Skip to main content
Tomi is Topsort’s AI agent for retail media ad operations. It lets retail admins analyze and manage ad campaigns using natural language.
Tomi is available in the sidebar of the Marketplace Admin view and supports any language.

The Problem Tomi Solves

As retail media programs grow, ad ops teams are expected to support more advertisers, more campaigns, and more reporting requests without proportional headcount growth.
Teams often spend hours pulling reports, comparing performance across time periods, checking pacing, diagnosing ROAS changes, identifying underperforming products or keywords, and preparing vendor-facing explanations. Tomi helps turn that manual investigation work into a conversational workflow.
Tomi can also reduce campaign management overhead by helping users create, update, pause, and resume campaigns through natural language.

What Tomi Does

Tomi turns retail media operations into a conversational workflow. Ask questions in plain language to investigate performance changes, flag anomalies, and surface optimization opportunities across your marketplace data. When write actions are enabled, Tomi can also draft campaign changes.

Analytics

Identify best-performing products, understand category trends, and analyze purchase behavior.

Campaign Management

Create, view, update, pause, and resume Sponsored Listings campaigns through natural language.

Insights

View AI-generated insights that highlight anomalous trends in key marketplace metrics and debug with Tomi. Learn more

Audience Targeting

Target loyal or high-value customer segments, exclude recent purchasers, or boost specific audiences.

Key Benefits

1

Find issues faster

Tomi helps identify which vendors, campaigns, SKUs, keywords, categories, or placements are driving performance changes, so teams can focus on the areas that need attention first.
2

Reduce manual reporting work

Teams can investigate performance, summarize trends, and prepare vendor updates without repeatedly pulling, joining, and reformatting reports.
3

Turn analysis into action

Tomi can suggest next steps such as reviewing pacing, reallocating budget, pausing inefficient spend, checking tracking issues, or preparing a seller-facing explanation.
4

Manage more campaigns with less overhead

Teams can create, update, and optimize campaigns across more vendors through a single conversational interface, without rebuilding from scratch each time.
5

Stay in control

For any write operations, users review and approve before anything goes live. Nothing launches without human sign-off.

Getting Started

Prerequisites

  • Marketplace Admin access
  • At least one active vendor with products in the catalog
  • Tomi enabled for your marketplace (contact your Topsort account team if you don’t see it in your sidebar)

Accessing Tomi

Tomi is available directly within the Marketplace Admin view. Look for the Tomi entry in the sidebar to open the chat interface. No separate login or configuration is required.

Use Cases

Analytics

Diagnose underperforming campaigns

1

Find campaigns with poor performance

Prompt: “Find the campaigns where performance looks bad at the campaign level for vendor X, then drill down into SKU and keyword to identify the likely cause.”Tomi surfaces results across campaigns, SKUs, and keywords.
2

Drill into additional dimensions

Prompt: “Also dig into category and placement level data to see if that’s where the issue is.”Tomi shows additional results, revealing where the underlying issue is.

Investigate what changed for a seller

1

Identify the change

Prompt: “What materially changed for seller X over the last 7 days that explains the change in sales or ROAS?”Tomi compares recent performance against prior periods and highlights the most important changes across many metrics.
2

Drill into the likely drivers

Prompt: “Which campaigns, SKUs, categories, or keywords contributed most to the change?”Tomi breaks down the results to show where the change is concentrated.
3

Summarize the recommended action

Prompt: “Summarize the likely cause and what the ad ops team should do next.”Tomi turns the analysis into a clear root cause hypothesis and a list of recommended actions, such as reviewing pacing, reallocating budget, pausing inefficient spend, checking tracking issues, or preparing a seller-facing explanation.

Creating Campaigns

Create a campaign from top-performing SKUs in a category

1

Ask Tomi for top SKUs

Prompt: “What are the top 10 performing SKUs in the gear category?”Tomi will return a list of products ranked by performance.
2

Request campaign creation

Prompt: “Create a campaign for these products with a $10k budget running for the next two weeks.”
3

Select ad placement

Tomi will ask where the ads should appear: specific keywords, product categories, competitor product pages, or everywhere (always-on).
4

Review and approve

Tomi presents the proposed campaign configuration. Review and confirm to launch.
If the selected SKUs span multiple vendors, Tomi will suggest creating a separate campaign per vendor and confirm before proceeding.

Create a campaign with forecast validation

1

Request a forecast

Prompt: “Forecast performance for a campaign promoting our top home appliance SKUs for [vendor name] with a $10k budget over the next 30 days.”Tomi will return projected metrics including estimated ROAS, impressions, and spend.
2

Conditionally create the campaign

Prompt: “Create the campaign if projected ROAS is above 4.”If the forecast meets the threshold, Tomi will proceed to campaign creation. If not, it will report back with the projected performance and wait for further instruction.

Managing Existing Campaigns

Pause underperforming campaigns

1

Identify underperformers

Prompt: “What are my four most underperforming campaigns?”Tomi will rank campaigns by performance and return the weakest ones.
2

Pause them

Prompt: “Pause those campaigns.”Tomi will confirm before making any changes.