Top 10 AI Prompts and Use Cases and in the Financial Services Industry in Wichita

Wichita's financial services sector faces a practical imperative: craft prompts that make AI useful, reliable, and compliant for local banks, credit unions, and fintechs.

With a Mercury survey showing 68% of AI adopters expanding teams rather than cutting jobs, firms in Kansas are scaling AI-driven roles in finance, sales, and customer service (Mercury survey on AI adoption in startups), yet US CFOs still flag security and privacy as top barriers - a trust gap prompt engineering can help close.

The upside is huge: AI-in-finance forecasts and adoption rates point to rapid efficiency and fraud-detection gains (AI in finance statistics and trends), and practical training - like the Nucamp AI Essentials for Work bootcamp registration - gives Wichita teams the prompt-writing skills to pilot secure models, automate cash forecasts, and run real-time fraud detection systems that protect Wichita customers and cut losses.

Bootcamp	Details
AI Essentials for Work	15 Weeks; Courses: AI at Work: Foundations, Writing AI Prompts, Job Based Practical AI Skills; Early-bird cost $3,582; Register for Nucamp AI Essentials for Work bootcamp

“AI-focused skills will empower finance professionals to confidently work with AI technologies and bridge the trust gap by ensuring decisions made by AI systems are transparent and understandable. … By combining human expertise with AI's analytical capabilities, organizations can make more informed decisions.” - Morné Rossouw, Chief AI Officer, Kyriba

Selection balanced practical impact for Wichita institutions with regulatory sensitivity and technical feasibility: prompts were scored for measurable ROI (shortening loan cycle times, tightening fraud detection, automating cash forecasts), aligned to common community-bank workflows highlighted by nCino's 2025 analysis of targeted, workflow-level AI use (nCino AI Trends in Banking 2025 analysis), and weighted by a “sliding scale” of scrutiny that RGP recommends - higher oversight for credit, trading, and fraud use cases, lighter controls for back-office automation (RGP AI in Financial Services 2025 report).

Each candidate prompt also passed a regulatory-and-governance filter informed by recent GAO summaries of finance use cases and risks, ensuring data-quality, explainability, and human-in-the-loop checkpoints (Consumer Finance Monitor AI in Financial Services overview).

The result: ten prompts chosen for high local relevance to Wichita banks and credit unions, fast path-to-value, and an auditable compliance posture - so pilots can move from sandbox to production without getting stuck in regulatory limbo.

For Wichita treasurers and FP&A teams, a prompt that says “Refresh our 13‑week cash forecast for Wichita operations”

becomes an operational superpower when routed to an agent that sits on top of ERPs, bank feeds, AR ledgers and payroll schedules - joining GL and budget data, updating drivers, flagging stale assumptions, and returning a reconciled forecast plus narrative and exportable workbooks in seconds, not days; Concourse shows agents can make rolling forecasts continuous and turn a full analyst day into an on‑demand update (Concourse insights on AI agents for finance and rolling forecasts), while PwC highlights that agents can consolidate cash positions, predict near‑term inflows/outflows, recommend transfers or sweeps, and improve forecasting speed and accuracy (PwC cites up to a 40% forecast accuracy boost) - which in Wichita means fewer surprises for community banks and tighter liquidity for mid‑market firms, with audit trails and human approvals built into each action (PwC analysis of AI agents for treasury forecasting and liquidity management).

The “so what?” is concrete: what used to be a day of copy‑paste and model juggling becomes a repeatable, auditable prompt that surfaces risks (shortfalls, >90‑day receivables) and prescribes next steps in minutes, freeing teams to focus on strategy rather than spreadsheets.

Source	Key outcome
Concourse	Agents enable on‑demand forecast refreshes, eliminate hundreds of manual hours, and produce narrative + workbook outputs
PwC	Agents can consolidate cash positions, recommend liquidity actions, and improve forecasting accuracy by up to 40%

The prompt “Handle balance, transfers, and proactive alerts” maps neatly to Fargo's proven capabilities in the Wells Fargo Mobile app - voice or text handling of bill pay, transfers (including Zelle), balance forecasts, transaction searches and proactive spending alerts - so Wichita banks can see how a single, well‑scoped prompt turns routine customer requests into instant, auditable actions that reduce call‑center load and surface risks like rising subscription charges (42% of Americans admit they forget unused subscriptions).

