Brief

While AI agents can build and ship code at unprecedented speed, they lack the strategic context and product intuition to make meaningful decisions. Brief provides the missing "intention layer" - giving AI agents business context, user understanding, and institutional memory to make intelligent product trade-offs.

CEO and Founder

Mocksi

Mocksi is revolutionizing the way software companies sell, learn, and evolve their products through AI-powered simulations. Our platform empowers businesses to create dynamic, interactive demonstrations that showcase their software's full potential, gather valuable market insights, and continuously adapt to meet evolving customer needs.

CEO and Co-founder

ProductBridge

Principal

Burb

Founder

Skedulo

Advisor

Skedulo is the platform for intelligent mobile workforce management. Our web and mobile apps help enterprises intelligently schedule, dispatch, and track jobs in the field with real-time visibility and communication, creating a more productive mobile operation.

Chief Product Officer (Product, Engineering, and Design)

Fond, formerly AnyPerk

Fond helps organizations across industries strengthen their company cultures through an ongoing commitment to employee happiness. Our employee engagement products are intuitively designed and flexible so companies can tailor rewards, recognition, perks and sentiment measurement programs for maximum impact.

VP of Product, Engineering, and Business Development

Microsoft

Lead half of the PM organization within Yammer, including four product managers and three business development managers.

My team focused on everything from growth to mobile app development, web features, and API/platform development. My platform BD team evangelized the Yammer API to partners and trained them on Yammer's model of growth, retention, and engagement.

Principal Program Manager Lead

Yammer, Inc.

Conceived and executed many successful launches (press below):
• Ticker
• Universal Search
• Yammer Connect & Embed
• Salesforce
• SharePoint
• Many partner integrations (https://www.yammer.com/applications), SAP, Dynamics

Director of Product Management

SafeNet

•	Continuing work with the DataSecure Product Line
•	Assisting in development of the corporate Enterprise Data Protection strategy
•	Expanding knowledge of the Product Lifecycle Process, as it applies to midsized multi-nationals

Senior Product Manager, Data Protection

Ingrian Networks

•	Learned Product Lifecycle Management: from strategic vision, near-term milestones, to end-of-life
•	Maintained constant high-level customer interaction to drive requirements, strategic and tactical
•	Worked cross-functionally with Marketing, Engineering, Sales, and Support
•	Released two well-adopted new products and a series of upgrades

Product Manager

•	Performed all functions, pre and post-sales, relating to the deployment of the DataSecure product line
•	Supported global Enterprise and Government sales
•	Worked closely with upper management and technical decision makers of Fortune 500s
•	Studied the pain points of compliance/security in large enterprise

Systems/Sales Engineer

An experienced operator, with a background as an individual contributor in Design, Engineering, Sales, and Product and leadership experience in Product, Engineering, Data, Design, and Business Development.

Been holding off on a full Brief demo because we've been shipping so fast the UI changes every few days 👀 

But we finally hit a more backend-focused moment, so here's 10 weeks of Brief (a16z speedrun sr005).

The short version: Brief is product intelligence that moves at AI velocity.

Product managers currently keep a dozen systems in their head. Analytics, Slack, GitHub, Notion, Fireflies, Linear. It takes days to synthesize all that into strategy. Brief does it in seconds.

What you'll see in the video:
Strategy Agent - generates your full company context (customers, business model, competitive position, goals, dev process)

Work Agent - Shows your entire pipeline from idea to shipped feature

Persona Agent - Understands who your users are from analytics and call notes

CPO Agent - Monitors for new patterns, invalidated assumptions, and market opportunities

All of this feeds into Decisions. The atomic units we pass to coding agents so they build the right thing the first time.

Through MCPs, Brief orchestrates everything: writes PRDs, creates tickets, and gives your coding agents the context they need.

Brief is for teams that don't have time for quarterly planning cycles. Connect corporate strategy directly to your coding agents and ship 10x faster.

Andrew Dillon

About 100 years ago, a group of product leaders got together out of frustration that most product events targeted entry level PMs.

It was an incredible group, including product leaders like Jackie, Vrushali, Gadi, Louisa, Neil, Christina, Nat, Gemmy (we need to meet up!).

It was through PM Leads that I met Ely.

Even before he was an educator and startup advisor to the stars, I benefitted from Ely's pragmatic approach to product management.

Fast forward our careers through jobs, teaching, advisory, fractional work, Ely's still a great sounding board for product strategy. Just as smart about the operation of modern product development with agents as he was with user intuition growing Yelp and Chime.

Sounds like his dance card is pretty full, but if you get a chance, you should definitely try and get Ely as an advisor for your startup!

Wanna know the secret of 10x engineering performance?

....ready?

....really realllllllly wanna know?

Here it is: give them enough context to make the right decisions and then get the f- out of their way.

For every feature, engineers make dozens, hundreds of decisions that aren't in mocks, PRDs, or prototypes. 

With the right context, they make the right decisions. Without it, they're left to guess or have to interrupt their flow to ask.

