One graph behind every build.
Your standards on the inner ring. ServiceNow’s best practice on the outer ring.
An AI-native ServiceNow teammate that doesn’t just advise — she analyzes, plans, develops, tests, and deploys.
Why do P1 incidents keep getting reassigned?@ for records · / for commands
Run the platform, implement a product, upgrade, or assess and clean up — Phyllis does the work in your instance, grounded and safe.
The everyday work — backlog, fixes, questions and health — done in your instance and scanned after every change.
Explore RunITSM, HRSD, CSM and every product in between — planned, built, tested and ready to promote.
Explore ImplementSkipped records triaged, regression tested and released to test — on your approval.
Explore Upgrade| Finding ↓ | Records | Severity | Remediation |
|---|---|---|---|
| CIs without a servicecmdb_ci_server | 212 | Map to application services | |
| Duplicate server CIscmdb_ci_server | 87 | Reconcile on serial number | |
| Stale discovery sourcecmdb_ci | 1,340 | Retire or rediscover | |
| Class not in CSDM 5cmdb_ci_appl | 46 | Move to the business application | |
| Missing ownercmdb_ci_service | 19 | Assign a managed-by group |
cmdb_ci_server · no cmdb_rel_ci to a service · 212 rowsCMDB, tech debt, performance, licences and AI readiness — each ends in a report you can hand over.
Explore AssessWhoever you are on the team, she takes the work you’d rather hand off.
Turn a one-sentence brief — or a workshop transcript — into sized user stories with acceptance criteria, grounded in the live instance, not a workshop guess.
She’ll take requests likeBusiness rules, client scripts, flows and ACLs in scoped update sets — every write read back and staged for your approval.
She’ll take requests likeCorrect scoping, naming and ACL coverage across the whole estate — every decision cited to the record it came from.
She’ll take requests likeAn Instance Scan with hundreds of checks after every task, self-healing on approval — and controls that hold on production.
She’ll take requests likeThe implementation sequenced and synced to Jira or ServiceNow Agile 2.0, released to test on approval — so the team always knows what ships next.
She’ll take requests likeATF written and run for what changed, screenshots checked against the build — nothing moves until it’s green.
She’ll take requests likeClass audits, orphaned CIs, stale references — Phyllis reads the live CMDB, cites every finding, and stages the cleanup for your approval.
She’ll take requests likeFrom a sentence to a promoted, tested change — every step staged for your approval.
Phyllis grounds herself in the full context of your ServiceNow instance — its tables, ACLs, flows, update sets, configurations, and dependencies. Every answer is evidence-backed and traceable to the records it came from.
Explore AnalyzeSay “I want to implement ITSM” and Phyllis returns sequenced stories with priorities, points, and acceptance criteria — synced both ways with Jira or ServiceNow Agile 2.0.
Explore PlanPhyllis develops on the platform — from tables, business rules, flows, and ACLs to the records behind them — then verifies and tests every change.
Explore DevelopBatch and release update sets through chat — in the right order, with conflicts surfaced before you promote.
Explore DeployATF suites and screenshot validation, plus an Instance Scan with hundreds of checks across six categories — nothing promotes until it’s green.
Explore TestBefore Phyllis builds anything, she queries the Phyllis Knowledge Graph — your live instance, your team’s conventions, what worked last time, your backlog and meetings through Phyllis Plugins, hundreds of expert ServiceNow skill packs and checks, the official docs for your release and a curated learning corpus. So she reuses what exists, follows the standard, and leaves less tech debt behind.
Your standards on the inner ring. ServiceNow’s best practice on the outer ring.
Add a laptop refresh request with manager approval.
One sentence — no template, no ticket form.
Your instance, your standards, the right skill pack and the docs for your release.
Existing records reused and scan-clean before review.
Before Phyllis changes a field, a table or a script include, she maps everything that depends on it — including the references that only exist as names inside scripts and conditions.
Rename incident.u_business_unit
Half of them only exist as names inside scripts and conditions.
var unit = gs.getUser().getRecord().getValue('department'); answer = current.u_business_unit == unit;
A typed reference fails the moment a form opens. A name in a script fails later — so she finds both.
Controls that hold at the connection itself — whichever tool, chat or workflow makes the call.
rm_storyrm_epicrm_scrum_tasksys_scriptsys_hub_flowsys_security_aclsys_propertiessys_ui_policysc_cat_itemConnect production for the backlog alone. Agile management only is enforced on every request Phyllis makes — not just hidden from view — and new production instances start there.
Nothing is written without your approval. Every write is read back and checked, and every approval and access change lands in the audit log.
Choose which models are on, block tables and switch off tools per instance, and watch credits move live while Phyllis works.
Every person connects with their own OAuth token, so Phyllis can only do what they can — and owners decide who sees which instance.
Phyllis is ISO/IEC 27001 certified — the international standard for information security management. Independently audited controls across access, data handling and operations, so your security team has a recognised baseline to sign off against.
Phyllis Plugins connect her to the tools around your instance — your backlog, your meetings — so the work starts where the conversation happened.
Stories, points and acceptance criteria land on your Jira board — and status, comments and changes sync back.
The same backlog sync for teams planning in Azure Boards — user stories, tasks and iterations.
Bring in a Teams meeting transcript and Phyllis drafts the stories — each one traced back to the moment it was said.
Connect Zoom and Phyllis reads the recording’s transcript, pulls out the requirements and turns them into stories for your review.
No. Every change is built in a scoped update set and staged for review with the design doc, scan, and test results attached.
Promotion requires an explicit approval — there is no path where a change ships unattended.
The discipline of a Certified Master Architect, applied to every build: correct scoping and naming, scan-clean code, proper ACL coverage, and ATF where it applies — with a signed design document alongside each change, every claim cited to the record.
No — she sits alongside. Stories mirror into Jira or ServiceNow Agile 2.0 with two-way sync.
Adoption doesn’t depend on ripping out your existing stack.
All of them — ITSM, CSM, HRSD, ITOM, ITAM, SecOps, IRM, SPM, App Engine, Now Assist and the rest of the platform.
Every product has its own use case describing what she analyzes, plans, builds and maintains there — [browse the use cases](https://phyllis.app/#use-cases) for the full list.
She signs in as you. Per-user OAuth means she can only see and touch what your roles and ACLs already allow — nothing more.
No plugins to install, no shared super-admin account — she works entirely under your existing security model.
Nowhere. Zero data retention is the default on every tenant.
Nothing is kept to train a model, and findings live inside your instance.
The audit trail is written per operator, per action.
The full posture — encryption, authentication, sub-processors — lives on the [security page](https://phyllis.app/security).
The real records — catalog items, Flow Designer flows, business rules, client scripts, UI policies, ACLs and AI Agents — wired together, scoped correctly, and delivered in connected update sets. The catalog item, the flow behind it, the approvals and the ACLs land as one build, not fragments — across every ServiceNow product.