Independent AI workflow starter pack

Turn scattered AI prompts into a governed way of working.

A practical collection of skills, projects and specialist agents for presales, sales and customer success. It helps teams produce faster, more consistent work without losing evidence, judgement or human control.

Explore all 33 workflowsUnderstand the system

An independent starter pack, not an official Basware product.

33portable workflows
36specialist agents
7evidence states
1shared operating model

What it is

One operating model.
Three ways to use it.

Basware AI Operating System is a documentation-first collection of reusable AI workflows. The same core instructions can be loaded as a portable Claude web skill, used as a Claude Project workspace or run through specialist Claude Code agents.

It is designed to turn recurring commercial and customer work into repeatable systems while preserving professional judgement. Today it is a validated starter pack of prompts, contracts, examples and tests. It is not a deployed enterprise application or an approved product-claims library.

Who it is for

Built around the work teams already do.

01

Presales & value consulting

Qualify work, design discovery, shape value cases and build evidence-safe narratives.

02

Sales teams

Prepare engagements, coach objections, improve deal strategy and turn meetings into action.

03

Customer success

Diagnose adoption, tell usage stories, prepare QBRs and track value realisation.

04

Enablement & governance

Create reusable learning, control evidence quality and review high-risk outputs.

Why it is valuable

AI speed, with an operating system around it.

Useful AI work depends on context, evidence and review. The collection makes all three explicit.

Move faster

Start with a complete workflow, not a blank chat. Each skill defines the inputs, decisions and output shape.

Stay grounded

Claims are separated into verified facts, inferences, assumptions, hypotheses and missing evidence.

Work consistently

Every surface follows the same workflow contract, so quality does not depend on who remembers the best prompt.

Keep people in control

Consequential claims and actions stay behind explicit evidence checks and human approval.

How it works

A simple path from request to trusted output.

The orchestrator owns routing. Specialists own bounded analysis. The evidence layer controls factual permissions. Humans own consequential approvals.

01

Ask

A user brings a real task, source material and the intended audience.

02

Route

The orchestrator selects the smallest useful specialist workflow.

03

Ground

Evidence is classified, gaps are exposed and permissions are checked.

04

Create

The specialist produces a decision-ready artefact in a consistent format.

05

Review

High-risk content is challenged before a human approves external use.

06

Learn

Reusable decisions and lessons improve the next engagement.

Evidence by design

Every claim carries a status.

The language is deliberately plain so reviewers can see what is known, reasoned, assumed or still missing.

VERIFIEDSupported by an identified permitted source
INFERREDReasoned from verified facts
ASSUMEDAdopted without verification
HYPOTHESISA proposition to test
MISSINGRequired information is unavailable
CONFLICTEDCredible sources disagree
REVIEW REQUIREDA designated human must decide

Complete feature catalogue

33 workflows, ready to adapt.

Each workflow includes a purpose, required inputs, structured process, output contract, evidence rules and review boundary.

Presales & value consultingFrom intake and discovery to RFPs, demos and value cases13 workflows
  1. Value Consultant OS
  2. Deal Qualification & Presales Triage Coach
  3. Account & Deal Strategy Coach
  4. AE Meeting-to-Engagement Generator
  5. Discovery Architect
  6. Value Hypothesis Builder
  7. Value Case Builder
  8. RFP Buddy Plus
  9. Demo Storyline Builder
  10. Proposal & Executive Narrative Coach
  11. Competitive Deal Coach
  12. IT / Security / Integration Readiness Coach
  13. Deal Brief Generator
Customer successFrom adoption diagnosis to QBRs, plans and value realisation8 workflows
  1. Customer Success Value Realisation Coach
  2. Adoption Diagnosis
  3. Usage Report Storyteller
  4. Living Agenda Coach
  5. QBR Narrative Builder
  6. Customer Health & Expansion Signal Coach
  7. Customer Improvement Plan Builder
  8. Customer Success Cadence Architect
Horizontal, sales & governanceMeeting intelligence, evidence, handoffs, enablement and control12 workflows
  1. Basware Meeting Intelligence
  2. APAC GTM Intelligence
  3. Sales-to-CS Handoff Builder
  4. Presales Intake & Capacity Control Tower
  5. Objection Coach
  6. Proof & Evidence Desk
  7. Output Governance & Red-Team Reviewer
  8. CRM Hygiene Assistant
  9. Weekly GTM Brief & Target Account Signals
  10. Offshore Handoff Translator
  11. Enablement & Learning Pack Generator
  12. Value Realisation Tracker

Ready from day one

Start at the level of tooling you have.

01

Portable web skills

Upload a packaged skill directly into Claude web chat and use it in any conversation.

02

Project workspaces

Set up a no-code Claude Project with starter prompts, knowledge templates and working conventions.

03

Specialist agents

Use the same contracts with a shallow, governed Claude Code agent workflow when deeper orchestration is appropriate.

Available now

A complete starter system.

  • 33 canonical workflow contracts
  • 33 portable Claude web skill packages
  • 36 specialist Claude Code agents
  • Project workspaces and starter prompts
  • Synthetic examples and blank templates
  • Deterministic packaging and validation tests

Not yet included

Future platform capabilities.

  • Production data connectors
  • A database or canonical record service
  • A deployed multi-user application
  • Autonomous writes to external systems
  • An approved product-claims library
  • Production monitoring and access controls

Proposed roadmap

Prove the workflow before building the platform.

This sequence is a direction for evaluation, not a product or delivery commitment. Each phase should earn the next.

01
Now

Evaluate & calibrate

Test the workflows against synthetic and approved internal scenarios. Tune quality, risk and usability.

02
Next

Build the evidence foundation

Align terminology, evidence sources, claim ownership and approval rules.

03
Pilot

Personal read-only use

Use the strongest workflows with approved context and no autonomous external writes.

04
Team

Governed small-team pilot

Add shared evaluation, review ownership, auditability and measured adoption.

05
Platform

Choose the runtime

Decide whether a multi-user application is justified by the evidence from pilots.

06
Scale

Controlled expansion

Add approved integrations, observability and change management only after governance proves durable.

The idea in one line

Reusable AI workflows that know what they know, show what they do not, and keep people accountable.

Back to the top