This is why it's such a massive unlock for engineers to take part in user research. They will never ever forget what a user said in a research session.

Now extrapolate that to coding agents that jam out thousands of lines of code in minutes. There millions of decisions happening daily that aren't in cursrorrules, claude.md, or agents.md files.

Without context, they're made incorrectly. The result is either rework or delivering the wrong thing for the user and the company.

You're hamstringing your efforts at modernization and pouring token money down the drain if you aren't giving your coding agents product intelligence at the speed they operate.

Had coffee with a friend yesterday who showed me an AI app they're trying out. On one page, they had a "product manager agent" as one option in a list of 20+ virtual coworkers.

Every Brief (a16z speedrun sr005) demo, people see we can write documents, create tickets, write code, etc. and want us to branch out into other domains.

The obvious question: why build specialized agents when one can do so much?

Depth and context.

𝗗𝗲𝗽𝘁𝗵
Foundational models are better accountants than me. Better lawyers than me. But this doesn't make them particularly good accountants or lawyers.

When you talk to your lawyer about a case, there's back-and-forth. They pull from past cases that seem unrelated to yours. They have ways of bucketing legal concepts that shape how they see your situation. They apply precedent through frameworks built from years of practice.

You're not just getting information plugged into a workflow. You're getting a multivariate understanding of how to think through a problem. That's depth.

Ways of working through problems that come from deep expertise, not just being really smart about something.

𝗖𝗼𝗻𝘁𝗲𝘅𝘁
Context means more than a big context window. Real context is both volume *and* attention.

Critical product decisions necessitate an understanding of customer needs, user personas, 6-month goals, company mission, values, shipping velocity.

You can dump all that into a million-token window. But the right approach gives an agent tools to understand the challenge plus a goal-seeking drive derived from deep experience.

Specialized agent swarms work because each sub-agent provides individual context. The orchestrator pulls from these layers to make better decisions.

---

General-purpose systems can be extremely powerful. Much like LLMs themselves, society hasn't fully priced in the value of really easy workflows like we haven't priced in the value of really good autocomplete.

But there's a gulf between virtual coworker platforms and delivering critical expertise in a deep domain.

Straight cookin' w/Opus 4.5 (and fewer calls due to the holiday week 😅).

This was last night, on another one today...

"What should we build next?"

You know the answer. It's somewhere in Linear. And GitHub. And that call transcript from last week. And the decision you made in Slack two months ago about not building that feature.

So you tab over to Linear to see what the team's working on. Check GitHub to see what actually shipped. Skim through Granola transcripts to remember what users said. Try to recall which features you decided to prioritize and why.

You synthesize all of this in your head. Then you go back to the agent and give it an answer.

You just became a human API between your agent and reality.

That's what we're fixing. Brief pulls your product reality from where it already lives - Linear, GitHub, your research, your codebase - and makes it available to your agents through MCP.

Velocity tracking shows you what your team is actually shipping and how fast.
User intelligence tells you who you're building for based on what's in your features, marketing, and codebase.

Feature pipeline connects what you're researching, planning, building, and shipping in one view.

The agents get context. You stop being the bottleneck.

briefHQ dot ai - DM me to skip the line.

Founders don't just wake up one morning and say, "My product strategy sucks." 

Because a lack of product leadership actually looks like:
- Slow delivery times
- Features not coming out as expected
- Untamable bugs and decreasing quality
- Pissed off customers

And basically being stuck in minutia, when you're trying to deliver that big vision you promised investors.

Your AI keeps pitching features you shipped last quarter.

We built something to fix that.

AI tools for product management have zero memory of what actually exists in your codebase. They'll read a year-old spec and confidently suggest you build it.

Never mind that it's been live for months.

Brief (a16z speedrun sr005)'s new agent scans your existing code, catalogs what's already shipped, and gives AI the context it's been missing. So when you're planning your next build, you're not drowning in redundant suggestions.

Video shows how it works. (There might be a hint or two about what's coming next.)

s/o to Josh Sirota and Eragon for the fit.

Ripped from the headlines... of Stephan and I asking Brief (a16z speedrun sr005) how we should position Brief.

What do you think "product sense" too squishy?

View company: Brief (a16z speedrun sr005)

What is “Product Sense"?

Everyone talks about it. Few can define it.
I'm asking because we're positioning Brief as "Product Sense at AI Velocity”… but I'm not sure that's right.

Brief gives AI coding agents (Cursor, Claude, etc.) access to your business context: who your customers are, what you're building, why decisions were made. The stuff that keeps agents from shipping technically perfect features that miss the mark.

But "product sense at AI velocity"... does that land?

So tell me: What does product sense actually mean to you?

(Asking for a friend… That friend is me.)

The "AI only gets you to 80%" complaint is backwards.

I was 𝘵𝘩𝘢𝘵 engineer early in my career. The one who'd build something 80% of the way, then abandon it for the next shiny thing. I was never cut out to be an enterprise developer.

When I started coding again in January, something clicked. Engineering became fun again because the AI actually finished the last 20%.

