Walmart expands gen AI ad tools with Sparky and Marty
Retail media is shifting again, and Walmart is moving fast. Sponsored placements are entering Sparky, the in-app AI shopping assistant. On the brand side, Marty is rolling out to help advertisers plan, optimize, and report without living in spreadsheets.
If you're a creative, this matters. Your work now shows up inside conversations, not just carousels and banners.
- Ads are coming to Sparky, Walmart's AI shopping assistant
- Marty automates search ad recommendations for Walmart advertisers
- New AI creative tools aim to cut production time
- What marketers should know
Ads are coming to Sparky, Walmart's AI shopping assistant
Sparky answers product questions, summarizes reviews, and recommends items from Walmart's catalog. Now, sponsored prompts appear inside that flow. In plain terms: your ad copy becomes part of the assistant's reply.
Walmart says 81% of Sparky users check availability or research before buying. Search-style questions are the main use case, so sponsored prompts land at the exact moment of intent. Amazon did something similar with its AI assistant, Rufus, earlier in 2024, signaling where retail discovery is headed. See Amazon's Rufus overview.
How to write for Sparky placements
- Answer the question first, then suggest. Lead with a useful, one-sentence takeaway before any CTA.
- Mirror shopper language. If the query says "budget blender for smoothies," use that wording in your first line.
- Prioritize specifics. Feature, use case, outcome. Example pattern: "900W motor for thick smoothies, dishwasher-safe, under $100."
- Be transparent. Keep ad labeling clear and avoid hype. The assistant format punishes fluff.
- Use variations. Prepare multiple micro-answers for common missions (gift, budget, premium, fast shipping, small space).
Marty automates search ad recommendations for Walmart advertisers
Marty is a gen AI "super agent" for Walmart Connect. It analyzes performance, recommends bids and keywords, and ships quick reports. It also gives share of voice so teams can track competitive position.
Walmart reports that 97% of Marty queries are unique, which means advertisers are asking very specific, real-world questions. That's good news for lean teams that need focused guidance without an analytics pit stop.
Creative uses for Marty (beyond bids)
- Extract language that converts. Ask for top query patterns that led to purchases and fold those phrases into headlines and image text.
- Find feature gaps. Identify attributes shoppers ask about but your product pages don't clearly show (e.g., sizing, materials, bundle contents).
- Prioritize visuals. Use Marty to tell you which benefits drive CTR by category and map each to a specific image or background concept.
- Ship faster reports. Automate daily creative KPI snapshots to decide which variants to refresh, pause, or scale.
Prompt starters you can copy
- "List the top 10 shopper queries that converted for [brand/category] last week and the exact wording used."
- "Which attributes show up in high-CTR queries that our product titles/descriptions don't highlight yet?"
- "Create three short headline options for Sponsored Search that align with the highest-converting query themes."
- "Summarize our share of voice shifts by keyword group and suggest one creative angle per group."
New AI creative tools aim to cut production time
Walmart's Automated Creative Generator builds backgrounds and visual assets for product imagery. Walmart says it can reduce production time by up to 80%. That's huge if you run scrappy launches or juggle seasonal refreshes.
A fast workflow for your team
- Define the job: one benefit per visual. Don't cram five messages into a 1:1 square.
- Pick a mood board: three style anchors max (material, lighting, palette).
- Generate 10-20 variants, then shortlist 3 based on clarity at thumbnail size.
- Add text only if it survives small-screen tests. Keep it to 3-5 words.
- Run a quick accessibility pass: contrast, legible fonts, clean edges.
With connected TV now in the mix via Vizio, Walmart's pitch spans from assistant-led discovery to upper-funnel video. Expect more briefs that ask for cross-surface consistency with lightweight, channel-specific tweaks.
What marketers should know
1) Media is moving from placements to agents
AI agents turn product discovery into a conversation. Your copy needs to be helpful first, promotional second. Think utility content with a single, confident nudge.
2) AI support levels the field
Marty brings better recommendations to smaller teams, which means more competition in mid-funnel formats. The edge goes to teams who iterate creative weekly, not quarterly.
3) Performance and brand budgets are blending
As creative tools plug into optimization, lines blur. Plan measurement across formats with one scoreboard: assistant-sourced conversions, assisted view-through, and share of voice by theme.
Action checklist for creatives
- Draft five Sparky-friendly micro-answers per hero SKU (mission-based: budget, premium, gift, fast, small-space).
- Pull weekly "query-to-copy" insights from Marty and update headlines accordingly.
- Build a background library (light, dark, lifestyle, ingredient) you can reuse and adapt in minutes.
- Create a voice guardrail doc for assistant replies: tone, do/don't phrases, claim limits.
- Set an experiment cadence: two new variants live every week, retire the bottom 20% performers.
Metrics to watch
- Assistant-sourced conversions and CTR on sponsored prompts
- Share of voice by priority query groups
- Variant survival rate after 7 days (keep, iterate, kill)
- Production time per asset before vs. after AI assist
Bottom line: Treat AI assistants like new storefronts. Write answers that help first, design visuals that read at a glance, and use Marty's insights to keep the feedback loop tight. Ship small, ship weekly, and let the data steer the next draft.
Want structured, role-specific practice with AI for creative work? Explore courses by job at Complete AI Training.
Your membership also unlocks:
