Live sprint tracker - Sprint 3 kickoff August 5
Day 1 of 10 - August 5 → August 18, 2026 - Sprint 3 Kickoff · View Sprint 2 Review →
| Date | Event | Details |
|---|---|---|
| Aug 4 (Mon) | Program Session | Turbo Mode deep dive with Zun, Matthew, Adelina, Nathan. Swimlane requirements reviewed. Teams federation unblocked. |
| Aug 5 (Wed) | Sprint 3 Kickoff | Sprint planning with Joana. |
| Aug 10 (Mon) | Portfolio Meeting with Krishna + Kush | SOC for AI + Telemetry, SOAR/Swimlane EBITDA case, Turbo Mode preview. |
| Aug (mid) | Turbo Mode SaaS Alpha Access | Deloitte team (Zun, Matthew, Adelina) gets SaaS alpha access to evaluate Turbo Mode as Swimlane replacement. POV testing with MXDR test range + Graph API email ingestion. |
| Aug 18 (Mon) | Sprint 3 End + Review | Show-don't-tell evidence. |
| Sep (target) | Turbo Mode Self-Managed (SMK) Release | Self-managed/deployed version with Turbo Mode for Deloitte's on-prem environment. Tony confirmed September target. Enables testing against Jira behind firewall + full Swimlane workflow migration. |
| Aug 31 (Mon) | Charlie vacation starts | Returns Sep 23. |
| Nov (end) | Swimlane Decision Deadline | 90-day notice before Feb renewal. Deloitte must decide by ~Nov. |
Four-layer cadence: 2-week sprints, monthly co-design, monthly portfolio, quarterly deep dives
Execution cadence. Ship code, build agents, deliver results.
| Planning | Monday (every 2 weeks) |
| Standups | Mon/Wed/Fri (team only) |
| Review + Retro | Friday (every 2 weeks) |
| Tony attends | Planning + Review only |
Show results to executives. Video evidence + working links.
| First meeting | Aug 10 (confirmed) |
| Attendees | Tony, Krishna, Adelina, Nathan, Charlie, Joana + Kush |
| Format | Show, don't tell (see below) |
| Note | Portfolio/Design layer — monthly cadence |
In-person design sessions. Strategic alignment + release planning.
| Duration | ~5-7 days |
| Q3 location | Houston, TX |
| With | Team + Ron + key engineers |
| Output | Quarter release plan |
Joint design sessions with Deloitte team. Quick status + 2-3 deep dives per session.
| Cadence | Monthly (schedule 1-2 ahead) |
| Format | Quick status items + 2-3 deep dives |
| T&C | Tony, Charlie, Joana, Victor |
| Deloitte | Kush, Krishna, Nathan, Zun/Adelina |
| First session | Jul 10 ✅ (90 min, virtual) |
| Next | ~August (TBD, waiting on Kush's agenda) |
Confirmed in the Jul 30 cadence email. Meetings organized in three layers by scope and attendees:
Tactical (weekly)
Delivery teams — Joana, Nathan, Zun, Matthew, Brandon/Marcos. Standup stays weekly (<5 tickets/week volume).
Program (biweekly)
Joana, Charlie, Nathan, Adelina, Zun/Matthew.
Portfolio/Design (monthly)
Tony, Krishna, Adelina, Nathan, Charlie, Joana. Next: Aug 10.
| Artifact | Cadence | Description | Status |
|---|---|---|---|
| Sprint Backlog | Updated daily | Items committed for current sprint, status, blockers | Sprint 2 |
| Burndown | Real-time | Story points or items remaining vs time | Sprint 2 |
| Sprint Review Deck | End of sprint | Video evidence + working links of shipped work | Sprint 2 |
| Definition of Ready | Standing | Criteria for items entering sprint (see below) | Defined |
| Definition of Done | Standing | Criteria for items being complete (see below) | Defined |
We're moving at 5× traditional speed. Words can't keep up. Ron didn't understand net new revenue agents until the third time Tony explained it - and Ron is the sharpest person in the room. If Ron needs three reps to absorb a verbal summary, everyone else needs more.
The old way (what we stop doing):
The new way (what every portfolio item needs):
| Month | Location | Notes |
|---|---|---|
| August | Houston | Rescheduling to Aug 6 (tentative) |
| August | San Francisco | Krishna's base |
| September | Austin | Tony's base |
| October | Los Angeles | Charlie's base |
Deep dives aligned with monthly portfolio locations to avoid over-travel (Q3 → Houston, Q4 → Los Angeles).
Current team, Deloitte counterparts, and Value First hiring plan
| Role | Count | Region | Focus | Status |
|---|---|---|---|---|
| AI PMO / Soft Skills | 1-2 | LatAm (preferred) | Requirements gathering, stakeholder mgmt, verification | Interviewing |
| Engineer | 1-2 | Eastern Europe or LatAm | MLflow, Kindo agent configuration, integrations | Planning |
Process: Invoice → Charlie → Ron. Charlie vets technical candidates. LatAm for soft-skills (live meetings), Eastern Europe for code (Charlie's preference).
