Recruiting software pivots from fairness to fraud

Fraud became the shared brief

The recruiting software message hardened over the fortnight. On 18 Sep, Greenhouse added a blog post on detecting candidate fraud and product pages on global hiring and better hiring experiences. On 21 Sep, it added pages on AI candidate screening and hiring fraud, plus product pages for voice AI integrations with 60 ATS platforms. On 25 Sep, it added more content on AI skills for recruiters and recruiting software for small businesses.

Lever moved in the same direction on 18 Sep, with new posts on AI in candidate and resume screening and a product page for a cost of a bad hire calculator. SmartRecruiters followed on 25 Sep with articles on modern recruiting strategies and AI-generated resume fraud, after launching its September 2026 product release page on 20 Sep and event pages for its Intelligent Hiring Lab series.

The visible positioning has changed. The public surface is less about workflow elegance and more about trust in the funnel: fraudulent resumes, AI screening, candidate engagement and the cost of hiring mistakes. Greenhouse also removed a blog page on reducing hiring bias with structured hiring on 18 Sep, a notable juxtaposition in a week when fraud and AI screening content accelerated.

Greenhouse had the loudest cadence, with reliability in view

Greenhouse was the most visibly active vendor. On 21 Sep, it added PetVet Care Centers as a customer logo and published a case study saying PetVet reduced managers’ time spent recruiting by 67%. On 25 Sep, it added Bath Fitter as a customer logo and published a case study saying Bath Fitter reduced time-to-fill by 36%. The same day brought three new blog posts, one new customer story page and eight new product pages, including events and newsroom announcements.

The company also showed hiring intent in technical areas. On 21 Sep, Greenhouse listed four new engineering roles focused on machine learning in Ontario and British Columbia. On 25 Sep, it listed three further roles across locations including Northern Ireland and London.

That release pace sat alongside two status-page events. On 23 Sep, Greenhouse reported three incidents: a minor partial service disruption affecting Indeed applications lasting about 166 minutes, a minor degradation in job board update performance lasting about 1,087 minutes, and a major outage in which MyGreenhouse job search was unavailable. On 27 Sep, it reported a minor Report Connector incident, resolved after about 52 minutes. In a category now selling confidence in the hiring pipeline, reliability disclosures become part of the product narrative, not a side note.

Jobvite rebuilt its catalogue in four days

Jobvite’s website made the sharpest structural move. On 15 Sep, sitemap changes showed the removal of 99 provider-related product pages and one careers page. On 19 Sep, its website featured 99 new provider-related product pages and a new careers page. On 26 Sep, Jobvite launched a new product page for CiChecked.

That four-day gap matters. This was not a slow brand refresh spotted after the fact. It was a removal, a reappearance and then a named product page arriving within eleven days. For operators, the signal is not only the new page. It is the sequence: catalogue pruning, catalogue restoration, then a fresh offer surfacing.

Jobvite also kept the paid channel warm. On 20 Sep, 35 new Google ad creatives were observed for Jobvite, with 40 creatives running in total. On 27 Sep, another six new creatives were observed, again with 40 creatives running in total.

Culture Amp pushed vertical proof and engineering hygiene

Culture Amp’s fortnight looked like a company broadening from engagement into sector-specific employee operations. On 18 Sep, it published content on AI in professional services, high-performance culture, technical challenges faced by employees and offboarding, and introduced a product page for its AI at Work Benchmark 2026. On 21 Sep, it added five blog pages across topics including aged care in Australia and New Zealand, private equity and gender in professional services, while removing two pages related to aged care and Permira. On 26 Sep, it added pages on continuous learning, professional services and on-demand webinars about employee engagement.

The engineering trail was unusually visible. On 15 Sep, Culture Amp released kaizen-design-system version 3.3.12, fixing usability issues in SingleSelect and MultiSelect components inside overlays, and created two repositories, example-go-buildkite-plugin and kafka-connect-plugins. On 21 Sep, it released @kaizen/tailwind 1.5.20 with dependency patches and created protect-event-parameters. On 26 Sep, it released version 3.0.0 of protect-event-parameters, moving the GitHub action to Node 24 from Node 20 while leaving parameters and masking behaviour unchanged.

Hiring supported that product motion. On 24 Sep, Culture Amp had listed 11 new roles over the preceding seven days, including sales, engineering and marketing roles. Its paid activity was also active: 29 new Google ad creatives were observed on 22 Sep and five more on 26 Sep, with 40 creatives running in total on both dates.

The ad shelf was full across the category

Paid distribution was not quiet. Greenhouse had 15 new Google ad creatives observed on 18 Sep and 12 more on 26 Sep, with 40 creatives running in total on both observations. Lever had nine new creatives on 18 Sep, six on 22 Sep and three on 26 Sep, also with 40 running in total each time.

SmartRecruiters showed 23 new creatives on 19 Sep, nine on 23 Sep and five on 27 Sep, with 40 running in total on each date. Jobvite and Culture Amp showed the same 40-creative running total on their observed ad checks. The pattern is simple: every named vendor in the set was observed at that 40-creative mark at least once during the period.

The sharper read is simultaneity. Between 18 Sep and 27 Sep, the category refreshed content on AI, fraud, vertical use cases and product releases while keeping paid creative inventory visibly stocked. This is a market fighting on message repetition as much as feature depth.