The linting. The testing. The pixel pushing. All gone. No more diving 20 layers deep in div tags, I just told the agent what the user needed, and it handled the grunt work.

Building Brief (a16z speedrun sr005) from idea to 6 figures happened fast because we're shipping exactly what Brief needs to be.

So what's really happening with all these "80% solutions" flooding the market?

Building something 𝘪𝘯𝘵𝘦𝘳𝘦𝘴𝘵𝘪𝘯𝘨 is easier than ever. Way more products exist now than could have existed before. But understanding what users actually need and executing on that vision? That part hasn't gotten any easier.

The "flash-in-the-pan" demos people complain about never would've been built in the first place. Or they'd have been built so slowly that nobody would've noticed. Now they're everywhere because the barrier to entry dropped.

That's not a bug. That's more experimentation, more ideas, more shots on goal. The ecosystem benefits from abundance, even if 90% of attempts fail.

The hard part was never the coding. It was always the product sense, the iteration, the understanding of what people need and building it incrementally for a niche before expanding.

If you're one of those builders using AI to ship faster, I'd genuinely love to show you how Brief helps nail that last 20%.

When you become a manager for the first time, it feels like doing open-heart surgery with oven mitts on.

You've spent years building muscle memory for the actual work. Now you're guiding someone else through it. Imprecise, frustrating, but ultimately more impactful because you're multiplying output across a team.

Managing managers is like adding dishwashing gloves inside those oven mitts. Another layer of abstraction from the hands-on work. A whole new communication skill set.

The meme is that AI has turned us all into managers. We prompt agents, they execute, we guide them to do better. Partially true.

There's a fundamental difference in AI coding though, the communication chain is broken.

In traditional product development, PMs communicate seek out context and communicate it around the organization. User needs, tradeoffs, why we're building something. Engineers use that context to build.

With AI coding agents, the PM can't talk directly to the agent. The engineer is the manager and the sole point of communication, translating product context into prompts. The agent can't ask clarifying questions about users or goals. It just executes.

So agents constantly drift off course: building for the wrong user, ignoring long-held decisions. Engineers get frustrated by rework. The product suffers.

The missing piece isn't better code generation. It's giving these agents the same context that makes all engineers effective. User needs, product tradeoffs, the why behind decisions.

We built Brief (a16z speedrun sr005) around this moment.

When Stephan, said "It's like you're in my development environment!" I knew we were onto something.

The agent finally had the context it needed. Not just to write code faster, but write the right code.

Product management has always been a "why" job, where "how" was a necessary byproduct.

How became so ossified that you could fill arenas with PMs who only ever operated process and never really engaged with the why of their work.

Coding agents have detonated process and a lot of folks are scratching their heads as to what Product looks like in the future because of it. But process is just how and how's never been constant.

Why is more important than ever.

The future of software development are engineering, product management, and design agents coordinating work, driven by small teams with taste and intuition.

It's pretty easy to extrapolate where this goes with coding agents. Come get a demo of Brief (a16z speedrun sr005) if you want to see "the future of product management."*

I posed this question to Stephan, "What does a design agent look like?"

The industry response has been design systems and prototyping. Which was where my head was when I asked Stephan, as he'd led design system efforts at Autodesk.

But his answer pointed back to something much more elemental. Color.

Non-designers don't think about color a lot. Throw in a hex code or RGB and move on. But hex codes are calculations, they don't render the same everywhere.

This has a compounding effect when your hex-based design system meets tailwind tokenization and then meets an application framework. The same color might be converted 3-4 times from your designs in Figma.

Compound this with coding agents' propensity to just make up a new shade or inject inline literals and your color palette asplode.

So Stephan's answer to "what does a design agent look like?"

Was that it would start with something that helps him work across application and cross context and standardize seemingly simple concepts like color.

We've got our plates full, but excited to see and partner with folks building this!

--

* actual quote from a new angel investor working at a foundational model lab

Used to be people who tried to connect with me without an intro message were trying to sell me something.

Now they're all trying to get into speedrun.

taplio

Andrew Dillon's Linkedin Analytics

Andrew Dillon

Performance ScoreAndrew Dillon's LinkedIn Performance

Andrew Dillon's Best Posts (last 30 days)

Ready to grow your LinkedIn presence?

taplio

.css-6ytdur{background-image:linear-gradient(284deg, rgba(39, 43, 72, 1) 0%, rgba(33, 47, 183, 1) 100%);color:transparent;-webkit-background-clip:text;background-clip:text;}Andrew Dillon's Linkedin Analytics

Andrew Dillon

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Andrew Dillon's LinkedIn Performance

Andrew Dillon's Best Posts .css-1emoo84{font-size:var(--chakra-fontSizes-md);font-weight:400;color:var(--chakra-colors-gray-600);}(last 30 days)

Ready to grow your LinkedIn presence?

Andrew Dillon's Linkedin Analytics

Performance ScoreAndrew Dillon's LinkedIn Performance

Andrew Dillon's Best Posts (last 30 days)