
But a good profile wasn't enough. Once it was done, there was little reason to come back. If companies didn't find the profile, no opportunities followed. Even when they did, the searching, applying, interview prep, and follow-up still fell on the user.
Resume 2.0 could explain a person deeply, but it couldn't act on their behalf. It was a great record — not yet an engine for action.
Pivot three: from a profile that shows you to an agent that moves for you
The question shifted again — from “How do we show who this person is?” to “How do we find the right opportunities for this person and turn them into real action?”
If Resume 2.0 explained a person's past and present, Sage uses that understanding to find their future.
Stage four: Sage — handing the full-time job of job hunting to AI
Today's job search is a string of disconnected chores. You cycle through job boards, wade through recommendations, and weed out mismatches, stale postings, and dead links. When you finally find something promising, you research the company, rework your résumé, and write the application. And nothing moves until you do.
Unless you do the work, nothing happens.
Generative AI lowered the cost of applying — and created a new problem. Anyone can now tailor a résumé in minutes and auto-apply, so applications have exploded. Candidates apply more but don't feel more seen. Companies receive more applications but aren't more confident in their judgments. And every submission disappears into a black hole where you can't even tell whether anyone read it.
CK isn't trying to solve this with a better search engine. What he proposes is a personal recruiter who works for exactly one person: you.
What Sage actually does today
The texts Sage sends already show an early form of this vision. It doesn't just send a job title and a link. It summarizes the company's funding stage, recent hiring activity, compensation, location, and the shape of the role — then explains which parts of your experience fit.
Hightouch: Sage noted its YC S19 and Series D background, its hiring activity over the past 30 days, and a business that combines customer data with AI — and tied it to my track record of translating complex technology into measurable business value.
Kong, enterprise role: Sage flagged compensation of roughly $320K–$400K, account expansion, and the value of open-source, API, and infrastructure software experience — and noted the posting had gone up just an hour earlier. Then it asked whether I wanted my résumé tailored for the role.
Arkestro, Solutions Consultant: Sage laid out the salary range, remote setup, and the context of a procurement platform built on AI and game theory, and showed the fit with my experience translating technology into business value and leading commercial cycles.
More interesting: Sage also tells you the downsides. In one recommendation, it said my technical sales and AI GTM background fit well — but the level might be a step below where I am. A good agent doesn't push you to apply to everything. It explains the opportunity cost along with the upside.
After each recommendation, Sage asks for feedback on role, location, pay, and overall fit. Those answers become signals that teach it my preferences.
The early-product seams still show. International texts arrive split into fragments, and carrier warnings about overseas senders interrupt mid-sentence. But even through that clunky screen, the shift is unmistakable.
I never went to a job site — and the opportunity came to me first.
I wasn't someone waiting on search results anymore. I was someone with an agent watching the market on my behalf.
Not 25 search results — a few opportunities worth your time right now
A good personal recruiter starts by understanding the person. What you've done and what you're good at. Where and how you want to work, what level and pay you're targeting. What you want from your next role and what you won't compromise on. Whether you're actively looking or just open to the right thing.
Then it watches the market for you. Instead of dumping dozens of algorithmically ranked postings, it picks the few worth your time and tells you the moment they appear — so you can move before you're applicant number 300.
Once you decide to apply, the agent takes on the repetitive work: researching the company and role, identifying the hiring team and decision-makers, tailoring the résumé, surfacing relevant work and stories, drafting the application, following up, preparing for interviews, scheduling, and supporting salary and offer negotiation.
Saywise now defines Sage as a fully agentic job search and application platform. The future of job hunting isn't a better search bar. It's having an agent that does the searching, preparing, and repetitive work for you.
Sage's bigger future: from job search agent to Career Management Agent
Job searches happen in bursts across a career. If Sage's role ends the day you're hired, you'll leave until you need it again. Subscription retention weakens, and everything Sage learned about you goes cold.
So Sage's long-term vision has to go beyond a Job Search Agent and become a Career Management Agent. The job search is just one scene. A true Career Management Agent stays with you across the whole arc — before the search, during it, after you're hired, and into your next move.
