We tried to replicate PreGTM's workflow in Apollo
PreGTM performs two core tasks. It finds candidate companies and evaluates their websites to verify genuine account fit.
Our first article compared Apollo's API against live Google search queries.
In this test, we evaluated two additional discovery workflows inside Apollo: manual Boolean UI search and the AI list assistant. We passed all generated lists through the same website crawler and classifier.
Target profile: EU importers and wholesale distributors of commercial batteries above 2 kWh.
Test brief used
Eu importers / wholesalers of baterries above 2kw that will be subject to the upcoming digital passport requirements from the new EU legislation
Evaluation results
| Discovery method | Attempted | Strong fit | Total fit (strong + possible) | Site failed to load |
|---|---|---|---|---|
| Search (pooled baseline) | 600 | 4.8% | 20.3% | 13.8% |
| Apollo API | 300 | 4.7% | 15.3% | 25.3% |
| Apollo UI (Boolean AND) | 258 | 1.6% | 8.9% | 9.3% |
| Apollo AI Assistant | 146 | 4.1% | 6.2% | 16.4% |
Key findings
1. Search outperformed Apollo on complex specifications
After crawling and reading company websites, the Apollo UI list yielded 8.9% qualified accounts. The AI assistant list yielded 6.2%.
Live search queries produced a 20.3% qualified account yield on the same profile (p < 0.0001).
High-level database industry tags like "Wholesale" or "Electrical Manufacturing" cannot capture specific capacity thresholds (e.g. ">2kWh") or trade compliance terms.
2. Apollo UI searches produced very healthy domains
The manual Boolean UI list achieved the lowest failure rate in the benchmark: only 9.3% of sites failed to load.
This outperformed the 13.8% failure rate of live search queries.
3. Qualification requires deep site reading
Both platforms require reading the live website to verify account eligibility. Keyword matching creates a candidate list; deep page text confirms the business model.
Two approaches to website qualification
| Dimension | Apollo Qualify / AI Research | PreGTM Classify |
|---|---|---|
| Mechanism | In-app prompt updating a record field | Multi-page site crawl and quote validation |
| Pages analyzed | Apollo record data and common site paths | Full sitemap and catalog pages within crawl budget |
| Output | Yes/No field or picklist value | Fit rank tier with verbatim quoted evidence |
| Workflow role | Qualify directly inside the outreach CRM | Export verified account CSV before enrichment |
How to combine both platforms
- Keep Apollo as your sales workstation: Manage contacts, email sequences, CRM synchronization, and pipeline inside Apollo.
- Use PreGTM as an upstream account filter: When targeting technical or niche B2B roles, filter raw domain lists through PreGTM before spending enrichment credits in Apollo or Clay.
Test constraints
One campaign profile, one classifier, 404 total domains evaluated in this test. The classifier did not read PDF documents.
Annex: Seed parameters
Apollo Boolean search filter (Germany)
Exclude keywords:
software, saas, consulting, consultancy, law firm, market research,
news, publishing, trade association, certification, testing laboratory,
trade fair, recycling
Employees: 1-1000
Germany query:
("Batterien" OR "Akkus" OR "Lithium-Ionen-Akkus" OR "Traktionsbatterien"
OR "Industriebatterien" OR "Batteriespeicher" OR "Stromspeicher"
OR "Energiespeichersysteme")
AND
("Großhandel" OR "Importeur" OR "Distributor" OR "Fachhandel"
OR "Vertrieb" OR "Handelsvertretung" OR "Großhändler" OR "Lieferant"
OR "B2B-Handel" OR "Einfuhr")Apollo AI assistant filter parameters
Location: 27 EU member states
Industry: Wholesale + Electrical/Electronic Manufacturing
Employees: 10-200
Keywords (any): energy storage battery, battery distribution,
battery importer, battery import, energy storage, battery supply
Excluded: retail