Fieldguide
Also known as: Field Agents, Fieldguide MCP, fieldguide.io
Agentic AI platform for audit and advisory firms that automates workpapers, testing, and engagement workflows.
Fieldguide is an AI native platform for audit and advisory firms that puts a workforce of Field Agents alongside practitioners across the whole engagement, from planning through sign off, and is used by half of the top 100 US accounting firms including KPMG, RSM US, Baker Tilly, BDO, Grant Thornton and Forvis Mazars.
Five named agents do the work: Field Planner ingests prior year files and engagement history, assesses risks and builds a ready to review plan; Field Auditor analyses evidence gaps and runs testing end to end across population, sampling, extraction and evaluation; Field Reviewer checks completed workpapers for accuracy and compliance and surfaces exceptions before partner sign off; Field Financials drafts financial statements and disclosures with every figure tied to its source; and Field Orchestrator takes a direction in plain language and coordinates the others across planning, execution and review with dependencies sequenced automatically, working every engagement in parallel while context carries forward between them.
Practitioners choose when agents pause for review, change that cadence at any time, and approve or redirect work at every checkpoint, and plans and execution follow the firm's own methodology across financial audit, SOC, IT audit, employee benefit plan, review and compilation, ESG, risk advisory and internal audit engagements, with a library of frameworks from SOC 2 and HIPAA to PCI, HITRUST and ISO.
Fieldguide MCP connects the same engagement data to the Copilot, Claude or OpenAI assistant a firm already uses, so a practitioner can ask what is open and overdue or start an engagement from a template with their existing permissions carried over. Fieldguide holds ISO 42001, SOC 2 Type 2 and AIUC-1 certification of its agents, enforces multi factor authentication, encrypts data at rest and in transit, hosts on AWS in the United States, and backs a 99.9 percent uptime SLA.
Founded in 2020 and valued at $700 million after a $75 million Series C led by Growth Equity at Goldman Sachs Alternatives in February 2026, it is sold by demo. It does not offer customer chosen models, self hosted deployment or a public developer API.
Vendor details
Canonical URL
https://www.fieldguide.com
Category
Enterprise operations agent
Subcategory
Audit and advisory agentic workpapers
Funding status
Independent, founded in 2020 by former CPA Jin Chang. Raised a $75 million Series C led by Growth Equity at Goldman Sachs Alternatives in February 2026 at a $700 million valuation, bringing total funding to $125 million, with participation from Bessemer Venture Partners, 8VC, Geodesic, and Thomson Reuters.
Company status
independent
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
Centralizes the engagement on one cloud native platform, from client requests and evidence intake through testing, documentation, and financial statements. Field Agents ingest and validate uploaded client evidence and execute substantive procedures across audit, SOC, ESG, and risk advisory practice areas, spanning frameworks such as SOX, HIPAA, HITRUST, and PCI DSS.
In practice
Your audit practice cannot hire fast enough as senior CPAs retire. Fieldguide's agents run the first pass on testing and evidence review, letting a smaller team deliver the same engagement volume.
You are sending testing and preparation work offshore and losing control of quality. Fieldguide keeps that first pass onshore in one platform, with agents executing and your people reviewing.
A SOC or risk engagement is buried in client evidence and PBC back and forth. Fieldguide reviews uploaded evidence for completeness up front, flagging missing items before they stall the engagement.
