Penn grad | AI x investing 🩵 | Community & Growth Building, testing, and learning @getDrivenAI (👉 https://nitter.cf/t.co/OwAgXKUQr3)
Joined September 2025
- Tweets280
- Following135
- Followers80
- Likes360
We’ve wanted @RobinhoodApp in Driven for a while.
A lot of our users already invest there, and we wanted to make it easier for them to bring that part of their investing workflow into Driven too.
Really happy to finally see it live.
If you use Robinhood, this one’s for you.
An earnings filing hits the SEC. Your Robinhood trade flow is already in motion.
Driven now connects directly to @RobinhoodApp for live trading.
With event-triggered Automations,
Driven can start the analysis automatically
when a new earnings filing appears,
then take you straight from insight to execution.
SEC filing → analysis → trade. $HOOD
👇 Watch the full flow
特别有质感的设计，来自Founding Members的夸奖☺️
Introducing Driven Founding Members. ✦
Among the first. Always part of the story.
Some of our earliest and most dedicated users have been with us from the beginning, exploring Driven, sharing honest feedback, and helping us grow.
Today, we're recognizing their support with something special.
100 Founding Member seats. By invitation only.
Each member receives a uniquely numbered Genesis Edition card, closer access to our team, and exclusive member benefits.
This is our way of saying thank you to those who believed in Driven early and continue to be part of our journey.
And to everyone who's supported Driven along the way, thank you.
This is just the beginning.
The Most Expensive AI Model Isn’t Always the Best: Introducing DrivenBench youtu.be/gssyvMPfLr8?si=NK1f…
Where Is Institutional Capital Moving? Analyze 13F Data with Driven youtu.be/9mGBrjrO7kI?si=WfU_… via @YouTube
Investment research is becoming something you can install.
In the Agent era, Skills are starting to look less like prompts and more like packaged capabilities.
A good investment Skill can bundle the research framework, rules, examples, data calls, tools, scripts, dependencies, and output structure behind a task.
So instead of figuring out the right prompt, model, workflow, or setup, you just need to know:
What problem does this solve?
What will it produce?
How do I use it?
How are others using it?
That matters even more in investing, where many tasks can’t be solved with one prompt.
Valuation, earnings analysis, portfolio risk, options strategies, IPO research, macro monitoring — these are repeatable investment methods, not one-off questions.
Prompts are instructions. Skills are investment methods you can install, run, improve, and share.
That’s the idea behind the Driven Skill Store.
driven.ai/skills
Yes, we heard you!!
Many users asked for better Hong Kong access, and now it’s here: a smoother Driven experience for Hong Kong IP users and Hong Kong stock investors.
Explore Hong Kong stocks with Driven.
Research with real-time market data, dig deeper into company insights, and let your Agent monitor what matters.
🤩Driven is now also available to users in Hong Kong.
Connect supported Hong Kong brokerages to bring your real holdings and portfolio context into Driven, and trade where supported and authorized.
@futufriends @moomoo @WebullGlobal @RockFlowApp
We saw many users uploading portfolio screenshots to Driven.
Now, no more screenshots.
Connect your account and let your Agent access it directly. @longbridgehk @IBKR @moomoo @futufriends @RockFlowApp
Don't Worry...Only with your permission. Always under your control.
Still uploading portfolio screenshots?
Connect your brokerage account to Driven instead.
Link @longbridgehk , @IBKR , @WebullGlobal , @moomoo , Snowball Securities, @futufriends , or @RockFlowApp via MCP, then ask your Agent to check markets, screen stocks, analyze your real holdings, and trade where supported.
Your brokerage, now inside Driven.
Connect yours at driven.ai
Driven's new era is here!!🫡
Stop uploading portfolio screenshots.
Connect your brokerage account instead.
‼️Driven Connectors are now live.
Connect your broker and let your Agent work with your real holdings, account data, and trades directly.
Already pay for professional financial data?
Connect it to Driven and use it in your research.
Need to turn research into something usable?
Connect productivity tools to create presentations, mind maps, organized knowledge, and more.
16 Connectors are available now, including @IBKR, @WebullGlobal, @moomoo , @FactSet , Fiscal.ai, @NotionHQ , @XmindHQ , and more...
Connect what you use. Extend your Agent.
It’s here.
Yesterday I teased a Driven Skill using $RDDT earnings as the example.
US-Earnings-Insight was built by a Driven user who read 50+ earnings reports and calls this season, then turned his deep-read process into a reusable Skill.
It helps investors look beyond the earnings headline: growth quality, margin signals, management commentary, market reaction, and overlooked details.
