Resources / System Agent

System Agent Prompt Catalog

120 ready-to-use prompts for the Synerise System Agent, organized by what the agent can do and who uses it.

A practical reference for marketers, analysts, merchandisers and developers. Find a prompt, replace the placeholders, then use it in your workspace. This page does not run prompts or connect to your data.

Browse the catalog

Private preview

The agent is in private preview on selected workspaces; open it from Synerise Agent in the top navigation bar. Capabilities follow the Synerise System Agent documentation.

Read the documentation(opens in a new tab)

A map of the workspace

Start with the job.

Every prompt has a stable ID, such as REC-04, and a two-level home: family → area. Four tags cut across the catalog: intent, persona, output and level.

01

DiscoveryWhat do people see?

Recommendations (REC) · Search (SRC) · Experiments (A/B/X) (EXP)

44
02

DataWhat does the platform know?

Product feed (FEED) · Catalogs (CAT)

18
03

InsightWhat is happening?

Analyses (ANL)

20
04

BuildWhat happens next?

Automation (AUT) · IQL (IQL) · Jinjava (JIN)

30
05

LearnHow does it work?

Documentation (DOC)

8
Understand the tags and capabilities

Intent

Explore
List objects, inspect configuration, check status
Measure
Get performance numbers, trends and comparisons
Preview
See results before they reach production
Build
Draft a campaign, workflow, filter, analysis or template
Tune
Fix, troubleshoot or improve something that exists
Learn
Get an explanation or an answer from the documentation

Persona

Marketer
Campaigns, journeys, communication, loyalty
Analyst
Data, reporting, experiment results
Merchandiser
What users see in search and recommendations, across app, web and stores
Developer
Feeds, catalogs, filters, templates, API

Output

Table
Metrics, campaign lists, analytics results, catalog records, search results
Chart
Line, column, bar, area or pie chart for trends and comparisons
Carousel
Product cards with image, title and details such as price
Action list
Rows of objects (campaigns, catalogs) with next actions
Guided form
Structured input for setup and multi-step configuration
Answer
A written explanation, recommendation or documentation answer
Draft
A proposed workflow, rule or configuration to review and apply
Code
An IQL filter or a Jinjava snippet
CSV
An exported file

Level

Starter
One turn, works as written once placeholders are filled in
Advanced
More parameters, comparisons or follow-up turns

Areas and capabilities

REC — Recommendations
Inspect campaigns and the AI recommendation engine, measure performance, preview results before they go live.
SRC — Search
Inspect indexes, synonyms and query rules, run test searches, read search analytics, tune relevance.
EXP — Experiments (A/B/X)
Review recommendation and search tests, prepare new configurations, interpret variant results.
FEED — Product feed
Check item feed imports and look up items in item catalogs.
CAT — Catalogs
Inspect key-value data catalogs, their records, jobs and event enrichment; export data.
ANL — Analyses
Ask about platform data in plain language: reports, metrics, trends, funnels, segmentations, histograms, aggregates, expressions, sankeys.
AUT — Automation
Describe a use case and get a workflow diagram to review and apply in Automation Hub.
IQL — IQL
Build item filters from plain language; validate, fix and explain existing ones.
JIN — Jinjava
Generate, explain, validate and fix Jinjava for message templates and recommendation inserts.
DOC — Documentation
Get answers sourced from the Synerise User Guide, API Reference and use case library.

Before you paste

Writing prompts that work

01

Name the object.

Use the campaign, index, catalog, product, rule or experiment name or ID. If several objects match, the agent shows the options and asks you to choose.

02

Give a time range for performance questions.

"Last 7 days", "this month" or exact dates. Without one, the agent picks a default and tells you which.

03

Go step by step on complex tasks.

Refine as you go: "use the Personalized type instead", "make it 12 items", "show that as a chart".

04

Review before you confirm.

The agent asks for confirmation before sensitive changes. Where it can't apply a change itself (for example some synonym or query rule changes), it gives you the configuration to apply manually.

Text in [square brackets] is a placeholder. Replace it with a name, ID, value or pasted code from your workspace.

