Where a single flow fires on one trigger and runs a chain, a process orchestrates many flows across days with human approvals, parallel paths, SLA timers, and sub-process reuse. Think "customer onboarding" with eleven handoffs, not "assign the lead" in one step.
A single Strkr Flow is enough for lead routing, Monday hygiene, or renewal flagging. A process is the right primitive when the motion crosses three or more teams, pauses on human decisions, and runs across business days. Customer onboarding, order fulfillment, employee offboarding, incident response: these are processes with branches, pauses, and reconvergence points that a single flow would strain to model.
Swimlanes
One lane per owning team.
Process canvases organize nodes into horizontal swimlanes, one per owning team or role. A customer onboarding process has lanes for Sales, Success, Implementation, Billing, and the customer. A hand-off from one lane to another makes the ownership transition visible, which is the exact information that gets lost when a process lives in a 44-row spreadsheet instead.
Phases
Named stages with entry and exit rules.
A process has named phases: Kickoff, Setup, Live, Review, Close. Each phase has entry and exit conditions configured as rule sets. The process can only enter Setup when Kickoff is marked complete and the ICP form is signed. Exit from Live requires success-plan sign-off. The phase model gives management a dashboard view of where each live process sits.
SLA timers
Clock on every handoff.
Each handoff and each phase carries a configurable SLA. The clock starts when the handoff lands in the receiving lane and stops when the receiving team closes the handoff task. SLA breach escalates according to the configured policy: notify manager, re-assign, raise priority. The dashboard shows live SLAs with time-remaining color coding.
Parallel paths
Two lanes work at once, then rejoin.
A process node can fan out to parallel paths that run independently in their own lanes. Classic use: in onboarding, Implementation configures the production tenant while Success writes the kickoff deck. Both complete before the Kickoff phase ends. The join node waits for all parallel paths to finish, which keeps the process stuck in Kickoff only until the real blocker clears.
Human approval
Pauses with named approvers and timeouts.
Approval nodes pause the process and wait for a configured approver. Timeout rules escalate: 24-hour inaction notifies the manager, 48-hour inaction re-routes to a backup. The approval captures the approver identity, the timestamp, the comment, and files to the process audit log. Approvers see the context (the record state, the prior steps) without opening a separate tool.
Sub-processes
Reusable chunks called from many parents.
A sub-process is a named process shape callable from a parent at a node. Classic reuse: the DPA signature sub-process is called from onboarding, from renewal, and from account expansion. One canonical shape, three parents. Edits to the sub-process propagate; the parent processes do not duplicate the DPA logic three times.
Process variables
Named state scoped to the run.
A process defines variables scoped to the run: the ARR value, the implementation complexity rating, the customer success tier. Nodes read and write variables. Variables persist across days and nodes, which lets a decision made in Kickoff (Tier 3 implementation) drive a routing decision in Setup three weeks later without re-querying the record.
Compensating actions
Roll back cleanly on failure.
If a process fails at step 7 of 12 and needs to abort, compensating actions roll back the earlier side effects: cancel the provisioned tenant, delete the kickoff deck, revoke the Slack channel. Compensation is explicit per step, not inferred. Teams that need enterprise rollback semantics get them without writing custom cleanup scripts after each failure.
Where processes live
Six orchestrations that touch five departments.
Six repeatable processes cover most of what mid-market B2B companies orchestrate across departments. The ones below each span three to seven teams and run across one to twelve weeks. Each entry names the phases and the typical SLA budget per phase.
Customer onboarding
Signed deal to first value.
Phases: Kickoff (3 days), Setup (10 days), Launch (5 days), Hypercare (14 days), Steady State. Lanes: Sales, Success, Implementation, Billing. Measured outcome across nine customers: median time-to-first-value fell from 38 days to 19 days after the orchestration shipped. The savings were in the handoffs, not the work.
Employee onboarding
Signed offer to productive week two.
Phases: Pre-hire (7 days), Day One, Week One, Week Two review. Lanes: Hiring Manager, IT, People Ops, Finance, the new hire themselves. Measured impact: three customers dropped new-hire first-day-unblock rate from 61 percent to 97 percent after moving to the orchestration. IT and People Ops stopped missing each other.
Order fulfillment
Signed quote to delivered product.
Phases: Contract Clear, Provision, QA, Deliver, Invoice. Lanes: Sales, Fulfillment, QA, Finance. Measured: fulfillment lead time fell 32 percent across the sample and billing holds (missed invoice) went from 7 percent to zero. The orchestration moved billing from a trailing step to a parallel path.
Incident response
Alert to postmortem published.
Phases: Detect, Triage, Resolve, Communicate, Postmortem. Lanes: On-call, Engineering, Customer Success, Comms. SLA timers on Triage (15 min), Resolve (per severity), Communicate (30 min), Postmortem (5 business days). The process surfaces at-risk SLAs on the status dashboard, which keeps the right people awake at the right times.
Renewal motion
Sixty days out to signed addendum.
Phases: Risk Review, Expansion Discovery, Proposal, Negotiation, Close. Lanes: Success, Sales, Finance. Measured: renewal close rate rose from 83 percent to 91 percent after four quarters on the orchestration. The structured risk review caught churn risks that previously surfaced two weeks before renewal, when nothing could be done.
Procurement review
Vendor request to signed contract.
Phases: Request, Security Review, Legal Review, Signature. Lanes: Requestor, Procurement, Security, Legal. Parallel paths on Security and Legal reviews cut cycle time in half for standard vendors. SLA timers escalate stalled reviews. The process writes to the vendor record and the procurement audit trail.
The reliability model
What happens when a step fails on day three.
Processes that run across days have more failure modes than single-trigger flows. A step that passes the sandbox can still fail live because of a change in a connected record or a transient integration outage. The reliability primitives below exist to catch these failures without the operator holding a watchlist.
Phase SLAs
Breach surfaces to the owning manager.
Each phase carries an SLA budget. A breach notifies the owning manager, raises the process to high priority on the dashboard, and starts the configured escalation chain. The alert is not a Slack message that gets lost; it is a visible color-coded entry on the dashboard the manager opens every morning.
Dead-letter per step
Failed steps go to a reviewable queue.
A failed action within a process writes to the process dead-letter queue scoped to that process type. The ops lead opens the queue, inspects the failure, applies a fix (bad field, dropped integration, invalid webhook target), and replays. Replay picks up at the failed step without re-running the earlier successful steps.
Idempotency
No double-ship on retry across days.
A step that sends an email, creates a Jira ticket, or provisions a tenant carries an idempotency key derived from the process run id and step index. Retry during a transient outage never double-ships. The engine does exactly-once semantics so the operator does not reason about it after a day-three failure.
Diff on edit
Schema changes do not break live runs.
If the process shape changes while runs are in flight, each run completes against the shape it started on. The next new run picks up the new shape. The ops lead can also opt to migrate live runs to the new shape explicitly, with a diff view showing which steps change. No silent drift mid-run.
Pause on human
The process never ghosts an approval.
Approval timeouts escalate per the configured policy; a stuck approval never silently hangs. The dashboard shows pending approvals, aging buckets, and the approver names. Managers run a weekly "stuck approvals" review from one surface instead of chasing twelve Slack threads.
Audit chain
One ledger for the whole run.
Every node, every variable update, every retry, every approval, every escalation writes to the process audit log with timestamps and user identities. Audits that cross three teams and twelve days land in one ledger, not three. SOC 2 Type II (in progress 2026-Q4) activity logging reads from this ledger.
What this replaces
The three tools a process orchestration layer eats.
Teams running mid-market B2B processes typically stitch together a project tool, a workflow tool, and a BPM or Kissflow-style orchestrator. Strkr Processes replace the orchestration layer and share state with the project and workflow surfaces. The consolidation savings land in both the subscription bill and the operator time.
BPM suites
Enterprise BPM at mid-market price.
Pega, Appian, Camunda sit at the enterprise end with multi-six-figure licenses and professional-services implementations. Strkr Processes cover the orchestration primitives that mid-market teams need (swimlanes, phases, approvals, SLAs, sub-processes, compensating actions) inside the per-seat subscription. The savings are the implementation line as much as the license line.
Project tool
Not a replacement for issue tracking.
Strkr Projects covers issues, sprints, backlogs, and roadmaps. Strkr Processes orchestrates handoffs between teams across days. They share state and link bidirectionally, so a process node can create an issue in the delivery team and wait on completion. The two surfaces complement rather than compete; teams run both.
Workflow tool
Flows are nodes; processes are the parent.
A Strkr Flow runs a single trigger and a chain. A Process orchestrates many flows across time with human pauses and SLA timers. Internally, process nodes call flows as sub-process atoms. The two abstractions nest cleanly, which keeps the mental model consistent as the operator moves between them.
Not Zapier
Middleware does not scale to multi-week processes.
Zapier and Make are optimized for single-trigger task chains. A multi-week process with human approvals, SLA timers, and parallel paths is outside the design envelope. Strkr Processes runs the orchestration natively. We do not recommend a middleware tier for jobs that cross teams across days.
HIPAA posture
BAA covers process audit and PHI handling.
Healthcare teams running patient-intake processes sign under the BAA with per-tenant encryption on the process state and the full audit chain. PHI fields on the process variables inherit the tenant encryption key policy. SOC 2 Type II attestation is in progress for 2026-Q4.
Open API
Vertical workflows integrate as code.
A vertical SaaS tool (healthcare EMR, legal practice management, construction PM) wires into a Strkr Process via native OAuth plus webhook. The process surfaces trigger events (phase entered, approval granted, SLA breached) on the webhook contract. Partners build in the record, not inside a middleware console.
Five processes to ship first
A sequencing plan for a 60-day deployment.
Teams that get fastest value from process orchestration sequence their rollout in a order: pick the single highest-stakes process that already has a working spreadsheet shape, orchestrate it first, then expand. The five below are the sequence we see work most often across the ninety-one deployments on record.
Order one
Customer onboarding first.
The onboarding process has the clearest ROI measurement (time-to-first-value) and the clearest cross-team pain. Teams see the handoff savings in the first live run. We recommend starting here across every mid-market team because the win is visible to the exec sponsor by week two.
Order two
Renewal motion second.
The renewal process shares records (Account, Opportunity) with onboarding and reuses several sub-processes (DPA signature, legal review). Starting second lets the sub-process library compound rather than start from scratch. Expect three to four weeks of effort for a sixty-percent faster renewal cycle.
Order three
Order fulfillment third.
Fulfillment uses the same approval primitives as the first two but adds billing integration. The teams that run onboarding and renewal first have the approval policies dialed, which shortens the fulfillment configuration time. Measured outcome: a 32 percent lead-time cut across the fulfillment sample.
Order four
Employee onboarding fourth.
HR and IT processes land fourth because the records (Users, Access, Equipment) are different from the customer-side ones. The orchestration primitives are the same, but the ops lead learns the HR data model after the customer processes are live. The savings are the first-day-unblock rate climbing to the high 90s.
Order five
Incident response fifth.
Incident response is the most time-sensitive process (SLA timers in minutes, not days). Starting fifth means the SLA primitives are proven at the longer cadences first. By the time incident response goes live, the dashboard shows mature SLA color coding and the escalation chains are tested at scale.
Walk through a live onboarding process from kickoff to steady state.
Bring your current onboarding playbook. We will orchestrate the first phase on screen and show the handoff math.
A project is a bounded delivery effort with a scope, a timeline, and a set of issues. A process is a repeatable orchestration shape that runs over and over against different records. Customer onboarding is a process: the same shape runs for every new customer, with the instance being a run of the process against that customer record. Project delivery for one specific customer is a project. Processes often create projects as sub-steps; projects do not run processes.
How many steps can a single process support?
The engine supports processes with up to 240 nodes in a single canvas. Beyond that count, the right pattern is sub-process decomposition: factor repeated chunks into named sub-processes called from the parent. Most real-world processes land between 20 and 60 top-level nodes with 2 to 8 sub-processes. Nothing in the engine encourages the ten-page spaghetti canvas that drives BPM implementation disasters.
Can process variables reference custom fields on linked records?
Yes. Process variables bind to any field on the triggering record and any field on linked records up to three hops. Custom fields are first-class. A process variable can bind to "the owner of the parent account of the primary contact on this deal" with the full linked chain. The variable captures the value at bind time and updates on each referenced record change if the variable is configured to track live.
What happens if we change a process shape while 300 runs are in flight?
Each live run completes against the shape it started on. The 300 in-flight runs continue to the shape they knew. New runs started after the shape change use the new shape. The ops lead can optionally migrate in-flight runs with a diff view showing exactly which nodes change and which run states need interpretation under the new shape. Explicit migration prevents silent drift; the engine never rewrites a running process without permission.
How do SLA timers interact with business hours and holidays?
SLA timers can be configured as calendar time (24-hour clock) or business time (business hours plus the tenant work calendar). Business-time SLAs pause overnight, on weekends, and on configured holidays. The work calendar supports per-user and per-team overrides for the processes that span regions with different holidays. Each SLA breach surfaces the breach time in both calendar and business formats, so the manager sees the business reason not just the clock reason.
How does this handle regulated-industry processes like HIPAA intake?
Healthcare teams run patient-intake processes under the BAA with per-tenant encryption on process state. PHI fields on process variables inherit the tenant encryption key. The process audit log captures access events in the format HIPAA activity logs require. Signed BAA from the Strkr side; the operator keeps the compliance workflow inside the record instead of exporting to a separate audit tool. SOC 2 Type II attestation is in progress for 2026-Q4.
No sales call, no migration consultant, no four-month implementation. Enter your card, get 14 days of the full Pro tier, cancel any time before day 14 with zero charge. Spin up a workspace, import your CSV, and have something useful before lunch.
We use cookies.
Essential cookies keep the site working. If you accept, we also enable Google Analytics
so we can see which pages help and which don't. Reject to opt out entirely. Details in our
Privacy Policy.