That work runs as a single loop:
Capture → Structure → Prove → Discover → Act → Grow → Learn
Step
What Sage does
What the person does
Capture
Surfaces experience missing from the résumé through questions and short check-ins
Answers honestly and corrects
Structure
Organizes experience into problem, judgment, work, and result
Decides the facts and what can be shared
Prove
Links work, references, and outside validation; diagnoses market position and evidence gaps
Chooses the next goal
Discover
Watches markets and companies continuously; shortlists the right opportunities
Judges which ones are worth pursuing
Act
Builds evidence-based applications, follow-ups, interview and negotiation prep
Approves what to emphasize; makes the final call to apply or accept
Grow
Records post-hire results, feedback, and learning
Does the actual work and grows
Learn
Feeds user and employer responses into the next recommendation
Decides whether to stay or move
Capture & Structure — remembering your career
People forget what they've accomplished. When a project ends, the problem, the calls, the small contributions all scatter. A year later, rewriting a résumé, only a faint outline of the result remains.
Sage can be your career memory. Instead of reconstructing your roles, key decisions, wins and failures, feedback, metrics, and lessons all at once, it builds them up continuously through short check-ins.
“What was the most important decision you made this week?”
“What did you personally produce on this project?”
“Is there a person or artifact that can verify the result?”
What accumulates isn't a résumé rushed together before a job change. It's living Career Evidence. The Live AMA lesson — good questions draw experience out — is at work again here.
Prove — building evidence and diagnosing where you stand
Sage shouldn't just recommend postings; it should explain your position in the market. Which roles are you strong for, and where do you fall short? Is there a gap between the salary you expect and what the market pays? What evidence is missing for the next step?
An agent shouldn't push everyone to switch jobs. Sometimes the best advice is: “Before applying anywhere, put numbers on what you're delivering in your current project.”
Discover — sensing the market, continuously
You don't have to check job boards every day; Sage watches for you. New roles that fit, companies raising and hiring, rising demand for specific skills, new job categories, shifts in pay and work arrangements, even moves within your network. Even if you're not actively looking, it tells you when something good appears.
What matters is precision, not volume. One posting that could change the direction of your life is worth more than a hundred you'll never open.
Act — arranging evidence for each application, and managing what comes after
Sage should be an Evidence Organizer, not a résumé generator. Rather than inventing plausible sentences from a posting's keywords, it should find relevant evidence in the life you've actually lived.
Analyze the core problem the role is meant to solve.
Find relevant experience across your profile, projects, writing, and work.
Separate verified facts from self-reported claims.
Show your strongest evidence alongside the gaps.
Let the user choose what to emphasize.
Reframe the chosen evidence in the posting's language — and disclose what changed from the original.
Submit only what the user has approved.
This is the Insight Organizer principle from the 2025 Conversation experiment, reborn inside a hiring product.
Most job tools stop the moment you click Apply. But real hiring starts after that. Sage should manage application status and deadlines, recruiter contact, follow-ups, reference requests, interview scheduling and interviewer research, and even relationships after a rejection. The goal isn't to send more applications. It's to create more conversations where companies and candidates actually understand each other.
The same goes for interviews and negotiation. Sage should help candidates evaluate companies, not just help companies evaluate candidates. It prepares likely questions and the evidence to answer them, honest framing for experience gaps, and a read on the business model and its risks. When an offer arrives, it compares base and equity, title and real authority, team stability, the impact on family and health, and what the role sets up next. The highest-paying offer isn't always the best career move.
Grow & Learn — staying with you after you're hired
A true Career Management Agent doesn't disappear once you land the job. It helps with your 30-60-90 day plan, a map of new stakeholders, early wins, the evidence you'll need for promotion, quarterly career reviews, learning plans, and weighing internal moves against outside offers.
That's when Sage stops being a job search agent and becomes a career operating system. You focus on your work; Sage remembers the evidence that would otherwise scatter and prepares your next choice. Each turn of the loop, Sage understands you more deeply. And the deeper that understanding, the clearer the boundaries of authority and responsibility must be.
Decisions stay human. The work goes to the agent.
The most important line in CK's vision may not be the slogan at the end. For me, it's this:
You would still make the decisions. But you wouldn't have to do all the work.
