Haast AI compliance software reviews marketing content before it goes live and monitors it after release. It can support a brokerage’s advertising review when its documented scope fits your workflow. It doesn’t audit mortgage files or make an ad legally compliant. A named person in your brokerage still approves every ad, and the value of Haast depends on the approval record you build around it.
Assess Haast for Advertising Review
Haast describes itself as compliance AI for enterprise marketing and legal teams in regulated industries. Its financial services page names insurers, banks, fintechs and funds as customers. Its rule library covers ASIC and the ACCC. Haast’s pages describe no features specific to credit, lending or mortgage broking.
Haast fits a brokerage when advertising volume is high enough that a structured review queue saves your reviewers time. That usually means a brokerage with several brokers, a marketing person or partner content to watch. It is a weaker fit for a solo broker who posts a few times a month.
Haast has three documented functions, and each answers a different question:
| Function | What Haast does | What it doesn’t do |
|---|---|---|
| Pre-publication review | Checks a draft ad against configured rules and suggests fixes before release | Approve the ad or confirm the claim is true |
| Live monitoring | Watches published assets and channels, including partner and influencer content | Decide what your brokerage does about a flagged post |
| Regulatory horizon scanning | Scans regulatory updates daily across the US, EU and APAC, then links changes to existing reviews and policies | Decide how a change applies to your brokerage |
None of these functions audits a loan file. Haast doesn’t check best interests duty (BID) notes, serviceability or credit assistance records. It doesn’t give an assurance that an ad complies with the law either.
Check What the Product Actually Reviews
The Haast platform page lists PDFs, Word documents, images, video and audio as review inputs. Haast’s homepage adds PowerPoint, Excel and email files such as EML and MSG. Drafts can come through Word, Google Docs, Figma, Notion, Workfront, monday.com, a Chrome extension or its API.
Haast’s legal engineers build the rules with your team from the applicable regulations, your internal playbooks and your risk appetite. Rule sets and risk thresholds can differ by product line, region, team or campaign. Each compliance task has its own agent with its own risk tolerance. Lawyers’ feedback inside Haast refines the rules over time.
Each flag shows the rule applied, what triggered it and why, with a suggested fix. Content owners fix the flags and submit the asset. Reviewers then approve it, add guidance or give feedback on Haast’s analysis. Haast’s homepage lists single sign-on and granular admin controls, but its platform page doesn’t describe individual permission roles.
For monitoring, the financial services page names social, web and partner or reseller sites. Haast also publishes results such as an 80% reduction in manual review. Treat those as vendor claims, not outcomes measured for a brokerage like yours. Your pilot produces the only figures that apply to you.
Two gaps matter for brokers:
- Content outside the supported formats or connected channels never reaches review. A claim a broker makes on a phone call, or a post in a channel nobody connected, goes unchecked.
- Context-dependent claims need evidence Haast doesn’t see. “We cut a client’s rate by 0.7%” is accurate only if the borrower file shows it. Haast reads the ad, not the file behind it.
Build a Broker Advertising Approval Record
An approval record links each ad to its source evidence, its review and the person who signed it off. The example below is illustrative: a fictional case-study post moving through a review workflow.
The draft post reads: “We refinanced the Taylors from 6.79% to 6.09% and cut their repayments by $310 a month.”
- Source evidence. The broker links the claim to the borrower file: the old and new loan documents showing each rate and repayment.
- Permitted use. The Taylors give written consent to use their story and their surname, or the post drops the name. RG 234 says testimonials from individuals should be attributed to them and be authentic (RG 234.130).
- Review. The content owner submits the draft. A review against ASIC’s rules would flag the interest rate without a comparison rate. The National Credit Code requires a comparison rate whenever an ad states an interest rate (RG 234.56).
- Correction. The content owner adds the comparison rate and its warning, and qualifies the saving as this client’s result. Each change creates a new version with a named owner.
- Sign-off. A named reviewer checks the final version against the file, the consent and the placement, then approves it.
- Retention. The approved version, its evidence and the approval decision go into the record your licensee or aggregator can inspect.