Jul 8 discussion: Finding resources with AI expertise + complex environment + T&C speed is extremely difficult (Shanzo example - impressed by Kindo's AI use). Charlie's suggestion: hire for training capacity, not existing expertise. Training requires investment but broadens the candidate pool.
| Team | Status | Expected Back | Impact |
|---|---|---|---|
| Agent Runtime (Madison) | Returned ✅ | Returned | Back and active. Multi-agent, agent features unblocking. |
| Agent Runtime (Sean) | Returned ✅ | Returned week of Jul 7 | Returned. Ramped back up. |
| Core Kindo (Brian Van) | PTO | Mid-July (expected) | Core platform changes blocked - expected back during Sprint 2 |
| Charlie (Agent Runtime lead) | Active | - | Shipping memory, system prompts, trace export. High velocity. |
Sprint 2 implication: Madison and Sean returned ✅. Brian Van expected mid-July. More code work possible this sprint. Engineering-dependent items now feasible.
Agent portfolio - SOC for AI POC delivered, expanding across three pillars
| ID | Agent | Status | Revenue Class | Blocker |
|---|---|---|---|---|
| A.1 | Threat Monitoring | PROD | Contracted | - |
| A.2 | Threat Intel | PROD | Contracted | - |
| A.3 | Threat Hunt | PROD | Contracted | - |
| A.4 | Detection Engineering | PROD | Contracted | - |
| A.5 | CTEM | BUILT | Contracted | Deployment pending |
| A.6 | Vitals Dashboard | DEPRIORITIZED | Alliance | Kush shifted to SOC for AI |
| A.7 | Quality Audit Agent | CO-DESIGN | Alliance | Digital twin approach - L2/L3 first use case |
| A.8 | Cloud Security Agent | PLANNED | Alliance | - |
| A.9 | IR Agent | PLANNED | Alliance | - |
| A.10 | IoT/OT Monitor | DEPRIORITIZED | Alliance | Kush shifted to SOC for AI |
| A.11 | Custom Client Agents | REQS | Alliance | Shadow & document method |
| A.12 | Identity Agent → IdaaS | PLANNED | Alliance | Tim Corder engagement |
| A.13 | GRC Agent → GRC aaS | PLANNED | Alliance | Nathan Ellis engagement |
| NEW | SOC for AI | EXPANDING | Alliance | Krishna: “I can see us taking this to clients” |
| NEW | SOAR/Swimlane Replacement (Turbo Mode) | EVALUATING | Alliance | Alpha in Aug. Nov decision deadline. |
Charlie's framework (confirmed by Kush/Krishna Jul 10). Three pillars of AI SOC discovery. Sprint 2 focus: Pillar 1 deep dive (Anthropic + Azure Foundry). Pillars 2 & 3 moved to research - Pillar 2 gated on positioning decision, Pillar 3 requires Deloitte SME guidance.
Pillar 1 - Platform Integrations
Reach into Anthropic, Azure Foundry, Copilot, ServiceNow etc. API-level interrogation of AI platforms to audit agent usage, models, and configuration.
Pillar 2 - LLM Gateway / Choke Point
Kindo as inference proxy (2 lines of code to redirect). LLM is the firewall. Agents inside vs outside org walls. Kush's philosophy: the choke point IS the governance layer.
Pillar 3 - Network/Endpoint Monitoring
Sweep for shadow AI (long tail). Network and endpoint-level discovery of unauthorized AI usage beyond managed platforms.
Krishna's tiered model: Basic (detect + respond post-action) → Standard → Elite/Advanced (preventative blocking). First version: "less intrusive" - detection and response post-action.
All 5 governance objectives confirmed at Jul 10 co-design session. Kush, Krishna, Nathan asked detailed technical questions.
| # | Deloitte Objective | POC Status | Jul 10 Status |
|---|---|---|---|
| 1 | Unauthorized agent deployment | Covered | ✅ Confirmed |
| 2 | Unauthorized tool/data connections | Covered | ✅ Confirmed |
| 3 | Agent action drift vs SOPs | Gap | ✅ Confirmed - biggest eng need |
| 4 | Guardrail/policy changes | Partial | ✅ Confirmed |
| 5 | Multi-tenant data breach | Partial | ✅ Confirmed |
Strategic Position (Jul 10)
Kush/Krishna confirmed alignment with co-develop model. Tiered model (Basic/Standard/Elite) - CrowdStrike analogy. Menu of discovery options. Post-discovery actions platform-dependent, policy-driven (remediate, notify, or block). Need platform compatibility matrix.