The reference library

Find your next prompt.

120 prompts / 5 families / 10 areas

Replace highlighted [square brackets] with values from your workspace.

Results: 120 / 120

  1. REC-01

    DiscoveryRecommendations

    List all active recommendation campaigns.

    Intent
    Explore
    Persona
    Merchandiser
    Output
    Action list
    Level
    Starter

  2. REC-02

    DiscoveryRecommendations

    Show the configuration of campaign [campaign name]: recommendation type, feed, filters and number of items.

    Intent
    Explore
    Persona
    Merchandiser
    Output
    Table
    Level
    Starter

  3. REC-03

    DiscoveryRecommendations

    Which recommendation types are available for feed [feed name]: personalized, similar items, complementary, frequently bought together, visual similarity?

    Intent
    Explore
    Persona
    Merchandiser
    Output
    Table
    Level
    Starter

  4. REC-04

    DiscoveryRecommendations

    How is campaign [campaign name] performing this month?

    Intent
    Measure
    Persona
    Marketer
    Output
    Table
    Level
    Starter

  5. REC-05

    DiscoveryRecommendations

    Show clicks, CTR, revenue and conversions for all recommendation campaigns in the last 30 days, sorted by revenue.

    Intent
    Measure
    Persona
    Analyst
    Output
    Table
    Level
    Starter

  6. REC-06

    DiscoveryRecommendations

    Plot daily CTR for campaign [campaign name] over the last 8 weeks.

    Intent
    Measure
    Persona
    Analyst
    Output
    Chart
    Level
    Advanced

  7. REC-07

    DiscoveryRecommendations

    Compare revenue from [app home screen campaign] and [web homepage campaign], last month vs. the month before.

    Intent
    Measure
    Persona
    Marketer
    Output
    Chart
    Level
    Advanced

  8. REC-08

    DiscoveryRecommendations

    Which products generate the most revenue in campaign [campaign name] this quarter?

    Intent
    Measure
    Persona
    Merchandiser
    Output
    Table
    Level
    Starter

  9. REC-09

    DiscoveryRecommendations

    Which recommendation campaigns had CTR below 1% in the last 14 days?

    Intent
    Measure
    Persona
    Merchandiser
    Output
    Action list
    Level
    Advanced

  10. REC-10

    DiscoveryRecommendations

    Preview what campaign [campaign name] would recommend for product [product ID].

    Intent
    Preview
    Persona
    Merchandiser
    Output
    Carousel
    Level
    Starter

  11. REC-11

    DiscoveryRecommendations

    Show personalized recommendations for customer [UUID]. I want to see what they get on the app home screen.

    Intent
    Preview
    Persona
    Marketer
    Output
    Carousel
    Level
    Starter

  12. REC-12

    DiscoveryRecommendations

    Preview what campaign [Articles for you] returns for reader [UUID].

    Intent
    Preview
    Persona
    Marketer
    Output
    Carousel
    Level
    Starter

Showing: 12 / 120

One conversation, several steps

Connect the prompts.

The agent keeps context within a conversation, so prompts from the catalog chain into longer tasks. Run them in order in one conversation. Select an ID below to reveal its full prompt; this clears any conflicting catalog filters.

  1. 01

    Fix searches that return nothing

    Cut the zero-result rate in one index

  2. 02

    Launch recommendations on a new app screen

    Go from feed check to a running test

  3. 03

    Diagnose an underperforming campaign

    Find out why CTR dropped and what to change

  4. 04

    Win back lapsed customers

    From audience to a personalized message

  5. 05

    A product never gets recommended

    Trace the problem from catalog data to filters

Catalog at a glance

Different roles. One workspace.

Intent

Explore
28
Measure
29
Preview
8
Build
31
Tune
12
Learn
12

Persona

Marketer
35
Analyst
23
Merchandiser
44
Developer
18

Output

Table
32
Chart
13
Carousel
11
Action list
5
Guided form
2
Answer
27
Draft
13
Code
16
CSV
1

Level

Starter
66
Advanced
54

Counts are calculated from the prompts in this catalog.