That sentence is both Sage's product principle and its ethical boundary. An agent can find, organize, prepare, and execute on your behalf. It must not decide the direction of your life and career for you.
What kind of work do you want?
Which city will you live in?
How will you weigh family and health?
Which company values can you live with?
How much risk will you take on?
What will you share, and what will you keep private?
Those decisions have to stay with the person. The more work AI takes on, the more strongly people's rights to choose, to correct, and to refuse must be designed in.
A good Sage shouldn't be measured by how many applications it fires off. It should be measured by how accurately it understands you, how faithfully it holds to your standards, and how transparently it explains why it recommended an opportunity.
What “Never search for a job again” really means
It doesn't mean you'll never have to think about your career. It isn't a call to hand every decision to AI. It's closer to a promise: you won't burn out on searching and busywork before you ever get the chance to prove what you can do.
AI does the searching; you make the choice. AI helps you apply; you live the career.
And someday, the line may stretch wider:
Never start your career from scratch again.
A world where you never have to explain your career from zero again. If Sage keeps remembering your experience, judgment, work, and growth, you won't have to reinvent yourself in front of a blank page every time you change jobs. The time you've already lived becomes the starting point for your next opportunity.
When three pivots connect into one agent
Now Saywise's four stages read as a single flow.
What Live AMA left behind: experience comes out when it meets a good question. (Capture)
What AI Content Producer left behind: AI shouldn't replace a person's voice; it should structure scattered words. (Structure)
What Resume 2.0 left behind: ability shows up more clearly in projects, judgment, work, and stories than in titles and keywords. (Prove)
What Sage has to do: connect that evidence to real opportunities, action, and long-term growth. (Discover → Act → Grow → Learn)
Once the loop is complete, the three pivots aren't a graveyard of discarded products. They're layers laid down, one after another, to build a Career Evidence Infrastructure that understands a person over the long run.
What 100 entries now have to prove
The 100-plus entries I left on Saywise are one test case for this vision. If Sage only reads an existing résumé and spits out a tailored version, it's not much different from any other AI job tool. But if those 100 entries help it understand the following, the story changes.
How I connect technology and markets
Which problems I keep coming back to
Why I arrived at the identity of GTM & Evidence Builder
How I see the Korean and U.S. markets
What work I kept doing through failure and recovery
Which organizations and roles make my experience land hardest
What values I want to protect in my next role
When that understanding turns into sharper matches, more honest résumés, more persuasive interviews, and better career choices, those 100 posts stop being self-promotion. They become evidence that improves Sage's judgment.
The question is simple: Do users with rich records actually find better opportunities and make better career decisions? When Saywise can answer that with results, Resume 2.0 and Sage become one.
Three pivots, one path
Saywise's new beginning isn't about launching a fourth product. It's that, for the first time, everything learned from the first three can connect to a person's entire career: the questions from Live AMA, the structure from AI Content Producer, the evidence from Resume 2.0, and the action and growth Sage creates.
The question Saywise started with is still here: how do we draw out the wisdom inside people that hasn't yet been expressed? Now a bigger question follows it: how do we help that wisdom and potential become a real next opportunity — and that opportunity become new evidence of growth?
From words to records. From records to evidence. From evidence to opportunity. From opportunity to growth. And from growth, back to better choices.
Changing course three times doesn't mean the ship returned to port. It crossed the same sea in different ways and learned what it was meant to carry. A pivot isn't a confession that the past was wrong. It's what happens every time you find a more important problem and rewrite the question.
What Saywise has to prove now isn't the number of features or the volume of auto-applications. It's that understanding people more deeply leads to better opportunities and better career decisions. The day that's proven, Saywise's three pivots will read not as wandering, but as a voyage to find one path.
And maybe the real Saywise story is only beginning.
The one-line takeaway — Saywise's three pivots turned human wisdom into expression, expression into evidence, evidence into opportunity, and opportunity into lasting career growth.
#Saywise #SageAgent #CareerManagementAgent #AgenticAI #NeverSearchForAJobAgain #NeverStartFromScratchAgain #StartupPivots #BuildInPublic #Resume2_0 #EvidenceGraph #HumanInTheLoop #FutureOfWork #FutureOfHiring #CareerGrowth