TidalWave
Agentic AI mortgage point-of-sale platform for US lenders: SOLO guides borrowers through the application around the clock while agents extract loan data, check documents and run both agency underwriting engines at the point of application.
Tidalwave is an agentic AI point-of-sale platform for US residential mortgage lenders and brokers. Its assistant, SOLO, answers borrower and loan officer questions at any hour in more than 30 languages, builds a personalized task list from each borrower's financial situation, and guides the application step by step, while loan officers see what the borrower sees and can step in where help is needed.
Behind the application, AI extracts borrower information from emails, texts and phone recordings to start and update loan files without manual data entry, reviews loan documents for errors and inconsistencies, and analyses bank statements to surface risk indicators and eligible assets. Tidalwave runs Fannie Mae's Desktop Underwriter and Freddie Mac's Loan Product Advisor side by side at the point of application, compares the findings and selects the better outcome for each borrower, then issues verified pre-qualification letters. Income, employment and asset verification run through Plaid, Argyle and Truv.
The platform is delivered as cloud software. Tidalwave describes a model-agnostic framework that combines general-purpose language models with domain-specific ones, and a governance layer with structured input and output protocols, context controls and audit trails. It is built for mortgage origination rather than as a general agent builder, and pricing is quoted through sales.
Vendor details
Canonical URL
https://www.tidalwave.com
Category
Enterprise operations agent
Subcategory
Agentic AI mortgage origination platform
Funding status
Independent, headquartered in New York, founded by chief executive Diane Yu, a two time founder who previously built FreeWheel, sold to Comcast, and served as chief technology officer at Better.com, alongside co founders who were engineers at both companies. Tidalwave raised a twenty two million dollar Series A in November 2025 led by Permanent Capital, with D.R. Horton, the nation's largest homebuilder, joining as both investor and customer, and a follow on from Engineering Capital, bringing total funding to about twenty four million dollars. It was named to the 2026 CB Insights AI 100 and is deployed by lenders including DHI Mortgage, NEXA Lending, Flat Branch, First Colony, and Mortgage Solutions.
Company status
independent
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
Tidalwave runs Fannie Mae Desktop Underwriter and Freddie Mac Loan Product Advisor at the point of application and verifies income, employment and assets through Plaid, Argyle and Truv, writing results back into the loan file.
In practice
A borrower texts pay stubs and a question at 11pm. SOLO answers in their language, the AI pulls the figures into the application, and the loan officer opens a file that is already populated in the morning.
A loan officer wants the best approval path without running each agency engine by hand. Tidalwave submits the file to Desktop Underwriter and Loan Product Advisor together, compares the findings and issues a verified pre-qualification letter.
A processing team keeps catching document problems late in the pipeline. Tidalwave reviews loan documents and bank statements as they arrive, flagging inconsistencies and risk indicators before the file moves on.
Sources & related URLs
Agentic Index coverage score
5.0 / 14 capabilities · 36%
| Integrations & Tool Calling | Full |
|---|---|
|
Loans go to both US agency underwriting engines and to verification services, and the results are written back into the loan file. Tech-specs says Tidalwave runs "Fannie Mae's Desktop Underwriter and Freddie Mac's Loan Product Advisor" at the point of application and "generates DU and/or LPA verified pre-qualification letters instantly", and that "AI updates loan files without manual data entry"; income, employment and asset verification run directly through Plaid, Argyle and Truv. Submitting to both agency engines and returning verified results into the loan file are actions, not just reads. Credential scoping and rotation, and custom tools, are not documented. Connections to loan origination systems and CRMs appear only in third party sources. Sourcetidalwave.com/tech-specs (index copy); July first-party read;read 2026-09-16 |
|
| Workflow Orchestration | Full |
|
Both agency underwriting engines run side by side, with the findings compared and each borrower routed on the better outcome, while agents manage the borrower's tasks. Tech-specs says Tidalwave "runs Fannie Mae's Desktop Underwriter and Freddie Mac's Loan Product Advisor simultaneously at the point of application", "compares findings from both in real time" and "selects the best outcome for each borrower automatically", then generates verified pre-qualification letters; SOLO "generates a personalized task list" from each borrower's financial situation and "autonomously manages tasks"; and how-it-works describes automation "from first contact and document review to submissions and closing". Deterministic underwriting runs combined with agent extraction and task management mix fixed and autonomous steps, and routing on the better of two findings is a branch. Whether lenders can configure the workflow, and how it is versioned, is not described. Sourcetidalwave.com/tech-specs, /how-it-works (index copies); readread 2026-09-16 |
