Connect your account and get complete Google ads copy from Juma, the Google Ads AI copy software that reads asset performance before writing.
Connect your Google Ads account, name the campaign and ad group, and Juma writes complete Google ads copy from your live asset performance data. No account connected? Share your target keywords and a landing page URL and Juma writes from that brief instead.
Same campaign, different ad group. Juma carries over the client context and pulls fresh asset performance data for the new group, so there is no need to re-brief the campaign goals or brand rules. You get a second RSA set optimized for this ad group's specific ratings and search term data.
Each ad group targets a distinct keyword theme, so the asset performance data will differ from the previous group. Juma reads the new ratings separately, checks the search terms report for this specific group, and writes replacement headlines tuned to those query themes. The retained assets and the new copy reflect the unique performance history of this ad group, not a paste of the previous output.
If the two ad groups have overlapping themes, Juma flags headlines that could serve both sets so you can decide whether to share them or keep the copy differentiated. This step typically takes less time than Step 1 because the client context is already loaded and only the ad group data needs to refresh.
With the responsive search ads complete, add sitelink and callout extensions that reinforce the ad copy. Juma reads the landing page and the existing headlines to write extensions that match the messaging and support click-through rate.
Sitelink extensions add navigational links below the main ad, giving users a direct path to specific pages: a product category, a pricing page, or a sign-up flow. Each sitelink needs a short headline (25 characters) and an optional two-line description (35 characters each). Juma writes the full set based on the destination pages visible on the landing page and the themes already covered in the RSA headlines.
Callout extensions add short benefit phrases that appear in a row beneath the ad text and do not link anywhere. They reinforce the value proposition with specifics: free shipping, 24/7 support, a money-back guarantee. Juma draws these from the landing page content and the approved value propositions in the project, keeping extensions consistent with the ad copy written in Steps 1 and 2.
Once the new assets have gathered enough impressions, pull fresh performance ratings to see how they are tracking. Juma flags any headlines now rated LOW or UNRATED for the next replacement round, keeping ad strength high between refreshes.
Google typically assigns BEST or GOOD ratings within a few weeks of an asset going live, provided the ad group has sufficient traffic. Assets that remain UNRATED after a full data cycle usually signal a targeting or relevance issue rather than a copy quality problem. Juma surfaces those edge cases in the re-audit and notes whether the asset is likely underserving due to pinning, low query match, or a headline too similar to others in the set.
Run this re-audit monthly rather than waiting for a manual review. Ad strength degrades gradually as Google shifts spend toward top performers and underserving assets accumulate. Scheduling the re-audit as a recurring step keeps the Google ads copy current and prevents a slow drop in ad strength from going unnoticed between campaign reviews.
Use the search terms report to identify query themes that current ad groups are not yet targeting. Juma surfaces these themes from the data and suggests new ad group structures, giving you a clear roadmap for expanding the campaign.
Search terms that trigger your ads but do not match any current keyword often reveal how users describe a need in their own words. When a theme appears repeatedly across multiple queries, it signals commercial intent strong enough to support a dedicated ad group with tailored copy and a matching landing page. Juma groups these themes by volume and relevance so you can prioritize which new ad groups to build first.
Each suggested ad group comes with a sample of the search terms that motivated it and a recommended keyword match type. You can take those suggestions directly into Google Ads or use them as the starting brief for the next round of RSA writing. Feeding this output back into Step 1 closes the optimization loop and keeps the account expanding based on proven traffic signals.
Search ad copy gets refreshed every cycle, but the brand rules, approved messaging, and current offers stay relatively stable. A Juma project keeps that foundation in one place so every round of RSAs starts from the client's actual voice and positioning, not from a blank brief. If the client project already exists, add these items to it.
Add a short description to each knowledge item in the project's info field so Juma knows what each file contains and when to use it. For example:
This Flow cuts RSA writing from two to three hours to under thirty minutes per ad group. Instead of manually reviewing asset ratings, checking the search terms report, and drafting new copy from scratch in separate browser tabs, Juma consolidates all three inputs and generates a complete replacement set in one session.
The manual process requires constant context switching: performance data lives in the Assets tab, search terms are in a separate report, and the creative brief is usually in a document or email thread. Consolidating those three sources before writing is what makes the output targeted rather than generic.
For agencies managing several accounts, the time savings compound further. A session that handles one account's asset review, gap identification, and google ads copywriting in under thirty minutes can be repeated across multiple clients in a single working session, without re-briefing from scratch each time.
Juma reads three data sources before writing a single headline: asset performance ratings (BEST, GOOD, LOW, UNRATED), the search terms report for the selected ad group, and your current live copy. Assets rated BEST or GOOD are preserved automatically. LOW and UNRATED assets are flagged for replacement.
The BEST and GOOD thresholds reflect Google's own assessment of how often an asset serves and how well it converts relative to others in the set. Keeping those assets preserves the performance signal your account has already built, rather than discarding earned data to start fresh with a blank brief.
Search terms showing query themes not yet covered by current headlines represent the highest-impact gap. These are real queries your ads already win traffic from, but without a headline that directly addresses the intent. Juma prioritizes these gaps when writing new headline candidates, making the output more targeted than any standard google ads copy generator.
A complete RSA set from this Flow includes up to 15 headlines and 4 descriptions in a downloadable document. Each asset has a character count, role label (benefit, feature, CTA, social proof), and pinning recommendation. The document shows which assets were retained from the existing set and closes with an ad strength assessment.
Character counts matter because Google truncates headlines that exceed 30 characters and descriptions that exceed 90 characters. Having the count visible in the document means you can review the full set before uploading without opening the Google Ads interface to check each asset individually.
The ad strength assessment at the end scores the full set against Google's criteria and flags any remaining gaps. If the set is missing a strong CTA or has too many headlines covering the same theme, the assessment identifies it with a specific recommendation before you upload rather than leaving it to post-upload discovery.
Most Google Ads AI copy software generates headlines from a brief and stops there. Juma reads your live asset performance ratings first, preserving BEST and GOOD assets automatically and replacing only what is underperforming. The output reflects what is already working in your account, which changes both what gets written and what gets kept.
The practical difference shows up in replacement logic. A standard ad copy generator treats every asset as a candidate for replacement regardless of performance. Juma separates the decision: BEST and GOOD assets are preserved, and the creative focus goes entirely toward replacing what is underperforming. This produces a tighter, more efficient asset set rather than a full rewrite that resets the account's performance history.
The search terms layer adds a second differentiator. Generic tools write to the keywords you provide; Juma writes to the gaps your account has already identified through real traffic. That distinction is why the output tends to produce stronger ad strength scores on the first upload.