Fargo's production design also shows how to scale safely: the assistant handled over 245 million interactions in 2024 while keeping sensitive data out of the LLM pipeline, and it supports English and Spanish text/voice interactions for broad accessibility (Wells Fargo AI assistant 245M interactions and privacy-first pipeline, Wells Fargo Fargo virtual assistant overview and features).

The “so what?”: a Wichita customer can get a reconciled balance, move money, and receive a proactive alert in one interaction - no transfers of PII to models, and the bank retains orchestration and audit control.

Metric / Capability	Detail
2024 interactions	245.4 million (production system)
Spanish usage	Accounts for >80% of usage in 2024
Core actions	Balance checks, transfers, bill pay, spending insights, turn on/off cards, Zelle
Privacy & architecture	PII scrubbed by internal SLM / orchestration layer; models used for intent only

“Fargo has brought a more simple and intuitive banking experience through its concierge-like experience. It helps customers meet their individual financial needs… Using Fargo, customers can get answers to their everyday banking questions and, rather than taking multiple manual steps to accomplish a task, they can ask Fargo to complete the task for them.” - Kevin Cole, Wells Fargo head of digital AI

AskResearchGPT turns a prompt like “Search proprietary reports and summarize findings” into a practical tool Wichita financial teams can use to cut research time: the Morgan Stanley release describes an assistant that sifts its 70,000+ proprietary reports, synthesizes answers with hyperlinks to source documents, and even exports findings into a client-ready email draft with one click - so a local M&A or wealth‑management team can swap hours of PDF hunting for a concise, auditable brief in minutes (Morgan Stanley press release about AskResearchGPT capabilities).

Built on GPT‑4 and rolled into advisor workflows, the capability mirrors broader industry gains in retrieval and meeting‑prep efficiency reported by CTOMag and others, making deep research accessible to regional banks and brokerages that lack large in‑house analyst desks (CTOMag coverage of AI adoption at Morgan Stanley); the “so what” is vivid: instead of juggling stacks of reports, a Wichita advisor can deliver a sourced summary and next‑step recommendations before a client's lunch break.

Metric / Feature	Detail
Launch date	Oct 23, 2024
Research corpus	70,000+ proprietary reports
Model	GPT‑4 (integrated into AskResearchGPT)
Output	Summaries with citations and one‑click email export
Primary users	Investment Banking, Sales & Trading, Research staff

“AskResearchGPT is emblematic of our tech‑forward philosophy in Institutional Securities.” - Katy Huberty, Global Director of Research, Morgan Stanley

In Wichita finance shops a Concourse/Stratpilot–style prompt - “Flag GL accounts with >10% variance and explain drivers” - turns routine flux checks into a practical watchtower: automated rules pick up material swings, attach percent and dollar context, and surface likely causes so controllers can investigate before auditors call; industry guides show this is both practical and expected - NetSuite explains percent‑variance formulas and why an 11% swing on a $100k line needs narrative, not just numbers (NetSuite guide to variance and flux analysis), Numeric highlights how AI and workflow tools auto‑generate explanations and speed investigations, turning a slow month‑end scramble into repeatable, auditable commentary (Numeric variance analysis and AI workflow guide), and Trintech notes the value of configurable thresholds and assignment workflows so a >10% flag immediately routes to an owner (Trintech guide to variance thresholds and month‑end close).

The “so what?” is vivid: what used to be a stack of spreadsheets and detective work becomes an on‑demand anomaly alert that names likely drivers, recommends next steps, and preserves audit trails - helpful for community banks, regional brokerages, and CFOs across Kansas who need timely, explainable GL controls.

Source	Key takeaway
NetSuite	Defines percent variance formulas and shows why narrative explanations are required for material swings
Numeric	Shows AI and tools can auto‑generate variance explanations and speed the variance cycle
Trintech	Recommends configurable thresholds and workflows to flag and assign variance investigations

A Master of Code Global–style credit‑decisioning prompt

Use alternative data to score Wichita applicants

frames a practical path for local lenders to reach credit‑invisible residents by combining non‑traditional signals (rent and utility payments, gig economy income, mobile and subscription payment patterns, and digital footprint signals) with bureau data to produce richer, explainable scores; the Kansas City Fed notes that fintechs and bureaus are already collecting alternative data to expand access Kansas City Fed: Using Alternative Data to Expand Credit Access, RiskSeal's 2025 guide shows how digital footprints and pay patterns can improve approval accuracy and fraud detection RiskSeal: Mastering Credit Scoring with Alternative Data (2025 Guide), and a cautionary note from Greenlining underscores the need for strong governance so those same signals don't replicate bias or invasive surveillance Greenlining: FCRA and Equity Risks of Alternative Scores.

