When Bolbi Liu says she’s building an agentic ad agency, she means it literally.
AdsGency, the San Francisco startup automating the entire performance marketing workflow — from creative strategy to placement to optimization — has raised $12 million in seed funding led by XYZ Venture Capital (with Streamlined Ventures, HF0, and Hat-Trick Capital joining in).
It’s not another “AI ad generator.” It’s an autonomous system that plans, creates, launches, tracks, and scales campaigns — all in hours instead of weeks.
The Why: Advertising Is a Mess
Today, running ads looks like juggling a dozen disconnected tools: Google Ads, Meta, TikTok, data dashboards, attribution platforms, CRM syncs, and human teams trying to hold the chaos together.
AdsGency’s premise is simple: kill the fragmentation.
A single LLM-powered platform connects every ad channel, learns what works, and self-optimizes.
Average customer ROAS? 8x.
That’s not hype. Brands like TAL Education and Mobvoi have seen 3–10x improvements in ad performance after switching from traditional agencies.
Why Investors Pounced
A founder who’s built ad systems at scale. Bolbi Liu designed multimillion-dollar internal ad platforms at AWS and DiDi.
A $400B+ market begging for automation. Traditional agencies can’t keep up with data velocity or ad fatigue.
Insane early traction. 650% revenue lifts for clients, global adoption via word-of-mouth, and early enterprise uptake.
Camille Ricketts (Partner at XYZ VC, ex-Notion CMO) calls AdsGency “an emerging giant — Cursor for advertising.”
That’s high praise. And dead-on positioning.
The Hypothetical 10-Slide Deck That Might’ve Landed $12M
(Even though, knowing Bolbi, she probably skipped slides and just let the demo do the talking.)
Slide 1: Hook — “Advertising, Rebuilt for the Agentic Era”
Content:
Headline: “Not another ad tool. A self-running ad agency.”
Subhead: LLM agents that plan, execute, and optimize campaigns across every channel — automatically.
Investor Lens: Sets ambition from slide one — new category, not incremental improvement.
My 2 Cents:
Most founders say “we use AI.” Bolbi says “we replaced humans.” It’s a stronger hook — and 10x riskier. That’s why it works.
Slide 2: The Problem — “Ad Chaos Is the Norm”
Content:
12+ disconnected tools per campaign.
Manual reporting, missed attribution, channel silos.
Teams overworked and insights underused.
Investor Lens: Chaos + cost = trillion-dollar inefficiency.
My 2 Cents:
Every investor has seen this pain up close. The magic isn’t showing chaos — it’s showing clarity after it.
Slide 3: Market Context — “$400B Waiting for Automation”
Content:
Global digital ad market = $400B+.
90% still managed manually or via slow agencies.
Fragmented workflows = 40% wasted spend.
Investor Lens: Huge TAM + inevitability of automation.
My 2 Cents:
If the pitch is “we make humans faster,” it’s meh. But if it’s “we make humans optional,” that’s seed-round bait.
Slide 4: The Product — “One Platform. Infinite Campaigns.”
Content:
Unified LLM stack: strategy → creative → execution → optimization.
Integrations: Google, TikTok, Reddit, Pinterest, Salesforce, Snowflake.
Output: 8x average ROAS.
Investor Lens: Full workflow automation = defensible moat.
My 2 Cents:
Don’t just show screenshots. Show the timeline shrinking — from 2 weeks to 2 hours. That’s what sells time.
Slide 5: The Tech — “Agentic Orchestration at Work”
Content:
Custom models trained on ad performance + creative data.
Proprietary agent coordination layer (multi-LLM stack).
Zero-party data system for privacy-safe attribution.
Investor Lens: Data moat + technical credibility.
My 2 Cents:
Every AI startup says “fine-tuned.” Few can say “self-improving.” That’s the line that lands the check.
Slide 6: Proof — “The Results Speak Ad.”
Content:
TAL Education → 3x ROAS.
Mobvoi → 10x profit increase.
Halliday → $3.3M revenue, 13x return.
Investor Lens: Hard ROI data = instant trust.
My 2 Cents:
If your metrics make VCs’ eyebrows go up, stop talking. Let silence do the work.
Slide 7: The Competition — “Agencies Are Dead Software Companies”
Content:
Matrix showing:
Agencies = slow + human-heavy.
Tools = fragmented + reactive.
AdsGency = unified + autonomous.
Investor Lens: Clean framing → “category creation” vibes.
My 2 Cents:
Never trash competitors. Just show a world where they feel obsolete.
Slide 8: The Team — “Built by Ad Ops Veterans Turned Engineers”
Content:
Bolbi Liu: Ex-DiDi, AWS.
Core team: researchers + growth marketers.
Advisors: ex-Google Ads, ex-Notion, data engineers.
Investor Lens: Deep domain + technical execution.
My 2 Cents:
Founder-market fit isn’t a buzzword when the founder built the systems the incumbents still use.
Slide 9: The Vision — “From Ad Automation to Ad Autonomy”
Content:
Phase 1: Ad creation + placement.
Phase 2: Budget management + attribution.
Phase 3: Autonomous campaign learning.
Investor Lens: The roadmap writes its own value climb.
My 2 Cents:
The phrase “ad autonomy” is sticky. It’s narrative shorthand for a trillion-dollar shift.
Slide 10: The Ask — “$12M to Build the First Agentic Ad Agency”
Content:
Raise: $12M seed.
Use: Engineering hires, model scaling, market expansion.
Goal: Replace agencies entirely within 5 years.
Investor Lens: Clear capital logic. Big vision. Measurable milestones.
My 2 Cents:
Seed rounds aren’t for building companies — they’re for building inevitability.
Why It Works
AdsGency isn’t just automating ad buying. It’s replacing the agency model entirely.
It owns the story because:
It feels inevitable.
It shows numbers, not adjectives.
And it’s run by someone who built this movie before — just inside Amazon.
That’s the rare founder–market–moment trifecta.