Now it’s one of the most-run Skills in Driven Skill Store.
This is what Skills are about:
turning real investment methods into reusable AI capabilities.
A Driven user read 50+ earnings reports and calls this season.
The takeaway: the real insight is often beyond the headline, especially inside the conference call.
So @zway_ai built US Earnings Insight to make that deep-read process repeatable.
The Skill helps investors analyze growth quality, margin signals, management commentary, market reaction, and overlooked details.
Built by a Driven user.
Now one of the most-run Skills in Driven Skill Store.
#EarningsSeason #InvestmentResearch #AIInvesting
$RDDT just gave us the perfect earnings case.
A strong quarter.
A sharp stock reaction.
And a lot hiding beneath the headline.
Who's holding $RDDT now🙋♀️?
Tomorrow I’m sharing a short video on US-Earnings-Insight, one of the most-run Skills in Driven Skill Store.
Built by a Driven user, it turns earnings reports into a buy-side style deep read.
Using $RDDT as the example. You'll find a new direction to see this earnings case.
Stay tuned 👀
I’ve recently started becoming active on LinkedIn again.
Planning to share more thoughts there on AI, investing, AI agents, and how we actually shift our work and thinking for the AI era.
One question I keep thinking about:
AI is clearly useful, but how do we move from using AI as a tool to letting AI agents become part of how we work?
I’ll also keep posting on @ProductHunt and meeting more builders thinking about AI-native workflows.
If you’re interested, connect with me here:
linkedin.com/in/zania-zhang/
So many of you asked for this! It's here. 😍
The Driven Connectors
Connect what you use. Extend your Agent.
Driven Connectors are coming soon 🚀
Connect your broker account &
Let Driven work with your real holdings, portfolio data, and trades.
Connect the financial data you already use.
Turn research into slides, mind maps, and more.
Powered by MCP, Connectors extend what your Agent can access and do beyond Driven’s built-in capabilities.
12 Connectors in the first release.
More soon 👀
If you’re just casually checking something, a general agent may be enough.
But if you take investing seriously, you need fresh market data, reliable sources, and tools built for real research.
‼️That’s why Driven exists:
to help serious investors work with more reliable data, go deeper, move faster, and access capabilities general agents can’t provide.
Chen retweeted
One of the easiest Skills to get started with. Super clear and genuinely useful.👍
I used it yesterday to break down $MU ’s Q3 income statement
share.driven.ai/zubew58avgyb…
Chen retweeted
"The AI Future Is for Everyone." Couldn't agree more with @finkd.
One piece I personally feel strongly about: Frontier AI needs to be for everyone, too. How many people outside tech/eng actually know what to do with the top models today—or even feel the difference when a new one drops?
Closing that gap is where the real future lies.
Give it a read: wsj.com/opinion/the-ai-futur…
2/ Execution raises the trust bar
One thing Product Hunt users made clear: once an AI agent moves closer to order workflows, they do not just want a final answer.
They want to know what data was used, which sources supported the conclusion, what changed, and what they should verify before taking action.
For investment agents, transparency is part of the trust layer.
3/ The messy middle is real
One comment that stood out: investors are already using AI to help with portfolio ideas, but the process can get messy fast.
The hard part is not getting one more AI-generated suggestion.
It is everything after: building conviction, keeping track of the thesis, and knowing when new information should change your view.
That is where investment agents can become more than chat.
🙋♀️Bookmark this if you’re building an AI product in a high-trust category.
I’m bringing a few Product Hunt learnings back to X.
X has a lot of investors.
Product Hunt has a lot of AI builders.
But AI investing products still sit somewhere between the two.
So when Driven launched on @ProductHunt, I was curious:
What would people question first?
Not what sounds exciting.
What needs to be clear before they trust an AI investment agent?
A few themes stood out:
1. Trust comes before automation
2. Source transparency matters more when execution is involved
3. Investors are already using AI, but the process is still messy
4. Market coverage is more than a feature list
5. Scheduled monitoring may be one of the most underrated behaviors
Our takeaway:
AI investing products will not win just by giving better answers.
They need to help users build trust, inspect sources, follow markets, manage portfolios, monitor changes, and move from insight to action without giving up control.
Curious what others think:
What matters most to you in an AI investing agent: trust, data, workflow, or execution?
1/ Trust first, automation later
Most people seem more comfortable starting where the agent reduces workload without taking over the final call.
That is also how we think about Driven.
The goal is not to tell users what to buy.
It is to give individual investors something closer to an investment team:
one that can help research deeper, monitor continuously, follow up on portfolios, surface source-backed context, and prepare next steps, while the final judgment stays with the user.