Pitch Builder
Creates target lists, runs comparables, and drafts pitchbooks for client meetings.
Browse and try the Frontier Agents you can leverage today.
Describe your task in plain language. We'll match you with the top 3 agents that fit best — from the frontier labs, the enterprise systems you already run, and the small-business suites.
Purpose-built agents from the frontier labs, the platforms, and the enterprise systems finance already runs — spanning research and client coverage, finance and operations, and the leaner stack a small business or portfolio company actually operates. Click any card to launch it.
Creates target lists, runs comparables, and drafts pitchbooks for client meetings.
Assembles a briefing on the person you're meeting — background, recent earnings-call commentary, the company's priorities, and where you add value — as a downloadable locked PDF.
Reads transcripts and filings, updates models, and flags thesis-relevant changes.
Creates and maintains financial models from filings, data feeds, and analyst inputs.
Tracks sector and issuer developments and synthesizes news, filings, and broker research.
Org-licensed ChatGPT with file connectors, data privacy, and audit trail — chat with internal finance documents, 10-Ks, board decks.
Formerly NotebookLM — upload earnings transcripts, 10-Ks and board materials; Gemini answers with inline citations and now runs code in a per-notebook cloud computer for grounded data analysis.
Agentic research feature in Gemini (now on Gemini 3): browses hundreds of sites plus your Gmail, Drive and Chat, then returns a multi-page cited report. Strong for sector primers and competitive landscapes.
Analyzes industry reports and competitor data to produce market and competitive intelligence for deal and planning work.
Reasoning research agent (GA June 2025) that runs multi-step research across your emails, meetings, files and the web, returning a source-cited report; now optionally backed by Anthropic models.
Data-analysis agent (GA June 2025) that reasons over raw data in Excel and Copilot chat — calculates statistics, finds trends and outliers, writes the Python, and explains the result.
Checks valuations against comparables, methodology, and the firm's review standards.
Reconciles general ledger accounts and runs net asset value calculations.
Runs the close checklist, prepares journal entries, and produces close reports.
Reviews financial statements for consistency, completeness, and audit-readiness.
Assembles entity files, reviews source documents, and packages escalations for compliance review.
Google Cloud's developer platform for building, scaling and governing enterprise agents — the renamed Vertex AI, with Agent Studio, the Agent Development Kit, Model Garden and BigQuery-grounded finance agents.
Runs collections, dispute resolution, and billing enforcement across the Q2C cycle — prioritizes high-impact accounts by real recovery likelihood. Ema cites ~60% faster dispute cycles and 3-5 day DSO gains.
Source-to-pay automation — validates invoices against POs and receipts, flags mismatches and duplicates before payment is released. Ema cites ~50% shorter AP cycles.
Automates the R2R cycle — reconciles transactions across ledgers and systems instantly and prepares journal entries for review.
Tier-1/2 customer support agent — answers billing, invoice, refund inquiries from customer portals, drafts responses, escalates complex cases.
Automates KYC checks and evidence collection, structures reviews, enforces policy, and maintains audit readiness at scale.
Resolves employee HR, IT and payroll requests autonomously across Teams and Slack, integrating 200+ enterprise apps to answer and execute in one surface.
Purpose-built finance agent in Excel, Outlook and Copilot chat — runs AI-matched reconciliations with audit-ready reports and drives ERP-connected collections.
Agentic system that plans and executes multi-step work across your emails, meetings and files, pausing for approval and returning finished artifacts like budget analyses and executive summaries.
Control plane to observe, secure and govern an organization's AI agents across Microsoft, SaaS and multicloud, with Defender and Intune enforcement.
Generally available. Microsoft's platform for building, deploying and governing agents — production SDKs for Python, JavaScript, Java and .NET, a tool catalog, and observability, in the Foundry portal.
Preview agent for continuous close in Dynamics 365 Finance — proactively raises subledger-to-GL exceptions and recommends the corrective action for each.
Turns manual back-office processes into governed agentic blueprints — 30+ templates including invoice auditing and PO rescheduling, run by specialized back-office agents with a full audit trail.