|
| Knowledge Grounding & RAG | Partial |
|
Answers are grounded in each loan's own data and documents. The home page describes SOLO as "mortgage-only AI" answering with "regulatory-compliant responses", grounded in each loan's data, documents and agency guidelines, and tech-specs says "AI extracts all relevant borrower information automatically and updates loan files" and gives "personalized education support" for mortgage questions "tailored to their specific situation". No maintained index of guidelines, overlays or product rules that a lender can update is named. The phrases "mortgage-only" and "domain specific models" describe training, so whether new customer knowledge can be added without retraining is not stated. Sourcetidalwave.com home, /tech-specs (index copies); readread 2026-09-16 |
|
| Human Oversight & Guardrails | Partial |
|
Loan officers can watch what borrowers see and step in, and the agents run inside structured guardrails, but no approval step is named. Tech-specs describes a governance framework with "structured input and output protocols and strict context controls", and says "loan officers see everything borrowers see in real-time, allowing them to jump in and help exactly when and where confusion occurs". Neither statement names an approval or checkpoint before an agent action, or a point where approvals can be inserted. Loan officers keeping control of decisions is a stated intent, not a built gate. Sourcetidalwave.com/tech-specs (index copy); readread 2026-09-16 |
|
| Security, Identity & Governance | Not documented |
|
Security details live in a trust center hosted on Vanta at trust.tidalwave.com (also trust.tidalwave.ai), with sections for frameworks, controls and subprocessors; the page renders client side, and whatever attestations and controls it lists are not stated on any other Tidalwave page. Tidalwave publishes no documentation, help, support, API, developer or status site. SOLO's regulatory compliant responses describe answer quality, not a control or an attestation. Sourcetrust.tidalwave.com; tidalwave.com; triedread 2026-09-16 |
|
| Observability & Auditability | Partial |
|
Audit trails are part of the agent framework, but inspecting individual steps is not described. Tech-specs describes a governance framework with structured input and output protocols, strict context controls "and audit trails". No view of each run's steps and tool calls is described, and SOLO's loan memory is the loan file, not a record of agent behavior. Sourcetidalwave.com/tech-specs (index copy); trust.tidalwave.com; readread 2026-09-16 |
|
| Memory & State Persistence | Partial |
|
SOLO keeps a full record of each loan for the team, but that record is the loan file rather than a separate memory layer. The home page says "SOLO builds comprehensive loan memory from day one, tracking every interaction, document, and decision to provide instant context for any team member". A record of every interaction, document and decision on a loan, read by team members, is the loan file itself, and deleting it would delete the business record. As a conversational assistant, SOLO holds conversation state within a session. No SOLO memory separate from the loan record, with a stated scope and lifetime, is documented, so there is no memory store to review, edit, delete or scope. Sourcetidalwave.com home; read July 2026, refusedread 2026-09-16 |
|
| Deployment & Data Residency | Not documented |
|
Tidalwave runs as SaaS it hosts itself, with no stated region or customer environment option. Its trust center describes TidalWave Tech Inc. as an "AI-driven, cloud native tech company" that "provides software as a service" to US residential mortgage lenders and brokers, and tech-specs describes "containerized microservices with infrastructure as code", which is how Tidalwave builds its own service, not a customer environment option. No data residency statement, region choice or customer hosted deployment is published. The trust center's subprocessor section could name hosting regions. Sourcetrust.tidalwave.com (served metadata); tidalwave.com/tech-specs (index copy); readread 2026-09-16 |
|
| Prebuilt Agents, Templates & Packs | Not documented |
|
Tidalwave sells one point of sale platform, and its capabilities are stages of one origination workflow rather than agents adopted separately. The home page describes one platform "automating document review, data entry, and loan processing with proprietary AI agents", with SOLO answering borrowers, error detection on loan documents and bank statement analysis as capabilities of that platform. They are sold together as a single product, not as units a buyer selects among, and there is no template catalog. Sourcetidalwave.com home (index copy); July first-party read;read 2026-09-16 |
|
| Triggers & Channel Coverage | Partial |
|
