Top AI builders for HR & People
Builders never pay to be listed, and order never depends on fees. What counts
On this page: Rankings · New this month · By model
Rankings
Top builders
- 1ModernEdge AITop use case: Decisions made on partial data because systems don't connect2solutions · 2 complete
- 2SectionTop use case: Training rebuilt for every role or location2solutions
- 3Bonsai LabsTop use case: Hiring throughput capped by screening and scheduling2solutions
- 4The AI LabTop use case: Quality depends on which site or which person2solutions
- 5B43Top use case: Training rebuilt for every role or location1solution · 1 complete
- 6AtherialTop use case: Senior time spent ramping juniors1solution · 1 complete
- 7OutcomeCatalystTop use case: Senior time spent ramping juniors1solution · 1 complete
- 8Portfolio Leverage CompanyTop use case: A specialist answering the same questions repeatedly1solution
- 9SudolabsTop use case: Senior time spent ramping juniors1solution
- 10Beyond the PromptTop use case: AI tools in place but unconfigured for the work1solution
Ranked by solutions on file. Ties go to the builder with more complete receipts: ones that say what changed in numbers, how long it took and what it cost. Once enough receipts are complete, this board will rank by those instead.
Buyers' picks
- 1
- 2
- 3How a Facilities Firm Cut RFP Turnaround 50-70%ModernEdge AI
- 4
- 5
- 6
- 7
- 8
- 9
- 10How One PE Operator Aced a Termination Talk in 30 MinBeyond the Prompt
Likes from buyers count first. Until more solutions have likes, the rest of the list is the receipts that tell a buyer the most: outcomes, a number, a time and a cost.
New this month
By model
Pick a model to see who builds on it and what for.
18 solutions use several models; 18 run on custom or proprietary models. A solution counts once for each family it names. GPT includes OpenAI, ChatGPT and Azure OpenAI; Gemini includes NotebookLM; Open-weight is Llama, Mistral and other self-hosted models.
Most solutions on any model
- 1ModernEdge AITop use case: Repeat customer questions that eat support capacity15solutions
- 2CorralDataTop use case: Marketing spend optimized on the wrong signal10solutions
- 3ElizaTop use case: Engineering throughput capped by hands-on-keyboard hours9solutions
- 4OdeTop use case: Intake documents retyped into a system9solutions
- 5PressWTop use case: Policies exist but nobody can find them mid-task9solutions
- 6Lazer TechnologiesTop use case: Validation and prototyping that takes months before a real user sees it8solutions
- 7Bolt GroupTop use case: Staff or capacity scheduled without visibility into volume6solutions
- 8North GroupTop use case: AI tools in place but unconfigured for the work6solutions
What it's used for
- 116solutions
- 215solutions
- 312solutions
- 411solutions
- 510solutions
- 610solutions
Share of new solutions, by quarter
| Quarter | GPT | Claude | Gemini | Other | Open-weight | Solutions |
|---|---|---|---|---|---|---|
| 2026 Q2 | 61% | 40% | 12% | 7% | 1% | 92 |
| 2026 Q3 | 43% | 44% | 12% | 9% | 5% | 104 |
| 2026 Q4 | 25% | 75% | 17% | 21% | 8% | 24 |
Of the solutions that say what they run on, the share naming each family. By when it went live where the builder told us, otherwise when it was added here. A solution naming two families counts in both, so a row can add up to more than 100%.