Reference¶
Look-up facts for the AI PDLC landscape: the flagship Freshworks figures, the GitHub Spec Kit and AWS Kiro fact sheets (versions, commands, spec files, pricing), the EARS requirement patterns, and a dated timeline of the org mandates. For the reasoning behind the pattern see Explanation. For tasks see How-to Guides.
Evidence labels used on this page
v = vendor-reported (single recorded session plus one live demo, no independent benchmark). x = externally verified against filings or press. m = long-standing marketing claim. d = taken from the product's own docs, repo or changelog.
Reported Outcomes And Benchmarks¶
| Metric | Before | After | Type | Conditions |
|---|---|---|---|---|
| Release cadence | Every 6 months | Every 2 weeks | v | Whole-PDLC change under one CPO tenure (Dec 2024 - Jul 2026) |
| Staffing ratio | 1 PM + 1 designer : 10-20 engineers | 1 PM : 1 engineer. Some teams no dedicated designer | v | Org of 200+ titled PM/designer/researcher staff |
| Agent round-trip latency | n/a | <= 10-15 s per step | v | Live demo on Grok models in Cursor |
| Evidence sweep in demo PRD | n/a | 4,358 active ITSM accounts aggregated | v | EX Agent Studio performance dashboard feature |
| Share of code generated with AI tools | n/a | "More than half" | v | CEO Dennis Woodside, May 2026 restructuring announcement (Benzinga) |
| FY2026 revenue guidance | ~$960M at recording | $963.5M-$966.5M (~15% YoY) | x | Raised with Q2 2026 results; Q2 revenue $237.4M, +16% YoY (Q2 2026 release) |
| Customer/end-user scale | — | 75,000+ customers, 300M end users | m | Long-standing company marketing figure |
Provenance: the episode writeup and live demo. Verification anchors: Q1 2026 results, Q2 2026 results.
Freshworks Timeline¶
| Date | Event | Type |
|---|---|---|
| 2024-11 | ~660-role (~13%) workforce reduction announced | x |
| 2024-12-02 | Srinivasan (Srini) Raghavan appointed Chief Product Officer (release) | x |
| 2025-12-31 | ~4,500 full-time employees | x |
| 2026-05 (Q1 results) | ~500-role (~11%) restructuring, ~$8M one-time charges; CEO cites AI code generation | x |
| 2026-07-28 | Ryan Manning (ex-CPO, BMC Helix) named Chief Product and Technology Officer, starting 2026-08-10 and reporting to CEO Dennis Woodside; Raghavan departs "to pursue an entrepreneurial venture"; CTO Murali Swaminathan reports to Manning (Nasdaq release, Inc42) | x |
| 2026 (mid-year) | Share price down roughly a third year-to-date (TIKR reported -32%; exact figure depends on the date) | x |
GitHub Spec Kit Fact Sheet¶
| Fact | Value | Source |
|---|---|---|
| Latest version | 1.0.11 (2026-09-24) | d — CHANGELOG, PyPI specify-cli |
| 1.0.0 release | 2026-08-21 (first anniversary) | d — History |
| First commit / first release | 2025-08-21 / 0.0.1 on 2025-08-22 | d |
| Release cadence | Patch releases every few days (1.0.0 to 1.0.11 in ~5 weeks) | d |
| License | MIT (Copyright GitHub, Inc.) | d |
| Install requirements | Python 3.11+, uv, a supported coding agent; Linux, macOS, Windows |
d |
| Package | specify-cli (CLI binary specify), published to PyPI |
d |
| Stewardship | Founded by Den Delimarsky and John Lam (Aug 2025 - Jan 2026); Manfred Riem lead maintainer since 2026-01-22 | d |
| Scale at 1.0.0 | 38 coding-agent integrations, 157 community extensions, 33 presets, 270+ contributors | d |
| Integrations table (2026-09) | 40 named agents plus generic (bring-your-own command directory) |
d — integrations |
| First-party processes | SDD (core), bug extension (0.9.5, 2026-06-05), assess extension (0.13.0, 2026-07-17) |
d |
| Composable primitives | Integrations, extensions, presets, workflows, workflow steps; bundles package them per role | d |
Spec Kit SDD Commands¶
Canonical order from the command reference: constitution -> specify -> clarify -> plan -> checklist -> tasks -> analyze -> implement -> converge. Only specify is strictly required before plan; clarify, checklist and analyze are optional quality gates.