Sources & related URLs
Agentic Index coverage score
7.5 / 14 capabilities · 54%
| Integrations & Tool Calling | Partial |
|---|---|
|
The platform takes in prior year files, client evidence and engagement data, and Fieldguide MCP lets engagement status flow into Teams, SharePoint and email through the practitioner's own assistant. Nothing shows the agents themselves taking authenticated action in an outside system; an Integrations page in the site navigation may cover that. SourceFieldguide, fieldguide.com/products/platform/fieldguide-mcpread 2026-09-06 |
|
| Workflow Orchestration | Full |
|
Field Orchestrator takes a natural language direction and activates Field Agents across planning, execution and review with dependencies sequenced automatically, Field Planner building the engagement plan from prior year files, Field Auditor running population, sampling, extraction and evaluation end to end, Field Reviewer checking workpapers and Field Financials drafting statements, across every engagement in parallel. Multi step, multi agent execution across a whole engagement on the vendor's own runtime is documented. SourceFieldguide, fieldguide.com/products/agents/field-orchestratorread 2026-09-06 |
|
| Knowledge Grounding & RAG | Full |
|
Plans and execution are grounded in the firm's own methodology and standards, applied consistently across every engagement. A Methodology and Frameworks module carries libraries for SOC, HIPAA, PCI, HITRUST, ISO and other frameworks, and Field Planner takes in prior year files and engagement history. SourceFieldguide, fieldguide.com/products/agents/field-orchestratorread 2026-09-06 |
|
| Human Oversight & Guardrails | Full |
|
Practitioners choose when agents pause for review, change that cadence at any time, and approve or redirect work at every checkpoint. Every Field Agent returns its work for review, and Field Reviewer surfaces exceptions before partner sign off. Fieldguide MCP actions are shown in their exact wording for review before anything changes. SourceFieldguide, fieldguide.com/products/agents/field-orchestrator and fieldguide.com/products/platform/fieldguide-mcpread 2026-09-06 |
|
| Security, Identity & Governance | Full |
|
Three attestations back Fieldguide's security: ISO 42001 certification, an annual SOC 2 Type 2 examination covering security, availability and confidentiality, and an AIUC-1 attestation of the agents. A trust center sits at trust.fieldguide.io, data is encrypted with AES-256 at rest, and every account uses multi factor authentication. Access follows engagement permissions: on the MCP page, each practitioner sees only the engagements and data they can already access. SSO, SAML and SCIM through the customer's identity provider are not named. SourceFieldguide, fieldguide.com/trust and fieldguide.com/products/platform/fieldguide-mcpread 2026-09-06 |
|
| Observability & Auditability | Partial |
|
Field Financials ties every figure to its source and flags exceptions, Field Reviewer surfaces exceptions in completed workpapers before sign off, and work returns for review at each checkpoint. What is missing is a record of the agents themselves: no run trace of an agent's steps or audit log of agent and user actions is described, and there is no export. SourceFieldguide, fieldguide.com/products/agents/field-orchestratorread 2026-09-06 |
|
| Memory & State Persistence | Partial |
|
Field Orchestrator works as a running conversation that builds on every prior decision, direction and result. Context carries forward, so a practitioner can jump between engagements and continue, and prior work, client responses and review notes feed every next step. That state lasts beyond a single session for each engagement, but no memory layer with its own lifetime and deletion scope apart from the engagement's records is described; prior engagement work is the application's own data. SourceFieldguide, fieldguide.com/products/agents/field-orchestratorread 2026-09-06 |
|
| Deployment & Data Residency | Not documented |
|
Fieldguide is hosted and managed on AWS data centers in the United States, according to its trust page, which also offers a Data Processing Addendum and Standard Contractual Clauses for EU to US transfer. A customer cannot select a region or bring its own cloud, and there is no on premises or dedicated tenant option. The region is fixed and disclosed, and legal cover for a transfer is not the same as keeping data in the EU. SourceFieldguide, fieldguide.com/trustread 2026-09-06 |
|
| Prebuilt Agents, Templates & Packs | Full |
|
Five named Field Agents are sold on their own product pages: Planner, Auditor, Reviewer, Financials and Orchestrator, each doing a distinct job in the engagement. Alongside them sits a Methodology and Frameworks library of templates across SOC, HIPAA, PCI, HITRUST and ISO. The agents stand alone: take away Field Financials and Field Auditor still runs testing end to end. SourceFieldguide, fieldguide.com/products/agents/field-orchestrator and fieldguide.com/trustread 2026-09-06 |
|
| Triggers & Channel Coverage | Partial |
|
