This review is part of a larger series of LinkedIn newsletters titled [The Human Side of AI: Cutting through the AI noise to show you how AI can be a powerful tool for your creativity, efficiency, and strategy].
Soulmates.ai builds synthetic audiences a researcher can question in place of fielding a study. It sells two things. Digital Twins are AI-modeled and category-based, and the vendor calls them directional. Brand Soulmates are 1:1 twins built from real recruited people, profiled on the HEXACO personality model, and validated to a claimed 93% fidelity against holdout respondents. The distinction matters because it decides how much weight a finding can carry. Either way, the tool replaces the sample. It does not replace the analysis.
Key strengths
Twins built on recruited humans
The Brand Soulmates path starts with paid recruitment on the channels an audience uses, screens hard to filter out AI-generated answers and low-effort respondents, then runs a 300-point instrument anchored by HEXACO-60. Twins are held out and tested until they clear the vendor’s fidelity bar. That provenance makes a finding easier to defend under client scrutiny than a persona prompted once from a demographic label. For a researcher who has to show the work, recruit-first is the differentiator.
Fast pre-testing before anything gets fielded
Digital Twins let a researcher stress-test positioning or a concept against a modeled audience in hours instead of briefing a vendor and waiting weeks. That fits early-stage work well: narrowing a field of concepts, or sanity-checking a questionnaire before budget gets committed to a real study. The vendor is candid that this tier is directional by design, which is the right expectation to set.
Confidential by default, and it fits the stack
Data handling is the gating question for anyone uploading client material, and Soulmates.ai answers it plainly: isolated per-client workspaces, with a stated position that the data a client generates is theirs and is never sold or rented. Access is gated by Okta SSO. Queries run through an MCP server, an API, and connectors for Slack, Teams, Salesforce, and HubSpot, so a twin can be reached from inside Claude or an existing workflow instead of a separate portal. A BYOD option enriches twins with a team’s own CRM and past research.
Where it could improve
The whole product leans consumer. Recruitment through paid social to find “your fans,” plus the emphasis on creative testing and social performance, points at B2C brand marketing. For B2B tech buyer research and long enterprise buying committees, the fit is thinner. Paid social is a strange place to go looking for a hard-to-reach IT director.
Then there is the 93% figure. It is the vendor’s own holdout number, and independent research on synthetic audiences and digital twins keeps finding lower accuracy on new, out-of-sample questions. Treat twin output as directional and validate any high-stakes finding against real people. It is also a single-method tool. There is no transcript coding and no synthesis of existing interviews, so it sits beside the qual stack rather than inside it. Pricing is entirely “get in touch,” on annual tiers with project-scoped twin builds, so there is nothing a solo researcher can trial cheaply.
Ratings
| Dimension | Rating | Rationale |
| Usability | 3.5 / 5 | Digital Twins are quick to query once set up and reachable from Slack or Claude, but Brand Soulmates run as scoped engagements, so onboarding is a sales conversation rather than a signup. |
| Power | 4.0 / 5 | On its core task, recruit-first twins with HEXACO profiling and holdout validation are more defensible than pure-synthetic rivals, though out-of-sample reliability caps how far to trust them. |
| Flexibility | 3.0 / 5 | Broad across brand, message, and concept work, even political and M&A, with wide integrations. But it runs a single method and skews consumer, so qual coding and synthesis stay untouched. |
| Cost | 2.5 / 5 | Opaque enterprise pricing on annual commitments, with twin builds scoped as separate projects. Real friction for individual researchers and small teams. |
Conclusion
Soulmates.ai is one of the more grounded entries in a crowded synthetic-audience field. The recruit-real-people-first design earns it a look for message and concept testing, especially when speed matters. It is built for consumer brand teams, so B2B buyer research will find the fit narrower. Treat the output as a starting point. Use the twins to narrow options and pre-test, then let real human data and a researcher’s judgment carry the decisions that count.
Last updated: 8/6/2026.