Keep these items for every approved ad:
| Record item | Who owns it | What it proves |
|---|---|---|
| Final approved version | Content owner | Exactly what was published |
| Source evidence reference | Broker on the file | The claim matches the client’s loan |
| Client consent | Broker on the file | The client agreed to the testimonial or case study |
| Review flags and responses | Content owner | Each issue raised and how it was resolved |
| Approval decision | Named reviewer | Who approved the ad and which version |
| Placement record | Marketing owner | Where the ad ran and how it appeared |
A Haast flag can’t replace the reviewer’s check. Haast compares the words with your rules, but it can’t confirm that 6.09% was the Taylors’ actual rate or that their consent exists. Nor can it see whether a story crop hides the comparison rate, which ASIC’s advertising guide says must be no less prominent than the interest rate (RG 234.58). ASIC places primary responsibility for an ad on the organisation placing it.
Monitoring adds one more duty. Someone else might post a testimonial on your page. If you leave it up once you know about it, ASIC treats you as responsible for it from then on (RG 234.187).
Evaluate Data Access and Operating Fit
Haast receives the ad drafts and everything inside them, plus document content from any integration you connect. Keep borrower details out of drafts unless the client has consented to their use. Work from anonymised examples wherever the ad allows it.
Haast’s privacy policy says data is encrypted in transit and at rest, and that Haast holds ISO 27001 and SOC 2 Type II certifications. It says data from connected Google accounts is never used to train models beyond delivering the features you request. The policy doesn’t say whether other customer content trains its models, or where data is stored.
Haast keeps personal information as long as necessary for the policy’s stated purposes, then deletes or anonymises it, though backup archives may keep it longer. Its published terms are a usage policy and don’t address content ownership, retention after you leave or export. Settle these points in the customer agreement before you send any content:
- Whether your drafts and approvals train Haast’s models, and where they are stored and processed.
- How long Haast keeps content after the contract ends, and how deletion is confirmed.
- Whether you can export approval records, flags and audit logs in a usable format.
- Which permission roles exist, so a content owner can’t approve their own ad.
- Who updates your rule set when ASIC changes its guidance: Haast’s legal engineers or your compliance reviewer.
Haast publishes no prices, so cost arrives as a quote after a product tour. The scope you bring to that conversation drives it: the teams and channels to cover, the rule sets to build and the integrations to connect. Haast’s FAQ puts implementation at 8 to 12 weeks, with its legal engineers translating your policies into rules. Budget your own compliance reviewer’s time for that work too.
Judge the fit against your advertising volume. An enterprise review process suits a brokerage or licensee approving many ads across brokers and partners. A small brokerage might get the same protection from a written checklist and its licensee’s approval process. The page on AI regulation in Australia covers your wider obligations when a brokerage uses AI.
Define a Meaningful Pilot and Its Limits
Run a pilot on approved, non-sensitive sample ads before you commit, and decide in advance what each test must show.
- Approved sample ads: send ads your licensee has already approved. Count the flags Haast raises on content you know is acceptable.
- Known problematic claims: include an interest rate with no comparison rate, “pre-approved” or “guaranteed acceptance” wording (RG 234.141), and “instant approval” or “no credit checks” claims (RG 234.140). Add a testimonial with no attribution and a savings headline whose qualification sits in small print.
- Expected review evidence: for each problem ad, write down the flag and rule you expect before the test. Compare Haast’s flag and reasoning with that answer.
- Escalation paths: name who decides when Haast and your reviewer disagree, and when an ad goes to your licensee’s compliance team.
- Approval bypass: try to publish a rejected ad, and change an approved ad after sign-off. Each must return to review before it can go live.
Adopt Haast when the pilot shows three things. Its flags catch your known problems without burying approved ads in false flags. Every approval records the named reviewer, the version and the evidence in a record you can export. Your team can still review and approve ads by hand when Haast is unavailable or a flag is unclear.
If any of those fails, keep the manual process and treat Haast as optional. For other ways brokers use AI, compare the AI tools for mortgage brokers.