For Wichita banks and credit unions the

so what

is tangible: a borrower who pays rent and utilities on time but lacks a bureau file can be included responsibly - if models are audited, consumers can dispute profiles, and data choices avoid proxies that entrench discrimination.

Alternative data type	Why it matters
Rent & utility payments	Signals payment consistency for applicants without credit files
Gig income & transaction patterns	Captures non‑traditional earnings and cash flow stability
Subscription & mobile payments	Indicates timely, recurring commitments
Digital footprint / phone & email lookups	Helps detect identity risk and fraudulent applications

For Wichita compliance teams, a prompt such as “Continuous transaction monitoring and SAR draft” can follow OCBC's playbook - pairing an Azure OpenAI–backed assistant with a secure, bank‑hosted orchestration layer so alerts and draft SAR narratives never expose raw PII to external models; OCBC reports the OCBC GPT pilot helped staff finish writing and research tasks about 50% faster and is already used to flag suspicious transactions as part of AML workflows, giving a practical blueprint for regional banks to automate triage while preserving human review (OCBC GPT pilot on Azure OpenAI and secure hosting, OCBC deployment of generative AI chatbot for staff).

The “so what?” is tangible for Wichita: instead of a compliance analyst assembling threads of transactions by hand, a governed agent can surface linked anomalous behaviors, propose lines of investigation, and draft a SAR-ready narrative for investigator sign‑off - accelerating response time without surrendering auditability or control.

Metric / Capability	Detail
Users / Scope	OCBC GPT rolled out to 30,000 staff (pilot ~1,000)
Productivity lift	~50% faster task completion in trials
AI decisions	~4 million AI decisions daily (risk, service, sales)
Security posture	Hosted in a secure, bank‑controlled environment; inputs not shared externally

“We are excited to be one of the first banks in the world to deploy generative AI tools at scale. We believe that these tools have the potential to transform the way our employees work…” - Donald MacDonald, OCBC Head of Group Data Office

A Wichita-ready prompt like “Generate tailored savings and retirement recommendations” maps cleanly to Bank of America's Erica capabilities: within the Mobile Banking app Erica already delivers proactive, personalized insights (bill reminders, spending & budgeting snapshots, Life Plan tools) and can surface eligibility cues or connect users to Merrill specialists for retirement conversations - while clearly stopping short of offering formal investment advice (Erica virtual financial assistant features and capabilities, Bank of America newsroom: Erica milestones and achievements).

For Kansas households and community-bank customers in Wichita, that means a single, authenticated in‑app interaction can highlight shortfalls, recommend savings targets tied to recent spending patterns, and route users to a human advisor for retirement planning - letting frontline specialists focus on complex cases while routine, personalized guidance arrives faster and with an audit trail.

Metric	Figure
Users (since launch)	Nearly 50 million
Total client interactions	3 billion+
Monthly interactions (avg)	~58 million
Proactive personalized insights	1.7 billion+
Users finding needed info	~98%

“Our clients appreciate Erica's ability to help them manage their spending, improve budgeting and increase savings. Erica is the bedrock upon which we've built an unmatched high-tech, high touch client experience.” - Nikki Katz, Head of Digital, Bank of America

A Concourse/RTS Labs–style prompt - “List top 10 overdue Wichita customers and recommended collection actions” - transforms collections from scattershot chasing into a focused recovery playbook: the agent ingests AR aging, payment history and dispute flags, ranks accounts by risk and value, drafts tailored dunning messages, and recommends next steps (soft reminder, phone outreach, payment plan or hold on new shipments) so Wichita banks and regional lenders can close the cash gap faster while preserving customer relationships.

AR automation best practices show this reduces manual errors, speeds reconciliation, and frees teams to work high‑impact accounts rather than spreadsheets; integration with ERPs and customer portals lets reminders hit inboxes with one‑click pay links, and machine‑learning prioritization surfaces the accounts most likely to yield quick wins (NetSuite guide to accounts receivable automation).

For midmarket Wichita firms, the “so what” is tangible: a system that used to require days of detective work becomes a prioritized, auditable action list delivered in minutes - paired with dashboards to track DSO and recovery progress (Versapay AR automation and collections solutions).