| Command (dotted form) | Agent skill | Purpose | Writes |
|---|---|---|---|
/speckit.constitution |
speckit-constitution |
Establish or update project principles | Constitution, synced templates |
/speckit.specify |
speckit-specify |
Requirements and user stories (what and why) | spec.md, checklists/requirements.md |
/speckit.clarify |
speckit-clarify |
Up to five targeted questions, answers encoded into the spec (formerly /quizme) |
spec.md |
/speckit.plan |
speckit-plan |
Technical plan: stack, architecture, constraints | plan.md and design artifacts |
/speckit.checklist |
speckit-checklist |
Requirements-quality checklists ("unit tests for your requirements") | Custom checklists (reviewer-owned) |
/speckit.tasks |
speckit-tasks |
Dependency-ordered tasks: Setup, Foundational, one phase per user story, Polish | tasks.md |
/speckit.analyze |
speckit-analyze |
Read-only cross-artifact consistency check of spec, plan and tasks | Report only |
/speckit.implement |
speckit-implement |
Execute tasks; reads checklist state as a gate | Code |
/speckit.converge |
speckit-converge |
Append-only check of code vs artifacts; appends missing tasks or reports Converged | tasks.md (append only) |
/speckit.taskstoissues |
speckit-taskstoissues |
Optional: convert tasks to GitHub issues (needs GitHub MCP tools) | GitHub issues |
Spec Kit Invocation Forms¶
| Integration or mode | SDD example | Extension example |
|---|---|---|
| GitHub Copilot default skills mode (also most skills-based agents) | /speckit-specify |
/speckit-bug-assess |
| Dotted notation used in the reference docs | /speckit.specify |
/speckit.bug.assess |
| Codex, Command Code, ZCode skills | $speckit-specify |
$speckit-bug-assess |
| Kimi skills | /skill:speckit-specify |
/skill:speckit-bug-assess |
Selected integration keys: claude (skills in .claude/skills), copilot (skills in .github/skills/ by default), codex, cursor-agent, gemini, kiro-cli (alias kiro), grok, opencode, qwen, zed, generic.
AWS Kiro Fact Sheet¶
| Fact | Value | Source |
|---|---|---|
| Vendor / license | AWS; proprietary IDE built on Code OSS | Introducing Kiro |
| Public preview | 2025-07-14 | Introducing Kiro; Forbes |
| General availability | 2025-11-17: Kiro CLI, property-based testing for spec correctness, checkpoints/rewind, multi-root workspaces, team plans via AWS IAM Identity Center | Kiro GA blog, SiliconANGLE |
| IDE 1.0 | 2026-08 (approx.): capability-based permissions, custom agents, experimental Agent Focus view, natural-language hook creation | IDE 1.0 changelog |
| Latest IDE version | 1.1 (2026-09-14): durable agent artifacts, native ARM64 builds for Windows and Linux, Code OSS 1.131 | IDE 1.1 changelog |
| Spec files | requirements.md (user stories + EARS acceptance criteria), design.md (architecture), tasks.md (trackable tasks) |
Specs docs |
| Spec types | Feature specs (Requirements-First or Design-First), Bugfix specs (root cause, what changes, what must not change), Quick spec / Quick Plan (no approval gates between phases) | Feature specs, Bugfix specs, Quick Plan |
| Hooks format | Structured v1 JSON in .kiro/hooks/ (IDE 1.0) |
IDE 1.0 changelog |
Kiro Pricing (2026-09)¶
| Plan | Price (per user/month) | Credits included |
|---|---|---|
| Free | $0 | 50 |
| Pro | $20 | 1,000 |
| Pro+ | $40 | 2,000 |
| Pro Max | $100 | 5,000 |
| Power | $200 | 10,000 |
| Enterprise | Custom | Custom |
Extra credits beyond the plan cost $0.04 each (in-plan rate works out to $0.02). Sources: kiro.dev/pricing, Introducing Kiro Pro Max, cross-checked with Morph's September 2026 summary. The preview-era plan (Pro $19 with 1,000 interactions, Pro+ $39 with 3,000) was replaced by credit pricing (new pricing plans and Auto).
Pricing moves often
Morph reports the 500-credit trial was removed on 2026-09-14 and that first-time upgraders get a $20 credit. Treat this as a secondary-source detail and re-check the pricing page before budgeting.
EARS Requirement Patterns¶
EARS (Easy Approach to Requirements Syntax) comes from Mavin et al. at Rolls-Royce (IEEE RE 2009). Kiro's docs use the event-driven form WHEN [condition/event] THE SYSTEM SHALL [expected behavior]. The other patterns below are the standard EARS set.
| Pattern | Template | Example (from the Freshworks-style dashboard feature) |
|---|---|---|
| Ubiquitous | THE SYSTEM SHALL |
THE SYSTEM SHALL show account counts per product line |
| Event-driven | WHEN |
WHEN weekly usage drops 20% THE SYSTEM SHALL list the affected accounts |
| State-driven | WHILE |
WHILE a data refresh is running THE SYSTEM SHALL show the last refresh time |
| Optional feature | WHERE |
WHERE drill-down is enabled THE SYSTEM SHALL link each account to its tickets |
| Unwanted behaviour | IF |
IF the warehouse query times out THEN THE SYSTEM SHALL show a retry message |
Org Mandates Timeline¶
| Date | Organisation | Event | Source |
|---|---|---|---|
| 2024-01-08 | Duolingo | ~10% of contractors cut as generative AI takes over content work | Bloomberg, TechCrunch |
| 2025-04-07 | Shopify | Tobi Lütke publishes memo on X: "Reflexive AI usage is now a baseline expectation"; teams must show why AI cannot do the work before asking for headcount; AI usage enters performance and peer review | CNBC |
| 2025-04 (late) | Duolingo | Luis von Ahn "AI-first" memo: gradually stop using contractors for work AI can handle; headcount only when a team cannot automate more | The Verge |
| 2025-04-30 | Duolingo | 148 new AI-assisted courses launched, more than doubling the catalogue | Duolingo press release (2025-04-30), TechCrunch |
| 2025-05 | Duolingo | Partial walk-back: von Ahn on LinkedIn: "To be clear: I do not see AI as replacing what our employees do (we are in fact continuing to hire at the same speed as before)" | Fortune, PR Daily |