Joule agents inside S/4HANA and SAP Concur — dispute root-cause analysis, expense report validation, and pre-submit expense audit, at mixed GA and early-adopter stages. SAP cites over 30 agents; the seven below are the finance ones its quarterly release notes have badged.
Agents embedded in Fusion Cloud ERP — Ledger, Payables, Expenses and Payments, generally available in 26B — plus the Collectors and Claims Settlement workspaces from the 12 Fusion Agentic Applications.
Agentic digital workers for the close and collections, coordinated by Vera — Verity Prepare reached general availability in July 2026, Accruals drafts auditable journal entries and Match resolves high-volume transactions; Collect and AR Management are in BlackLine's early adopter program.
Role-based agents across Workday Financials and HCM — Financial Audit, generally available, automates audit evidence collection; Planning is now available; Revenue Contract is in early access; Supplier Contract remains announced-only.
Mid-market finance agents — Accounting AI reconciles against PDF statements and flags anomalies, Finance AI builds board-ready management reports, plus Payroll and Project Management AI.
Build and orchestrate agents alongside robots and people — Agent Builder, Maestro orchestration and a named invoice dispute-resolution agent for finance operations.
Generally available. Build, deploy and operate agents at scale on any framework, model or protocol — runtime, memory, identity, gateway and observability, without managing the infrastructure.
Generally available since November 2025. Spec-driven agentic development in the IDE and terminal — turns prompts into executable specs, then builds across large codebases with parallel agents.
In preview. Answers cost questions, surfaces optimization opportunities, investigates cost anomalies to root cause, and runs recurring FinOps workflows on a schedule.
Available today on AWS and IBM Cloud. IBM's agentic control plane — one place to operate, govern and scale agents, with an agent catalog, workflow builder and observability traces.
Generally available in the Zurich release. An autonomous AI developer that builds ServiceNow applications from conversational prompts, inside the platform's own architecture, governance and data.
Announced generally available with the Zurich release. A single control point for governing AI agents across the enterprise — inventory, oversight and policy for agents wherever they run.
Prebuilt Agentforce skills for banking, insurance and wealth — service assistance, relationship and advisor support, and collections — grounded in CRM and Data Cloud.
OpenAI's toolkit for building and deploying production agent workflows — Agents SDK plus ChatKit and the Connector Registry for governed enterprise data access.
Codex-powered successor to custom GPTs — persistent team agents that connect Slack, Google Drive, Microsoft apps, Salesforce and Notion and complete multi-step work under org governance.
Enterprise platform (2026) for managing AI 'coworkers' with shared business context, execution environments, evaluations and permissions — the governance layer for running agents on real finance work.
OpenAI's work agent (July 2026, GPT-5.6) — gathers context across your connected apps, breaks a goal into steps, and stays on multi-hour tasks to deliver finished spreadsheets, documents, decks, and web apps. Think close packs, board decks, and models produced end-to-end.
Google's employee-facing agent app (which absorbed Agentspace) — search and act across Microsoft 365, Workspace, SharePoint and Jira, with prebuilt Gemini Notebook and Deep Research agents plus no-code Agent Designer.
Claude's agentic workspace — hand Claude a task and it works across your files, calendar, email, messaging and the web, with background work and scheduled tasks. Desktop is generally available; web and mobile are in beta. The delivery surface for the finance agent plugins.
Claude inside Excel, PowerPoint, Word and Outlook — builds and audits models cell-by-cell with traceable changes, drafts decks and memos in place. The add-ins shipped alongside the Anthropic finance agents.
Checks cash on hand against upcoming payroll, flags the gap before it becomes a problem, and shows which receivables would close it.
Walks the month-end close for an owner-operator — categorizes what's uncategorized, reconciles accounts, and explains the P&L in plain language.
A weekly pulse on the business — cash position, sales trend, pipeline movement and what needs a decision this week, in one read.
Plans and runs a marketing campaign end to end — audience, offer, channel copy and the assets to go with it.