Field Agents start when a practitioner directs them in natural language, through Field Orchestrator or through Fieldguide MCP from an outside assistant, and then work every engagement in parallel. Nothing shows an event, schedule or webhook starting agent work, not even client evidence arriving through Request Management. SourceFieldguide, fieldguide.com/products/agents/field-orchestrator and fieldguide.com/products/platform/fieldguide-mcpread 2026-09-06 |
|
| Model Flexibility & Routing | Not documented |
|
Fieldguide does not name the model provider behind its agents on its trust, MCP or Orchestrator pages, and no customer or admin can select a model. The trust page points to a limited number of subprocessors listed at fieldguide.com/subprocessors, which may name one. SourceFieldguide, fieldguide.com/trustread 2026-09-06 |
|
| APIs, SDKs & MCP Extensibility | Partial |
|
Fieldguide MCP is an MCP server through which a practitioner's own assistant, whether Copilot, Claude or OpenAI, or any MCP compatible tool, queries live engagement data and takes approved actions in Fieldguide, such as posting a comment or starting an engagement from a template. Each user's existing permissions carry over, and the call runs from outside into the platform. The technical reference is missing, though: no endpoint, auth scheme or configuration reference is published, and there is no REST API or SDK. SourceFieldguide, fieldguide.com/products/platform/fieldguide-mcpread 2026-09-06 |
|
| Testing, Debugging & Optimization | Not documented |
|
Nothing lets a customer test a change to an agent: the trust, MCP and Orchestrator pages show no evaluation harness or scored test cases, and no quality gate or optimization loop. AIUC-1 is a third party attestation of the vendor's agents, and the configurable review cadence gates each output rather than testing a change. The sub one percent error rate Fieldguide has quoted is its own benchmark of its own product. SourceFieldguide, fieldguide.com/trust and fieldguide.com/products/agents/field-orchestratorread 2026-09-06 |
|
| Browser & Computer Use | Not documented |
|
The Field Agents work over evidence, prior year files and workpapers inside Fieldguide's own platform, and outside tools reach them through Fieldguide MCP. No browser, desktop or computer control is mentioned. SourceFieldguide, fieldguide.com/products/agents/field-orchestrator and fieldguide.com/products/platform/fieldguide-mcpread 2026-09-06 |
|
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
Fieldguide introduced an MCP integration that connects its audit and advisory platform to AI tools such as Claude, Microsoft Copilot and ChatGPT. Users can query engagement data and check request status from those assistants.
Bears on: MCP / tool calling / API
View sourceFieldguide achieved FedRAMP Moderate Authorization in partnership with Knox Systems. This certification enables the rollout of a dedicated federal environment for the platform's AI agents.
Bears on: Deployment / data residency
View sourceFieldguide launched Field Financials, introducing three AI agents to draft and validate financial statements, alongside Fieldguide MCP. Built on the Model Context Protocol, the MCP integration connects live engagement data directly into external AI assistants like Claude, Copilot, and Gemini.
Bears on: MCP / tool calling / API
View sourcePricing
Contact for pricing
What is public
No public rate. Fieldguide publishes no list pricing and routes all commercial terms through sales.
Billing mechanics
Negotiated annual enterprise contracts scoped to engagement volume, practice areas, and firm size. Procurement runs through sales with product tours and demos rather than self serve signup.
Cost watchouts
Scope creep is the main exposure: adding practice areas or extending agents across more of the engagement lifecycle expands the contract at renewal. Confirm what agent capacity and practice coverage the base term includes.
Variable cost rationale
Priced as a negotiated annual platform contract scoped to engagement volume and practice areas; cost is predictable once scoped, but expanding practices or deeper agent adoption raises the next renewal.
Additional watchouts
With no public rate, benchmark across comparable firms before signing, and clarify how expansion into new practice areas or deeper agent adoption changes the renewal.
Sales call required
Yes, required for paid access
Free / trial
Request Demo on every page; no free tier or trial
Lowest paid plan
Not published
Key ambiguities
No entry rate, billing unit, or tiering is published; all pricing detail is negotiated under sales.
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 Fieldguide
The closest documented capability profiles to Fieldguide 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.
- Alloy.ai8.0 / 14Fuller documented coverage on Integrations & Tool Calling and Triggers & Channel Coverage
- Eudia7.0 / 14Adds documented Model Flexibility & Routing
- Findem7.0 / 14Fuller documented coverage on Integrations & Tool Calling
- HROne7.0 / 14Fuller documented coverage on Triggers & Channel Coverage
- Parachute6.0 / 14A lighter documented profile than Fieldguide
- Rivvun AI9.0 / 14Adds documented Model Flexibility & Routing
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded