How the catalogue is organised

Uptiq's agents are not a single product with modules. Each one is scoped to a specific workflow, which is deliberate: narrow scope is what keeps permissions defensible, makes configuration to institutional policy possible, and allows a deployment to be measured against a before-and-after number rather than a general impression.

That produces a catalogue rather than a platform diagram. The agents group into four families, and most institutions start inside one of them rather than across several.

Two things run underneath all of it. Document AI is the extraction and classification layer nearly every other agent depends on, and Integrations Cloud is the connection layer, more than 100 native integrations into core banking, origination, servicing, CRM and document management systems. Neither is usually bought on its own; they are what make the workflow agents deployable without a systems replacement.

The four families

Grouped by workflow, with the named agents in each. Coverage varies by lending product, so the lending families are listed by the product line they were built for.

FamilyAgentsWhat they do
IntakeIntake Agent for Commercial Lending, CRE Lending, SMB Lending, SBA Lending, Equipment Finance, and Consumer LendingClassifies incoming documents, checks the package against a configurable checklist for that product, and surfaces what is missing while the applicant is still engaged
UnderwritingUnderwriting Agent for Commercial Lending, CRE Lenders, Small Business Lenders, SBA Lenders, and Equipment Finance CompaniesApplies your credit policy consistently across files, calculates ratios your way, and flags exceptions before committee rather than during it
Analysis and preparationFinancial Spreading Agent, Bank Statement Analysis, Cash Flow Agent, Credit Memo Generation AgentThe document-to-decision layer: tax returns and financials spread into your template, statements analysed, memos drafted with citations back to source pages
Continuous monitoringContinuous Monitoring Agent for Commercial Lending, CRE Lending, SBA Lending, Equipment Finance, and SMB LendingThe post-close tickler lifecycle: chasing required financials, testing covenants as documents arrive, surfacing breaches within a day of receipt
Deposits and onboardingBusiness Deposit Account Opening, Consumer Deposit Account Opening Agent, Client Onboarding AgentIdentity and entity document handling, verification evidence, and the completeness work that decides whether an application is finished or abandoned
Risk, compliance and governanceEighteen agents including Sanctions OFAC Screening, KYC CDD EDD Ongoing, KYC and KYB Compliance Monitoring, Fraud Detection and Investigation, Credit Risk Monitoring and Surveillance, Loan Review and Credit QA, Model Risk Management, Regulatory Reporting, Regulatory Change Management, Consumer Compliance Fair Lending UDAAP, Internal Audit, Exam and Audit Management, Policy and Procedure Management, Third-Party Vendor Risk Management, Information Security and Cyber Risk, Enterprise and Operational Risk, Business Continuity and Resiliency, and Complaint Analytics and SurveillanceThe largest single family, aimed at the second and third lines rather than the front office
Wealth and advisory operationsClient Engagement Agent, Client Lending Agent, Trust and Estate Document Review Agent, Alternative Investments Operations Agent, Billing Reconciliation and Fee Validation Agent, Compliance and NIGO Review Agent, M&A Advisor Transition Agent, M&A Customer and Account Data Analysis Agent, and Proposal Generation AgentBuilt for wealth managers and RIAs, where the document volume sits in onboarding, statements and quarter-end paperwork rather than credit files
Growth and analyticsAbandoned Application Follow-Up and Outreach System, SMB Dashboard AnalyticsRecovering applications that stalled, and reporting across the SMB book

The pattern worth noticing is that the lending families are organised by product line rather than by function. An Intake Agent for SBA Lending and one for Equipment Finance do the same category of work against very different document sets and checklists, which is what makes each of them useful rather than generic.

For how the lending agents fit together across a single file, see AI agents for commercial lending workflows, and for what governing any of them requires, AI agents for financial services.

HOW THE LENDING AGENTS FOLLOW ONE FILEINTAKEClassify, check the list,chase what is missingANALYSISSpread, bank statements,cash flow, memo draftTHE DECISIONApprove, decline, structure.A person, every time.MONITORINGCovenants tested, financialschased, breaches surfacedEACH STAGE HANDS STRUCTURED OUTPUT TO THE NEXTWhich is why the second agent costs less to deploy than the first.Separate agents exist for commercial, CRE, SMB, SBA, equipment finance and consumer lending, becausethe document sets and checklists differ enough that a generic version would help nobody.
Agents on either side of the decision. Never in place of it.

Four things true of every agent in the list

These are platform-level properties rather than per-agent features, and they are the reason the catalogue is structured the way it is.

01

Scoped to one workflow

An agent that spreads financials does not hold credentials to change a loan status. Narrow scope is what keeps permissions justifiable to a reviewer and limits what can go wrong, and it is why the list has fifty entries rather than five.

02

Layered over your existing systems

Agents read from and write to the core, origination, servicing, CRM and document systems already running, through 100+ native integrations. A deployment is a workflow change rather than a platform migration, which is what makes the timelines possible.

03

Configured to your policy and templates

Your spread template, your ratio definitions, your checklist for that product, your memo structure. Output that does not match how your institution works gets routed around, whatever its accuracy.

04

Evidenced by default

Every extracted value cites back to its source page, adjustments are surfaced for a person to accept or reject rather than applied silently, and every override is retained with its reason and user. This is what makes review a check rather than a redo.

Deployment timelines follow from the second point. A single agent is typically live in about five business days and a multi-agent suite in around thirty, provided the scope stays one defined workflow and the configuration uses templates you already have.

Where institutions usually start

The catalogue is large; the sensible first deployment is usually not. This is where each segment tends to begin, based on where the document volume and the rework concentrate.

