# Filled sales battlecard example — Browse AI / Amazon S3


Our [Browse AI record](/companies/browse/) highlighted data-destination support. We followed that signal to Browse AI’s current [Amazon S3 page](https://www.browse.ai/integrations/amazon-s3) and [setup guide](https://help.browse.ai/en/articles/10715592-aws-s3-integration-guide). They provide a much stronger basis for a battlecard than a feature checkmark.

Browse AI documents CSV and JSON exports, S3 availability on all plans at no additional integration cost and regular robot-run credit usage. Its setup guide requires an AWS account with permission to create CloudFormation stacks and IAM roles, plus the destination bucket. Those are concrete facts to work with.

The following is a **worked card for a hypothetical competing data-extraction vendor**. The buyer scenario and proposed sales wording are illustrative. We have not tested either product in a customer environment and are not asserting that the hypothetical seller has an advantage.

### Buyer and fit

**Situation:** an operations lead wants recurring web data delivered into the company’s own S3 bucket. An AWS administrator controls permissions. The buyer is evaluating whether to use Browse AI or another extraction tool.

**Browse AI’s credible strength:** it already documents the required export destination and file formats. A buyer who wants data in S3 has a reason to shortlist it. Do not dismiss that strength as “just an integration.”

**Qualification:** find out which part of the workflow is difficult. Extracting the right fields, approving AWS access, delivering a predictable file or recovering from a failed run are separate jobs. Compete on the one the buyer actually needs help with.

**Poor-fit condition for a replacement:** if Browse AI already meets the extraction, access and downstream-processing requirements at acceptable cost, an integration checkbox is a weak reason to switch. Establish a real unsolved problem before proposing a migration.

### Discovery that leads somewhere

| Ask | Follow up | What the answer changes |
| --- | --- | --- |
| “What will consume the exported file?” | “Does it expect CSV or JSON, specific field names or a particular folder structure?” | Defines the sample output both vendors must produce. |
| “Who can approve the AWS connection?” | “Can they create the required IAM role and CloudFormation stack?” | Identifies an implementation dependency and the person to involve. |
| “How current must the data be?” | “What happens downstream if an export is late or duplicated?” | Sets an acceptance test for delivery and failure handling. |
| “Who owns this after setup?” | “Who gets notified and restores the process if it breaks?” | Opens a meaningful discussion about ongoing work. |

These questions do not imply that Browse AI fails any of the requirements. They establish what you need to verify before comparing the offers.

### Claim record and objection response

**S3-01 — usable, vendor-documented, checked September 10, 2026:** Browse AI documents an S3 integration with CSV and JSON exports. Its FAQ says the integration is included across its plans; normal robot-run credits still apply. Its guide describes the AWS permissions required for setup.

**Withdraw this hypothetical line:** “Browse AI cannot export to your own cloud storage.”

**Also avoid:** “S3 export is free, so there is no usage cost.” An integration fee and the cost of running the extraction are different things.

**Buyer objection:** “Browse AI already sends data to S3.”

**Proposed response:** “Yes, it documents that integration. Let’s compare the output your downstream process needs and who will maintain it. If both tools meet those requirements, the decision should come down to the remaining extraction, operating and commercial differences.”

That response acknowledges the buyer’s correct information. It also gives the rep a useful next step without inventing a feature gap.

### A fair proof exercise

Agree the requirements with the buyer before the demo. Use the same source page and required fields for both vendors, with authorization to extract the data and a nonproduction destination.

1. **Produce the sample.** Export the agreed fields in the buyer’s chosen format. Check completeness, field names and how missing values appear.
2. **Inspect delivery.** Confirm the file reaches the intended bucket and prefix. Check that the downstream process can consume the actual output.
3. **Review access.** Have the AWS owner inspect the permissions and setup steps. Record the work required from the buyer, rather than guessing from “no code” marketing.
4. **Test repeat runs.** With a controlled test source, change a value and rerun. Record how updates, file naming and duplicates behave.
5. **Review recovery and cost.** Ask how failures are surfaced and recovered, then capture actual usage for the test. Identify any support or recurring operating work that remains.

The pass criteria are the buyer’s required output, permitted access and agreed delivery behavior. We are proposing this test, not reporting a result. Put the actual evidence into the card when it exists.

**Next step:** the deal owner sends the sample specification; the buyer’s AWS administrator confirms access requirements; the product specialist records the two outputs and unresolved questions. Agree a date with those people rather than adding an arbitrary deadline to the card.


## Evidence and ownership

Claim ID: S3-01. Evidence level: vendor-documented. Verification date: September 10, 2026. Status: usable for the documented existence, formats and stated plan eligibility only; comparative performance remains untested.

Sources:
- https://www.browse.ai/integrations/amazon-s3
- https://help.browse.ai/en/articles/10715592-aws-s3-integration-guide

Card owner: assign a named owner before operational use. Buyer, seller, deadlines and evaluation results are not real deal records. Agree them with the actual participants. Recheck documentation and relevant plan before a live comparison. Withdrawn example wording is illustrative, not a claim that a real vendor used it.

Guide: https://competitortracker.io/blog/battlecards-go-stale-dossiers-do-not/