Outcome	Reported impact
Reduced DSO	85% of CFOs at firms automating >50% of AR reported improvements (NetSuite)
Time saved managing receivables	~50% less time (Versapay)
Faster payments / fewer past‑due invoices	~25% faster payments; ~30% fewer past‑due invoices (Versapay)

“We've transformed our finance operations from a manual mess to a streamlined, strategic advantage; saved the cost of a full‑time associate; and reduced DSO significantly.” - Danny Ng, VP of Finance, Crystorama

For Wichita banks and credit unions wrestling with dense policy manuals, regulatory filings, and weekly internal reports, an employee copilot modeled after Raiffeisen Bank's RBI ChatGPT offers a pragmatic playbook: hosted on Azure OpenAI Service and Azure AI Foundry, the assistant can ingest policy documents, extract the required clauses, and draft internal summaries or report skeletons that preserve citations and traceability - freeing compliance and HR teams to focus on judgment, not copy‑paste.

RBI's rollout shows this pattern scales: initial pilots aimed at automating legal and regulatory summaries and drafting internal content, while “yellowGPTs” let staff contribute dedicated knowledge bases for repeatable answers - an approach Kansas institutions can mirror to reduce review backlog and tighten audit trails.

Built‑in content safety and enterprise search components mean sensitive text stays governed, and Microsoft's broader AI playbook underscores how copilots boost employee productivity when paired with governance and role‑based access controls.

For Wichita, that translates into faster, auditable policy briefs, sharper investigator-ready reports, and a single, searchable knowledge layer for frontline staff and back‑office teams (RBI ChatGPT deployment on Azure OpenAI Service case study, Microsoft AI-powered customer transformation and Copilot guidance blog).

Metric / Capability	Detail
Platform	Azure OpenAI Service, Azure AI Foundry, Azure AI Search, Azure AI Content Safety
Adoption	Pilot to 2,000 users; active base grew to 20,000 (RBI)
Primary use cases	Summarize legal/regulatory docs, draft internal reports, quick customer summaries, content creation

“We hope to make everyone an AI power user. Every employee needs to understand how to interact with language models to remain competitive and provide the best customer experiences.” - Armin Woworsky, Distinguished Engineer, RBI

Generating Wichita‑like synthetic datasets lets local banks and credit unions train models without putting real customer records at risk: research from Google Research shows differentially private pipelines - using DP‑SGD and parameter‑efficient fine‑tuning techniques like LoRa or prompt tuning - can produce high‑quality synthetic training data while formally limiting privacy loss (Google Research: Differentially Private Synthetic Training Data), and academic work on privacy‑preserving deep‑learning releases reinforces the same approach (Harvard Kennedy School: Privacy‑Preserving Synthetic Data Release Using Deep Learning).

For Wichita use cases, that means pilots can spin up representative, auditable datasets that mirror local transaction patterns and customer segments for model validation, bias testing, and explainability checks - without exposing PII to external models.

The practical payoff is clear: safer model experimentation, faster iteration, and evidence‑backed explanations for regulators and boards; local teams should pair synthetic data with strong bias‑monitoring and explainability practices as recommended for practitioners (Nucamp AI Essentials for Work - model explainability and bias monitoring recommendations), turning privacy safeguards into a competitive enabler rather than a roadblock.

Getting started in Wichita is a pragmatic three‑step loop: pilot a single, high‑friction prompt (think cash‑forecast refresh or AR prioritization), lock in governance and human‑in‑the‑loop checkpoints from day one, then scale with a measured roadmap that ties pilots to KPIs - exactly the field playbook Cornerstone Advisors and Hapax recommend for community banks and credit unions (Cornerstone Advisors Playbook for Generative AI‑Driven Productivity).

Layer in policy and audit controls from the Consumer Bankers Association's white paper to keep compliance and consumer trust front and center (Consumer Bankers Association playbook for compliant AI in financial services), follow a phased AI roadmap (foundation, expansion, maturation) to avoid scattered pilots (AI roadmap guide for financial services), and invest in practical prompt‑writing and governance training - programs like Nucamp's AI Essentials for Work help local teams move from experimentation to repeatable, auditable outcomes (Nucamp AI Essentials for Work bootcamp page).

Start with one measurable win, keep regulators and auditors in view, and the pilot→govern→scale rhythm will turn one‑off gains into durable operational advantage for Wichita institutions.

Bootcamp	Length	Early‑bird cost	Register
AI Essentials for Work	15 Weeks	$3,582	Register for AI Essentials for Work (Nucamp)

“AI is about unlocking new growth opportunities for financial institutions.” - Ron Shevlin, Chief Research Officer, Cornerstone Advisors