Commercial banks

Intake and Financial Spreading for commercial lending, then Credit Memo Generation once the spread output is trusted. The analysis layer is where the hours are, and each step reuses the configuration from the one before it. Detail in financial spreading software.

Credit unions

Usually Consumer or Business Deposit Account Opening, because onboarding volume is high and the completeness problem is acute, or Intake for SMB Lending where member business lending is growing. Capacity rather than appetite is the constraint in this segment, so one workflow that demonstrably works is worth more than a broad programme.

SBA and small business lenders

Intake for SBA Lending, where the checklist is long, prescriptive and unforgiving, followed by Bank Statement Analysis. The document cost per file barely moves with deal size in this segment, which is why the return concentrates here.

Equipment finance companies

Intake for Equipment Finance, then Continuous Monitoring. High deal count and heavy documentation make this segment unusually well suited to the document layer; the specific obligations are in equipment finance documentation and compliance automation.

Wealth managers and RIAs

Client Onboarding or Trust and Estate Document Review, where the paperwork is dense and arrives in bursts. Compliance and NIGO Review is the other common entry point, because not-in-good-order rework is measurable and universally disliked.

What none of them do

Worth stating plainly in a page that lists fifty agents. Every one of these prepares, checks, drafts or monitors. None of them decides.

  • Credit decisions. Approve, decline, structure, price. These stay with your credit authority, and the agents change what that person is reading rather than whether they read it.
  • Customer-facing actions. Anything that reaches a customer or moves money is gated behind a person by default, and that gate is configurable tighter but not looser without a deliberate decision.
  • Judgment adjustments. Add-backs, normalisations, treatment of non-recurring items. Proposed with evidence attached, accepted or rejected by an analyst.
  • Policy. Agents apply the credit policy you configure. They do not set it, and a deployment should never move it as a side effect.
  • Regulatory filings and legal responses. Prepared and evidenced, never submitted autonomously.

Where Uptiq fits

The catalogue reflects how Uptiq is built: domain-trained agents for specific regulated workflows, with a Knowledge Team of former underwriters, bankers and analysts certifying extraction accuracy per document type rather than as a single blended benchmark. More than 150 financial institutions run on the platform, across banks, credit unions, non-bank lenders, wealth management firms and equipment finance companies. Agent pages for each entry above live in the Uptiq agent marketplace, where the individual capability, integration and configuration detail sits.

95%+ extraction accuracy certified per document type, 36% less time in financial spreading, 41% faster underwriting, and 63% less credit memo preparation time.Uptiq platform benchmarks across production deployments

How to pick the first one

The catalogue is a menu, not a roadmap. The selection method matters more than the breadth.

Score candidate workflows on four things

How often the work repeats, whether the rules already exist in writing, whether the output needs an audit trail anyway, and whether a mistake is recoverable. Workflows scoring on all four are where deployments have actually landed.

Pick by volume and rework, not by complaint

The loudest pain is often rare. Count files per month through each stage and how much of that work gets redone; the intersection is almost always intake or spreading.

Baseline before anything changes

Elapsed time by stage, touches per file, rework rate, on a representative sample. Without it the improvement becomes an anecdote and the second deployment is harder to fund than the first.

Prove it on one product line

Not the whole book. A defined segment gives a clean comparison and a realistic accuracy read on documents that look like yours.

Extend along the same file

Once intake is trusted, the next agent is the one immediately downstream, because it reuses the configuration and the governance already built. Spreading across departments before deepening in one is how programmes end up broad and unconvincing.

The sequencing detail is in automating underwriting and origination workflows, and the vendor diligence questions in SOC 2 Type II for commercial lending AI.

Frequently asked questions

How many AI agents does Uptiq offer?

More than fifty, across four families: the lending lifecycle (intake, underwriting, analysis and preparation, continuous monitoring), deposits and client onboarding, risk and compliance, and wealth operations. The lending agents are organised by product line, so there are separate intake and underwriting agents for commercial, CRE, SMB, SBA, equipment finance and consumer lending.

Do I have to buy the whole platform?

No. Agents are scoped to individual workflows and deployed one at a time. Most institutions start with a single agent in one product line, prove it against a measured baseline, then extend to the workflow immediately downstream, which reuses the configuration already built.

How long does deployment take?

A single agent is typically live in about five business days, and a multi-agent suite in around thirty, provided the scope is one defined workflow and the configuration uses your existing templates and policy. Timelines stretch when the pilot scope expands mid-project, which is the most common reason these programmes slip.

Does this replace our core or loan origination system?

No. The agents run alongside existing systems, reading and writing through more than 100 native integrations across core banking, origination, servicing, CRM and document management. This is what keeps a deployment a workflow change rather than a platform migration, and it is the single biggest predictor of whether a project ships.

Which agent should we start with?

Usually the one covering the stage with the highest file volume and the most rework, which is almost always intake or financial spreading. Score candidates on repetition, whether the rules are already written down, whether the output needs an audit trail anyway, and whether a mistake is recoverable.

Do the agents make credit decisions?

No. They prepare, check, draft and monitor. Approve, decline, structure and price stay with your credit authority, adjustments are surfaced for a person to accept or reject rather than applied silently, and anything that reaches a customer or moves money is gated behind a person by default.

This listing reflects the Uptiq agent catalogue as published as of September 2026 and is maintained as a reference rather than a specification; availability, naming and coverage by product line change, so confirm current scope with Uptiq before relying on any entry. Performance figures are platform benchmarks across production deployments and are not a guarantee of results at any individual institution.

Start with one workflow

Tell us which stage is consuming your team's week and we will show you the agent that handles it, running on the documents you actually receive.