Tracks who owes you and how long it's been overdue, then drafts the follow-up — escalating in tone as the days pile up.
Breaks down margin by product, service line or customer so you can see what is actually making money and what is quietly costing you.
Gets the books ready to close — chases uncategorized transactions, missing receipts and unreconciled lines before close starts.
Assembles what the accountant will ask for — categorized expenses, supporting documents and the schedule of what's still outstanding.
Reads a contract before you sign it and surfaces the terms that matter — payment timing, renewal, termination and anything unusual.
Sorts inbound leads by how likely they are to close and how much they're worth, so the first call of the day is the right one.
Plans what to publish and when — a calendar built around what your customers actually ask about, with the drafts to fill it.
Try it live runs a real KokoAI session on this page. See it work plays a scripted preview — illustrative, not the provider's own agent. How to run it opens the details, including the hand-off to your own Claude or the provider's app, where you bring your own account and data.
Parse, extract, and reason over PDFs of any size — financial filings, contracts, reports.
Upload, read, and write files within an Anthropic-hosted workspace. Foundation for document-heavy agents.
Live search the public web with citation-grade results. Critical for research, news, and recent events.
Pull a specific URL into context. Pairs with web search for deep reads of cited sources.
Sandboxed Python for analysis, modeling, charts. Stateful across turns within a session.
Drive a browser or desktop autonomously by interpreting screenshots and emitting actions.
Anthropic-shipped Skill: enforce brand voice and visual rules on generated docs and decks.
Generate well-formatted Word docs from structured inputs — board memos, controller letters, finance narratives.
Generate Excel workbooks with formulas, formatting, and multi-sheet structure — financial models, variance packs.
Generate PowerPoint decks from structured inputs — board packs, CFO command-center reviews.
Reference agent that answers questions about financial filings using PDF + code execution. Open-source on GitHub.
Reference implementation for a tool-using customer service agent with escalation logic.
Multi-step research with web search, source synthesis, and structured report generation.
Retrieval-augmented Q&A with vector store + reranking. Baseline pattern for internal-knowledge agents.
Coordinator pattern: lead agent delegates to specialist subagents, merges results, returns synthesis.
End-to-end Computer Use reference — Docker, screenshots, action loop, safety guardrails.
Fast read-only search agent for locating code: find files, grep symbols, answer 'where is X defined / what references Y'.
Software-architect agent for designing implementation plans. Returns step-by-step plans and critical files.
General agent for researching complex questions, multi-step tasks, and open-ended searches.
Independent reviewer for diffs — surfaces correctness bugs, with effort tiers from low to max.
Q&A about Claude Code itself, the Claude Agent SDK, and the Anthropic API.
Configure the Claude Code status line via settings.json.
Read/write GitHub issues, PRs, files, commits, CI status. Standard MCP server bundled with Claude clients.
Bridge code and design — read Figma into code, push code into Figma, manage design tokens.
Read, search, label, draft emails and manage threads.
Query and edit Notion databases and pages from inside Claude.
Post, search, and react to Slack messages across channels and DMs.
Manage Linear issues, projects, and cycles from agent workflows.
Run read-only SQL against any Postgres database — analytics, ad-hoc data pulls, schema inspection.
Query BigQuery warehouses including finance, planning, and operational marts.
Query Snowflake warehouses for finance and analytics workloads.
Connect bank accounts, fetch balances, transactions, and identity verification for finance agents.
Read and manage Stripe customers, invoices, subscriptions, and payment events.
Read/write files in a sandboxed directory — the simplest MCP server, useful as a building block.
Want the full catalog as a locked, board-forwardable PDF? Download an executive overview — coverage by provider, the finance-featured highlights, and the ecosystem catalog.
Margins, FCF conversion, ROIC, and the working-capital cycle (DSO/DPO/DIO/CCC), percentile-ranked against sector peers.
Assemble a company-specific, persona-framed executive deck from the site's own intelligence.
Cited answers across news, vendors & capabilities.