Understand its recommendations
Hero AI's responses look like polished prose. Underneath, they're built from data lookups and reasoning steps. Knowing which parts come from where helps you decide what to act on.
Where the data comes from
Hero AI pulls from three sources:
- Live Google Ads data. Performance metrics, keyword lists, ad copy, and extensions for your active account. Pulled fresh when you ask. This is the most reliable layer.
- Your product context. The product description and analysis from onboarding. Reflects what you wrote at setup, edited if you've updated it.
- General knowledge about Google Ads and B2B SaaS. Built into the model. Useful for context, but not specific to your account.
When Hero AI cites a number, that number came from layer 1. When it explains why something might be happening, that's reasoning over layers 1 and 3. When it references "your customers" or "your product", that's layer 2.
What's grounded versus inferred
Recommendations have two parts: observation and inference.
- Observations are facts pulled from your account. "Your CPA on the alternative to jira campaign is $245 over the last 14 days" is an observation. You can verify it in the dashboard.
- Inferences are conclusions Hero AI draws from the observations. "The high CPA is likely driven by a few high cost keywords that aren't converting" is an inference. It might be right, might not.
Trust observations. Verify inferences before acting on them, especially when they recommend a change.
How findings are weighted
Hero AI doesn't treat every blip as equally meaningful. Findings are sorted into three buckets so you can tell signal from noise:
- Patterns. Consistent across the current period, the prior period, and a longer historical baseline. These are usually structural and worth acting on.
- Variances. Single-period anomalies that may revert without any intervention. Treated as worth watching, not necessarily worth fixing.
- Inconclusive. Worth a closer look but not enough data to commit. Often surface as "investigate this" rather than as a recommendation.
Recommendations are weighted toward patterns. The supporting evidence section under each recommendation shows the timeframe data so you can verify the bucket yourself.
When a recent change in your account looks like it caused a metric shift, change-history events are scored as a primary suspect, secondary, or ruled out. This makes it easy to see what the engine is attributing the move to.
How to verify before acting
Three ways to sanity check:
- Cross check the dashboard. Hero AI's numbers should match what the dashboard shows for the same date range. Mismatches are usually a date range or filter issue.
- Open Google Ads. For per keyword recommendations, the search terms report and per keyword performance view are the source of truth. Hero AI summarizes; Google Ads shows the underlying data.
- Ask Hero AI to show its work. "What specific keywords are you talking about? Show me the numbers." It'll list the keywords and metrics behind the recommendation.
When recommendations are reliable
Hero AI gives more reliable recommendations when:
- The campaign has at least 100 clicks and 10 conversions of data.
- The time window is at least 14 days.
- You've stated a goal so it knows what to optimize for.
It gives less reliable recommendations when:
- The campaign is less than a week old.
- Volume is low (under 50 clicks per week).
- Conversions are low or zero (it'll often say "not enough data").
When recommendations are wrong
Hero AI is wrong sometimes. Common ways:
- It misses context outside Google Ads. A spike in CPA might be driven by an industry shift or a competitor's launch that Hero AI can't see.
- It overweights recent data. A weekend dip might look like a trend if Hero AI only checked the last 5 days.
- It can be confidently wrong. The polished prose can make a guess sound like a finding. Treat specific numbers as facts, but treat causal claims as hypotheses to test.
When Hero AI is wrong, push back. "Are you sure that's what happened? What's the underlying data?" usually surfaces the gap.
What to do with recommendations
Most recommendations come with one of two paths:
- AI-assisted. Some changes have an action button on the recommendation: pause a campaign, add a negative keyword, draft and apply new ad copy, redesign a landing page. You'll review and confirm before any change is made. See Actions Hero AI can take.
- Manual. For changes that need a human in Google Ads (restructuring targeting, editing conversion tracking, anything Hero AI can't safely automate), the recommendation gives step-by-step instructions you follow yourself.
A few rules of thumb regardless of path:
- For changes you're not sure about: test on one campaign first, give it a week, then evaluate.
- For big strategic recommendations (kill a campaign, restructure targeting, double a budget): get a second opinion. Ask Hero AI to argue the opposite case ("What's the case for not doing this?") or run it past someone on your team before clicking the action button.