Charlie's follow-up email to Deloitte team (Jul 10) - Kindo reaches governance outcomes on three existing open standards, no proprietary protocol adoption required.
Standard 1 - Inference APIs (Live Today)
Chat Completions, Responses, Messages, Models. 2 lines of config to route Claude Code / Copilot / agents through Kindo. Demo →
Standard 2 - Tools / Federated MCP Gateway (Live Today)
Single governed endpoint for tool/integration access. Federated MCP gateway provides real-time control over agent tool use.
Standard 3 - Telemetry / OpenTelemetry (Landing Now)
OTLP-based. Two directions: (a) customers route Kindo agent traces to their own backend (per-org isolated), (b) Kindo ingests OTel from third-party agents/apps (Claude Desktop, etc.). Foundation for LLM-as-judge, evals, self-improvement.
Tier Mapping via Open Standards
Quarterly in-person sessions + monthly co-design with Deloitte
These items were categorized under "design sessions" in Tony's meeting notes. Updated with Jul 10 co-design outcomes.
| # | Topic | Description | Owner |
|---|---|---|---|
| 1 | Net New Revenue Agents | Design + deploy agents that generate alliance revenue (Tier 2/3 packages) | Tony / Joana |
| 2 | Threat Remediation | Extend A.1-A.5 into automated remediation workflows | Charlie / Victor |
| 3 | Deloitte Roadmap (Azure/GCP) | Cloud platform alignment and multi-cloud strategy | Charlie |
| 4 | Institutional Knowledge / Memory | Skills, memory, compound learning flywheel. Don't equip Deloitte to build what we want to build (Charlie) | Charlie |
| 5 | AI Cyber Guard / Tower | Control plane co-development. Evan building on AWS: unify policy, enforcement, observability, guardrails | Charlie |
| 6 | Lifecycle Hooks | Generic lifecycle hooks - Kush says yes but NOT most important. Ship fast MVP, don't over-engineer. | Charlie |
| 7 | Workflow Acceleration | Accelerate deployment cycle (time-to-value for new Kindo customers) | Victor / Joana |
| 8 | SOC for AI | Three-pillar discovery, AI governance, platform compatibility matrix | Joana / Victor |
| 9 | Agent Discovery Choke Points | LLM gateway as firewall. Agent telemetry open standard. Kush's philosophy from Jul 10. | Charlie / Kush |
| 10 | AI Cyber Guard Convergence | Evan's AWS-based AI Cyber Guard (4 problem statements: unify policy, enforcement, observability, guardrails). Open standard for agent telemetry. | Charlie / Evan |
90 min, virtual. SOC for AI POC demo. Three-pillar framework confirmed. Digital twin pivoted to L2/L3 analysts. Tiered SOC model (Basic/Standard/Elite). Co-design cadence open (biweekly vs monthly TBD - needs Tony input). Teams integration Step 0 initiated. Triage agent 15→10 min, sub-agent split considered.
90 min, virtual. SOAR/Swimlane replacement moved from hypothetical to active evaluation. Charlie demoed Turbo Mode — Krishna and Adelina responded positively. Swimlane decision deadline confirmed: end of November (90-day notice before Feb renewal). Teams federation confirmed over Slack. Weekly standup stays weekly. Three-layer meeting cadence established (Tactical/Program/Portfolio).
Decision needed: Schedule dedicated SOC for AI deep dive with Deloitte before or after the Aug 10 portfolio meeting. From Jul 30 session — Krishna expressed excitement about SOC for AI + Telemetry convergence: "I can see us taking this to clients." Session should cover: SOC for AI expansion across three pillars, telemetry/observability integration with new Kindo infrastructure, and agent health monitoring (Adelina's ask).
Houston, TX. 5-7 day in-person session. Team + Ron + key engineers. Output: Q3 release plan, SOC for AI architecture, net new revenue agent designs.
Los Angeles, CA - aligned with October monthly portfolio (Charlie's base). Review Q3 results, plan Q4 releases, expand to service lines beyond D&RaaS (Identity aaS, GRC aaS).
Day 10/10 - Demo Day. Nine items delivered. SOC for AI POC built and tested - first run scored 3.8/10 with real findings. Cross-platform AI orchestration demonstrated. Co-develop story ready for Kush.
9 items delivered. SOC for AI POC built and demoed. Jul 10 co-design session held with Kush, Krishna, Nathan, Zun. All 5 governance objectives confirmed. Three-pillar discovery framework established. Digital twin pivoted to L2/L3 analysts. Co-design cadence open - Joana suggested biweekly (sprint-aligned), Krishna suggested monthly. Left open Jul 10. Needs Tony input. Next session ~2 weeks (Jul 25).
Jul 8 pivot: Charlie directed POC to demonstrate Kindo orchestrating across external AI platforms (Anthropic, OpenAI, Copilot), not just Kindo-internal governance. POC has two components: (1) Kindo-native agent governance monitoring, (2) cross-platform AI interrogation. Additionally, Charlie is shipping memory, agent-editable system prompts, and trace export - the foundational capabilities for institutional knowledge. Friday demo tells the co-develop story: SME digital twin → institutional knowledge → Kindo platform features.
| Item | Type | Owner | Effort | Dependencies / Risks |
|---|---|---|---|---|
| SOC for AI - Cross-Platform POC Build POC on Kindo SaaS demonstrating two capabilities: (1) Kindo-native governance agent (drift detection, policy monitoring), (2) Cross-platform AI interrogation - Kindo agent connects to Anthropic/OpenAI/Copilot to check agent usage, models, and configuration. Charlie's direction: prove Kindo can orchestrate SOC for AI across heterogeneous AI environments. Agent export/import available via Yash's portability feature. |
P0 | Victor + Warren | 2d | Built ✅ Triggered Agent on Kindo SaaS. First run: governance report covering all 5 objectives, scored 3.8/10 🔴. Real findings. Enterprise access resolved (Hannah). Demo ready. |
| SOC for AI - Audit Surface Map Map Kindo's current audit logs, telemetry, RBAC, DLP, MCP policies against Deloitte's 5 governance objectives. Deliver before Friday requirements call. |
P0 | Warren + Victor | 1d | Done ✅ |
| SOC for AI - Requirements Questionnaire → Krishna Generate SOC for AI questionnaire (same format as A.6). Send to Krishna who routes to right person. |
P0 | Tony + Joana + Warren | 0.5d | Done + Sent |
| SOC for AI - Detection Rule Library Define detection rules mapping each of Deloitte's 5 objectives to specific Kindo audit events and API endpoints. |
P1 | Warren + Victor | 2d | Carries to Sprint 2 Refining post Jul 10 feedback. |
Show-don't-tell: every criterion needs evidence, not a status update.
An item can enter the sprint when ALL of these are true:
| # | Criteria | Why |
|---|---|---|
| 1 | Clear outcome defined - what does "done" look like in business terms, not technical terms? | Victor's point: business value, not technical value |
| 2 | Owner assigned - single person accountable | No orphan items |
| 3 | Dependencies identified - blocked/unblocked explicitly tagged | Victor's framework: shoot where we're unblocked |
| 4 | Effort estimated - days, not points. Be honest. | Tony needs to know what to expect without asking |
| 5 | Classified soft-skill vs code - which work type? Determines who can execute. | ~80% soft-skills moves without Brian's team |
| 6 | Passes tenant filter (if integration) - tenant-scoped? data stays in tenant? SOC 2 II / BAA / DLP? | Tony's note #1: "Deloitte only" |
| 7 | Fits the sprint - total committed work ≤ team capacity | Don't overcommit then under-deliver |
| 8 | Acceptance criteria written - how will we verify it's done? | "Show don't tell" starts here |
An item is done when ALL of these are true:
| # | Criteria | Evidence Required |
|---|---|---|
| 1 | Acceptance criteria met - every criterion checked off with proof | Screenshots, video, or live URL |
| 2 | Business done, not just technical done - stakeholder can see and use it | Working link or deployed artifact |
| 3 | Evidence attached - "show don't tell" proof in the same message as the completion claim | Video walkthrough, screen recording, API response |
| 4 | Integrations pass tenant + compliance check | Tenant filter results documented |
| 5 | No open blockers or regressions | Verification report |
| 6 | Reviewed - at least one other team member has seen the output | Reviewer name + feedback |
| 7 | Documented so Warren can report status | Warren updates programmatically |
| 8 | Owner confirmed done | Explicit sign-off |
Sprint 2 complete. Agent telemetry prototype, platform matrix, and gateway architecture delivered. Discovery menu and OTel proposal in progress. Co-design items staged for Jul 30.
3 items delivered (Agent Telemetry prototype, Platform Compatibility Matrix, Gateway Architecture). 2 in progress. 4 awaiting co-design. Jul 30 design session held — SOAR/Swimlane replacement moved from hypothetical to active evaluation, Turbo Mode alpha committed for August, Teams federation decision confirmed. See Sprint 3 →
A working demo where Kindo discovers agents across Anthropic + Azure Foundry, takes a real action (notify non-compliance), and a CrowdStrike-style menu of 5–10 discovery methods Krishna can present to clients with “which approach fits your environment.” That's the show-don't-tell for the next co-design session.
Three deliverables:
Platform Matrix (Victor+Charlie, 2d)
├──→ Discovery Menu (Joana+Charlie, 2d) — BLOCKED until Matrix done
└──→ Pillar 1 Deep Dive (Victor+Warren, 3d) — runs parallel
└──→ Discover→Act Loop (Victor+Warren, 1d) — BLOCKED until Pillar 1 has Foundry working
Matrix is the bottleneck.
It feeds both the Menu and validates which platforms the Act Loop can target. Victor + Charlie need to start here Day 1.
Everything else (triage optimization, A7 scoping, Copilot/ServiceNow, logging) is off the critical path — nice to have in Sprint 2 but won't make or break the co-design demo.
Two external blockers that could slip the path:
| Item | Type | Owner | Effort | Risks / Dependencies | Status |
|---|---|---|---|---|---|
| SOC for AI - Pillar 1 Deep Dive (Platform Integrations + Foundry) Deepen POC on platform integrations pillar. Add Azure Foundry interrogation alongside existing Anthropic. Demonstrate discovery + act loop on at least one platform. Pillars 2 (gateway) and 3 (network/endpoint) moved to research - Pillar 3 is greenfield requiring Deloitte SME guidance, Pillar 2 gated on positioning decision. |
P0 | Victor + Warren | 3d | Foundry API access/credentials needed. Depends on Act Loop item for full demo story. | Prototype ✅ |
| Platform Compatibility Matrix Internal source-of-truth grid: per platform (Anthropic, Foundry, Copilot, ServiceNow, Bedrock...), what's discoverable and what actions the API supports (remediate/notify/block), and whether real-time or after-the-fact. Feeds the Discovery Menu of Options. Not a Deloitte deliverable - this is the evidence that backs the Menu. |
P0 | Victor + Charlie | 2d | Risk: Some platform APIs may have undocumented limitations. Requires hands-on probing, not just docs review. Blocker for Menu. | Complete ✅ |
| SOC for AI - Discovery Menu of Options (CrowdStrike Model) Krishna-facing deliverable: 5-10 discovery methods for finding agents across enterprise environments (platform API, gateway/choke point, network/endpoint sweep, central registry, etc.), each with pros/cons per company type and network disposition. Modeled on CrowdStrike's '5 ways to discover endpoints' approach. Derives from the Platform Compatibility Matrix - Matrix is the input, Menu is the synthesis. Low engineering lift, high strategic value. |
P0 | Joana + Charlie | 2d | Unblocked: Platform Compatibility Matrix complete. Tier mapping in progress. Charlie availability for technical validation. | In Progress |
| SOC for AI - Discover→Act Loop (Single Platform Demo) Krishna asked directly: 'once you discover an agent, what can you do?' Current POC demonstrates discovery only. This item adds a real remediation action to the POC on a single platform (e.g., notify agent author of non-compliance via Anthropic API). Proves the full loop: discover → assess → act. |
P0 | Victor + Warren | 1d | LLM judges in prototype address assess step. Full act loop deferred to co-design input on platform priority. | Co-Design |
| Triage Agent Optimization Work Session Zun provides generalized (sanitized) prompts. Charlie + runtime team analyze latency independently. Goal: sub-10 min. Consider splitting into basic triage + advanced investigation sub-agents. Kindo runtime team scope - requires agent runtime engineering. |
P1 | Charlie + Runtime Team | 2d | Requires work session with Zun. Ready to schedule when timing works for Deloitte team. | Awaiting Co-Design |
| A7 Quality Audit - Design & Scoping Only Digital twin approach for weekly ticket audit - design phase only. L2/L3 analyst collaboration requires Teams integration (Step 0 + Step 1), currently blocked on Deloitte IT approval. Sprint 2 scope: define ABC grading criteria, map audit workflow, identify pilot candidates. 60% subjective scoring problem still unanswered - needs dedicated research. Full collaboration blocked on Teams. |
P1 | Joana + Victor | 1d | Depends on digital twin approach + design session with Krishna's team. Ready to co-design in Sprint 3. | Awaiting Co-Design |
| SOC for AI - Copilot & ServiceNow Interrogation Extend discovery to Copilot and ServiceNow platforms. Scope boundary: Pillar 1 Deep Dive covers Anthropic + Azure Foundry depth; this item covers Copilot + ServiceNow breadth. OpenAI removed from scope (not relevant to Deloitte's environment - theirs is Anthropic, Foundry, Bedrock, ServiceNow, Copilot). |
P1 | Victor | 2d | Covered in Platform Compatibility Matrix research. 5 platforms analyzed with capability mapping and license gates. | In Matrix ✅ |
| Logging/Observability Gap Analysis Nathan/Zun flagged agent failure logs missing for troubleshooting. Context compaction being fixed in upcoming release (Mo confirmed). Assess current gaps. Kindo runtime team scope - platform-level logging requires agent runtime engineering. |
P2 | Charlie + Runtime Team | 1d | Depends: Runtime team capacity (Madison/Sean ramping). Brian Van PTO (mid-July). May slip if higher-priority runtime work (triage optimization) takes precedence. | Sprint 2 |
| Pillar 2 - Gateway/Shim Scoping & Design Kush explicitly endorsed Kindo as gateway shim (Jul 16 email: “I would like to explore inserting Kindo as a shim to be the gateway to encore controls and gather telemetry”). Scope: architecture doc for Pillar 2 gateway positioning — inference proxy + MCP + telemetry. Deliverable for August co-design session. Gate unblocked by Kush’s endorsement. |
P1 | Charlie + Victor | 2d | Three open standards documented. Kush endorsed gateway approach. Architecture ready for collaborative refinement. | Complete ✅ |
| Agent Telemetry — Native Pillar Strategy + OTel/MLflow Architecture Charlie’s strategic proposal (Jul 20): adopt MLflow headless + multi-tenant inside Kindo as native agent analytics — traces, LLM judges, sampling/retention/compliance. Working prototype demonstrated. Strategic fork: host telemetry natively vs. export to external tools (Galileo, Braintrust, Arize, LangFuse, Datadog). Completing the Inference + Tools + Telemetry trifecta positions Kindo to manage entire AI deployments, not just native agents. Kush aligned on both trace types. Two open Kush questions remain: (1) hook into existing customer MLflow equivalents, (2) non-Kindo agents tracing directly vs through Kindo. Resourcing gate: major surface-area expansion comparable to Sandboxes — requires minimum one dedicated engineer (effectively permanent). Charlie recommends locking execution + resourcing plan before bringing to wider engineering team. |
P1 | Charlie | 2d | Working prototype demonstrated. Trace export, LLM judges, eval configuration, span detail, error diagnostics. Strategy + resourcing decision required before wider team rollout. | Prototype ✅ |
| Telemetry Storage/Retention Architecture Define the two levers Charlie described (retention duration + sampling fraction) as concrete, customer-facing controls. Document cost model so Deloitte can present storage as “a dial, not an open-ended liability.” Feeds into MLflow architecture doc. |
P2 | Charlie | 1d | Included in OTel/MLflow two-tier proposal. Validated in development environment. | In OTel Proposal |
Same criteria as Sprint 1 (see Sprint 1 tab for full details):
Same criteria as Sprint 1 (see Sprint 1 tab for full details):
Day 1 of 10. SOAR/Swimlane replacement evaluation, Turbo Mode SaaS alpha, Teams federation completion, and the Aug 10 portfolio presentation to Krishna + Kush.
Turbo Mode SaaS alpha access for Deloitte to evaluate as a Swimlane replacement, a Swimlane workflow import proving Turbo Mode covers the core detection + triage automation, and a portfolio-ready SOC for AI + Telemetry demo with an EBITDA case for the SOAR replacement. The Aug 10 portfolio meeting with Krishna + Kush is the show-don't-tell milestone.
| Item | Type | Owner | Effort | Status |
|---|---|---|---|---|
| Turbo Mode SaaS Alpha Access Provision Turbo Mode SaaS alpha access for Deloitte to evaluate. Charlie committed August timeline. Foundation for SOAR/Swimlane replacement evaluation. |
P0 | Charlie | 3w | In Progress |
| Swimlane Workflow JSON Import + POV Zun to share Swimlane JSON workflows. Charlie to import and build against them in Turbo Mode, proving coverage of the detection + triage requirements. Turbo Mode currently handles 5 of 10 requirements. |
P0 | Charlie + Zun | 2w | Waiting on Zun |
| SOC for AI + Telemetry Portfolio Presentation Portfolio-ready demo for Aug 10 meeting with Krishna + Kush. SOC for AI expanding, Krishna: “I can see us taking this to clients.” Convergence with agent telemetry work. |
P0 | Joana + Charlie | Aug 10 | Planning |
| Teams Federation Completion Deloitte side exclusion added (Adelina confirmed Aug 4). Complete T&C/Kindo side integration. Slack confirmed dead — Teams federation is the path. |
P1 | Joana + Ken | Aug | Unblocked |
| Case Management Scoping Historical record search for ticket correlation (Zun requirement). Not just ticket management — correlation across historical case records. |
P1 | Charlie | 1w | Scoping |
| Multi-Tenancy Architecture Design for separating clients within a shared tenant. Supports SOAR replacement and portfolio-wide deployment model. |
P1 | Charlie | 1w | Scoping |
| SOAR EBITDA Analysis for Portfolio EBITDA case for the SOAR/Swimlane replacement across three vectors: SaaS fee replacement, manual playbook labor, manual triage labor. For Aug 10 portfolio meeting. |
P1 | Tony + Joana | Aug 10 | Planning |
| SOC for AI Session with Deloitte Dedicated SOC for AI deep dive with Deloitte — from Jul 30 session, Krishna: “I can see us taking this to clients.” Decision needed: schedule before or after Aug 10 portfolio meeting. Prep: SOC for AI + Telemetry demo (prototype exists), platform compatibility matrix update, agent health monitoring proposal (Adelina’s request), heterogeneous AI governance story. |
P1 | Joana | TBD | Needs Scheduling |
| QA/Account Manager Role Definition Value First hires positioned as technical QA bridging Deloitte↔Kindo engineering. Technical background required. Ongoing hiring track. |
P2 | Joana + Victor | ongoing | Hiring |
| Turbo Mode Visual Diagram View Visual diagram view for Turbo Mode workflows — aids Swimlane comparison and portfolio presentation. |
P2 | Charlie | 1w | Planned |
Same criteria as Sprint 1 (see Sprint 1 tab for full details):
Same criteria as Sprint 1 (see Sprint 1 tab for full details):
Items that produce a built artifact (agent, integration, code in Kindo) - ranked by strategic priority
Parallel track - program management work that needs an owner and date but doesn't produce a product artifact. Not sprint-rankable. Joana's track.
| Item | Owner | Target Date | Notes |
|---|---|---|---|
| Deloitte Training Cohort 2 Launch | Joana | Jul 10 ✅ | Launched. LMS technical review ✅ complete. Domain fix ✅ complete. |
| Scaling Story for Ron / Forge Point | Tony + Joana | During Sprint 2 | Capture while Tony is present. 3-5× revenue growth narrative for Forge Point VC. |
| Monthly Portfolio Prep (July) | Joana | Ahead of portfolio review | Video evidence + working links for shipped functionality. |
| Value First Hiring | Joana / Victor | Ongoing | Interviews in progress. Invoice → Charlie → Ron. |
| QA/Account Manager Hiring | Joana / Victor | Ongoing | Value First hires positioned as technical QA bridging Deloitte↔Kindo engineering. Technical background required. |
| Teams Integration Instructions → Krishna | Charlie | Sprint 2 Week 1 | Email tenant ID + setup instructions for company-to-company Teams collaboration. |
| Co-Design Session Scheduling | Joana | Sprint 2 | Schedule next co-design session. Cadence open - Joana suggested biweekly (sprint-aligned), Krishna suggested monthly. Left open Jul 10. Needs Tony input. Next session ~Jul 25. |
| Action Items from Jul 10 Session | Joana | Sprint 2 Week 1 | Circulate action items list to all attendees. |
Items requiring team input - updated July 13
Design Session #2 with Krishna (Jul 30) and strategic alignment (Jul 31). SOAR/Swimlane replacement moved to active evaluation.
90-min session with Kush, Krishna, Nathan, Zun. First formal co-design work session.
| Action | Owner | Status |
|---|---|---|
| Share Swimlane JSON workflows to Charlie (ASAP) | Zun + Adelina | Waiting |
| Schedule Turbo Mode deep-dive with Adelina + Zun for SOAR capability mapping | Charlie | Sprint 3 |
| Set up shared ticketing system, get Deloitte's preferred tool | Joana | Sprint 3 |
| Ensure Brandon and/or Marcos attend weekly standups | Joana | Ongoing |
| Retrospective on 05.10 release failure → send findings to Nathan | Joana | Mid-late next week |
| Confirm what's in 2026-07 release (context window compaction, observability, LiteLLM timeout fix) | Joana | In Progress |
| Explore AWS Bedrock quota increase for rate limiting | Adelina | Sprint 3 |
| Continue Teams federation setup (Deloitte side done, T&C side pending) | Krishna / Adelina | Unblocked |
| Reproduce latency issue in isolated prototype environment | Zun | In Progress |
| Designate Deloitte attendees for weekly who can cover all open tickets | Krishna | In Progress |
Full team (Tony, Charlie, Victor, Joana) prep for Jul 10 design work session with Deloitte.
110+ days of empirical data (Mar 9 - Jul 10, 2026) - what worked, what failed, what was killed
| Initiative | Grade | Key Evidence |
|---|---|---|
| Kindo × Deloitte Dashboards | 85% | 2 production dashboards, daily cron, 240 deploys. Owner (Joana) + same-day loop + cron refresh |
| Strategic Dashboards | 80% | Same-day delivery pattern. Revenue Map for Ron dinner = built same day Tony directed |
| Kindo LMS | 75% | T1-T4 deployed (T1/T2 Joana-confirmed). P1/P3/P4 visual re-times done. Cohort 2 launched Jul 10. |
| Tony CoS Dashboard | 55% | 136 deploys, CI/CD working. But: DM capture unverified, Phase 2 never started |
| Autonomous Pipeline | 40% | Architecture complete (43 routes). But: 0 agents dispatched for 42+ days. Pipeline became meta-work |
March-April: Build
Pipeline architecture, 67 routes, Sprint 0, first dashboards. Every problem solved by adding.
May: Peak Complexity
6 new Pages in 3 weeks. 5 AI eval crons. 4 daily dossiers. Maximum complexity = diminishing returns.
June: Subtract
7 systems killed. 0% dossier engagement. Silence First. Human-only review. Higher signal than anything added.
| # | Capability | Who Reacted | When |
|---|---|---|---|
| 1 | Thinking model / decision OS | Igor: "Jarvis not Siri" | May 15 |
| 2 | Knowledge extraction | Steve Ward: "floored" | May 25 |
| 3 | AIPMO / autonomous PM | Valent: SOW signed ($5K) | Apr-Jun |
| 4 | Dependency mapping | NFL corpus proof point | Apr |
| 5 | Sprint planning / estimation | Hector: "that's money" | Jun |
Warren = Autonomous Delivery Partner - the product is the method, Warren is the engine
Every load-bearing milestone (85% of outcomes) used these 5 components. Each Sprint 2 item should map to this template.
A human on the customer side driving the loop (the "Tony" equivalent). Sprint 2 check: Who is the owner for each item?
Customer's own words, not our interpretation. Sprint 2 check: Do we have source material in their language?
Deterministic rules, written - never probabilistic process docs. Sprint 2 check: Are the rules codified or still in people's heads?
Keeps the output alive after initial delivery. Sprint 2 check: Will this need automated updates or is it one-shot?
Output framed in the customer's framework, not VtKl's internal language. Sprint 2 check: Are we using Krishna's/Kush's words or ours?
01 - Owner + Same-Day Loop
The mechanism. Without a named owner driving the loop, nothing ships. Present in: Kindo dashboards, CDO thesis, Execution Plan + Revenue Map, Strategic Portfolio Design, Great Subtraction.
11 - Narrative (Recipient's Framework)
The communication layer. Without narrative framing, output ships but doesn't land. Ron needed visual revenue framing. Igor needed "Jarvis" framing. Krishna needed triage language.
E
Executive
Revenue Map for Ron dinner (May 21-23). Visual, revenue-framed.
S
Strategic
Sprint plan + GANTT + questionnaire (Jun 8-9). Full planning cycle in one thread.
O
Operational
Daily dashboard refresh via cron. 240 deploys. Deloitte logs in daily.
Each person migrated UP in altitude over 110 days. Sprint 2 should respect these altitudes.
| Person | Started As | Evolved To | Sprint 2 Altitude |
|---|---|---|---|
| Tony | Operator (every function) | Teacher → Subtractor | Strategic direction only. Biweekly. |
| Victor | Ops Support | Chief Operating Intelligence | Tactical execution + Warren calibration |
| Joana | Program Delivery | Delivery Authority | Scrum lead + agent design |
| Charlie | Builder | Platform Architect | Architecture decisions only |
| Warren | Engineering Tool | Autonomous Delivery Partner | Execution engine - absorbs operations |
"I don't want to do design partnership work where we equip them to build stuff that we want to build." Deloitte can execute faster than T&C on skills/memory if they know the HOW. Share the WHAT, never the HOW. This applies to all IK/memory sessions with Kush.
Channel activity and progress updates - most recent first
9ff54d70-924e-46c5-81b5-758bd96411cc. Direct webhook trigger. Two steps: (1) API Action → Anthropic /v1/models endpoint, (2) Claude Sonnet 4.6 governance analysis.🔵 #deloitte-agents-design
🟢 #client-kindo-deloitte
🟡 #client-kindo-training
agents/soc-for-ai-discovery/t-and-c/kindo-deloitte-internal created, deployed to kindo-deloitte-internal.pages.devAugust 2026 → January 2027 — team availability, holidays, and meeting cadence across the critical stretch
| Jul | Aug | Sep ⚠️ | Oct | Nov | Dec | |
|---|---|---|---|---|---|---|
| Maternity Leave |
||||||
| Kush OOO |
Out next week | |||||
| Charlie OOO |
💍 Sep 11 — Wedding (Ecuador)
|
|||||
| Tony Travel |
Tentative — dates pending | |||||
| Holidays 🇧🇷 Brazil |
🇧🇷🇺🇸 Sep 7 (Mon) — Independência + Labor Day
|
🇧🇷🇺🇸 Oct 12 (Mon) — Aparecida + Columbus Day
|
Nov 2 (Mon) — Finados
Nov 15 (Sun) — Proc. da República
Nov 20 (Fri) — Consciência Negra
|
🇧🇷🇺🇸 Dec 25 (Fri) — Natal + Christmas
|
||
| Holidays 🇺🇸 US |
↑ Sep 7 shared | ↑ Oct 12 shared |
Nov 11 (Wed) — Veterans Day
Nov 26 (Thu) — Thanksgiving
|
↑ Dec 25 shared | ||
| Co-design | Last session before Charlie OOO | |||||
| Portfolio | Krishna + Kush. SOC for AI + SOAR EBITDA. | Just after Charlie returns; may run remote | Single session for Dec (holidays) | |||
Last updated: July 20, 2026 · 10:30 AM PT