Borrowers and loan officers reach the agents by text, email, voice and chat, and every start described comes from a person sending something. How-it-works says borrowers "text or email their information, and our AI instantly extracts loan data", that SOLO answers questions any time in over 30 languages, and that it transcribes phone conversations into started applications; tech-specs says loan officers can "upload emails, texts, or phone recordings to instantly start a loan" and describes "event driven processing enabling real time agent responses". The phrase "event driven processing" does not say what event starts the agent, and no event or schedule that starts agent work without a person asking is named. Sourcetidalwave.com/how-it-works, /tech-specs (index copies); readread 2026-09-16 |
|
| Model Flexibility & Routing | Partial |
|
General purpose LLMs and domain specific models work together, and Tidalwave picks between them. Tech-specs describes a "model agnostic framework combining general purpose LLMs with domain specific models". That is more than one model class inside Tidalwave's framework; no lender or admin choice of model or provider is named. Sourcetidalwave.com/tech-specs (index copy); readread 2026-09-16 |
|
| APIs, SDKs & MCP Extensibility | Not documented |
|
The platform itself has no published API or SDK, and no API or developer site exists. Its connections to mortgage vendors and the agencies show what Tidalwave calls, not whether Tidalwave itself can be called. Sourcetidalwave.com/tech-specs; triedread 2026-09-16 |
|
| Testing, Debugging & Optimization | Not documented |
|
Nothing public shows a way to test or score the agents. Flagging errors in a borrower's loan documents checks the loan, not the agent's output. Press coverage reports a public accuracy benchmark with Columbia University, though no write-up of it from Tidalwave is published, nor is any QA or testing surface. Sourcetidalwave.com; tidalwave.ai/blog; triedread 2026-09-16 |
|
| Browser & Computer Use | Not documented |
|
Every action runs through direct agency and verification connections or through the borrower's own conversational application, not by operating screens. The agents connect to Desktop Underwriter, Loan Product Advisor, Plaid, Argyle and Truv, and tech-specs says "AI extracts all relevant borrower information automatically and updates loan files without manual data entry" from uploaded emails, texts and phone recordings. Since nothing drives a lender's or third party's screens, changes to UI elements have nothing to break. Sourcetidalwave.com/tech-specs (index copy); July first-party read;read 2026-09-16 |
|
The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Recent platform changes
TidalWave launched a bidirectional integration with the Total Expert CRM platform to automatically sync borrower contacts and application statuses. The integration passes real-time application activity from TidalWave into Total Expert as actionable Insights, and allows loan officers to initiate applications directly from their CRM.
Bears on: Integrations
View sourcePricing
Not public; quoted through sales
enterprise subscription scaled to loan volume and seats
What is public
No list pricing. Tidalwave sells its point of sale platform to lenders through enterprise engagement, but rates are not public.
Billing mechanics
No billing unit is published; whether Tidalwave prices per loan, per seat or as a platform subscription is not stated, and rates are quoted through sales.
Cost watchouts
Cost scales with loan volume and seats, and integration into a lender's existing origination stack and agency connections may add implementation effort.
Variable cost rationale
Priced against loan volume and seats, so cost grows with origination throughput and the size of the loan officer team, offset by the platform's stated reduction in cost per loan.
Additional watchouts
Confirm whether pricing is per loan, per seat, or a platform subscription, and how it scales with origination volume across your loan officer team.
Sales call required
Yes, required for paid access
Free / trial
Demo on request; no public free tier found
Key ambiguities
No rate or pricing unit is published on a reachable first-party page, so whether Tidalwave prices per loan, per seat or as a platform subscription is unknown.
Related vendors
- 11th Estate — Agentic platform whose AI engine scans markets, matches an…
- Adopt AI — AI CPA firm whose agents run reconciliations, close, AP and AR, and…
- Aera Technology — Decision intelligence platform where an always on agent executes…
- Akro AI — On premise operational intelligence that automates document heavy…
- alfred_ — AI executive assistant that triages the inbox overnight and scores…
- Alloy.ai — Commerce intelligence system for consumer brands that unifies four…
Alternatives to TidalWave
The closest documented capability profiles to TidalWave among enterprise operations agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- ConverzAI5.0 / 14Adds documented Security, Identity & Governance
- Shortwave6.0 / 14Adds documented APIs, SDKs & MCP Extensibility
- Synthetic3.0 / 14A lighter documented profile than TidalWave
- Vela6.0 / 14Adds documented Security, Identity & Governance
- alfred_5.5 / 14Fuller documented coverage on Knowledge Grounding & RAG and Human Oversight & Guardrails
- Balerion AI4.5 / 14Adds documented Security, Identity & GovernanceTidalWave vs Balerion AI →
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded