Thoughtworks × Tango Analytics · Engagement Proposal

Agentic Reinvention.

A 30-day ELT-ready plan. A working prototype in days. A production-grade agentic platform in 48 weeks. Gated, fixed-fee, co-built.

A working multi-agent system for lease renewals, built on AI/works to demonstrate the art of the possible. Not a mockup. Click to use it.

Try the working prototype → See the engagement plan →

30 days

ELT-ready implementation plan, before any build commitment

48 weeks

Three gated segments. Fixed-fee for outcomes.

2 + 3 agents

Two in production, three at production readiness. Portfolio curated with Tango in Segment 1.

The Moment

Every company today needs an agentic system strategy.

Ungoverned agentic AI is not a future threat. It is current state. The window to do this strategically is open now.

"Every company in the world today needs to have an agentic system strategy. This is not optional. This is existential."

Jensen Huang · NVIDIA GTC 2026

of enterprises already have AI tools in use without a governance model.

Source: IBM Cost of a Data Breach Report 2025

of agentic AI projects will be cancelled by 2027, driven by unmanaged cost and unmeasured risk.

Source: Gartner 2025

30–45%

productivity gains across knowledge-work functions by 2027, for the organizations that get this right.

Source: Gartner, McKinsey

The Destination

The Agentic Organization.

Becoming agentic is not a technology project. It is a category of company. The destination that gives every modernization decision a purpose.

Stage 01

AI Tourists

Enterprise ChatGPT accounts. Copilot licenses. No systems, no governance, no outcome model.

→

You are here

Stage 02

Agent Builders

Scattered agents, no control plane. The messy middle. Most organizations are stuck here, trying to contain agents instead of harnessing them.

→

Destination

Stage 03

Agentic Native

Agents writing, testing, releasing. Anthropic-class operating model. End-to-end observability, governance, and economics by design.

Three foundations beneath every agent - what Segment 1 proves.

Data readiness

Lineage, quality, and access controls across Core, WatchWire, AgileQuest, and Locatee, so agents can synthesize the data that is Tango's moat.

Agent economics

Token cost is a P&L line item. FinOps, model routing, and cost per outcome, so every agent's margin is known before it scales.

Governance posture

Control plane before agent sprawl. Policy, audit, and human oversight by design, and ASC 842 / IFRS 16 / SOC 2 by default.

The Approach

Three principles govern this engagement.

These are not values on a wall. They shape every commercial, architectural, and team decision in the work that follows.

Build to prove, not advise to plan

Every segment produces functional, demonstrable output against Tango's actual data, constraints, and customer requirements. Strategy is validated through delivery, not presentations.

Product paradigm, not platform rewrite

We are not rebuilding Tango on a modern stack. We are co-creating a fundamentally new product category, agentic AI for integrated workplace lifecycle management, grounded in Tango's proprietary data moat and domain expertise.

Capability transfer by design

Thoughtworks builds alongside Tango's founding team from day one. Product, engineering, platform, data. When we leave, your team owns the platform, the operating model, and every artifact.

Workshop signal · April 21 · Dallas

The Roof Reduction Agent surfaced as a high-value candidate alongside Lease Renewal. Final selection lands via the Segment 1 scorecard, with your team, not ours.

Discovery & Validation

Desirability is proven before commitment scales.

Every agent is prototyped, evaluated against synthetic data, and validated with design partners, before any production build commits.

Agent prioritization scorecard

Tango's portfolio is sequenced through a structured scorecard refined in Segment 1. Workshop signal: the Roof Reduction Agent surfaced as a high-value candidate alongside Lease Renewal, final selection lands as the scorecard matures.

Rapid prototyping via AI/works

The first agent gets built as a working prototype on AI/works in Segment 1: core workflow logic, constraint enforcement, escalation triggers. Independent of ADP infrastructure, against synthetic data, in days not weeks.

Customer desirability signal

The prototype goes in front of design-partner contacts to evaluate desirability and inform design and sequencing of the initial agent portfolio. Customer reaction shapes the roadmap, not just the agent.

Validate, then iterate

Segment 2 takes the proven concept into a closed beta on real customer data with 2–3 design partners. Capture feedback, measure outcomes against success criteria, iterate. Production scale only follows validated value.

Objective

Translate the reinvention proposal into a validated implementation plan. Define shared foundation architecture, AI platform target-state, and first-agent specification with detailed requirements. Rapidly prototype and evaluate the first agent (Lease Renewal Agent, or TBD given Roof Reduction Agent workshop signal) using AI/works against synthetic data, with initial design-partner feedback to evaluate customer desirability and inform design and sequencing of the initial agent portfolio, delivering the ELT-ready implementation plan within 30 days of engagement start.

Focus

Current-state immersion & data estate assessment

Immerse across Core, Edge, R&O, ML/AI, E&S
Map Oracle ADF stored procedure dependencies
Evaluate Redshift / DBT maturity and API readiness
Establish Track 1 cadence with Hope / Empire leads

Focus

Validate architecture & ADP blueprint

Shared foundation layer: data, customer, business rules, security, reporting bridge
AI platform target-state: framework, control plane, governance
Agent prioritization scorecard; first-agent selection
Synthetic data strategy

Focus

First-agent prototype via AI/works

Generate synthetic data from domain schemas
Build first-agent prototype independent of ADP infra
Evaluate vs success criteria
Present to design-partner contacts for desirability signal

Focus

Implementation plan & org readiness

ELT-ready implementation plan: budget, hiring, roadmap
Design-partner selection criteria & shortlist
Org readiness assessment, change plan
AI-First engineering operating model (v1)

Outcomes · Segment 1

An evidence-based foundation for reinvention: architecture, plan, operating model, and a working first-agent prototype against synthetic data.

Shared foundation blueprint

AI platform target-state architecture

First-agent prototype + evaluation

Agent portfolio roadmap & scorecard

ELT-ready plan (30-day)

Design-partner shortlist

AI-First operating model v1

Seg 2 execution plan

Gate · M1

ELT approval of blueprints, plan, prototype + measurable Segment 2 outcome criteria.

Approval gates initiation of Segment 2 investment. 30-day deliverable from engagement start.

First-agent candidate: Lease Renewal Agent or Roof Reduction Agent (workshop signal). Selection finalized via the agent prioritization scorecard during Segment 1.

Objective

Co-develop the shared foundation layer, rebuild the first agent (Lease Renewal Agent or TBD) as MVP on the ADP vertical slice with customer-facing validation via design partners in closed beta, prototype and evaluate the second agent on the ADP against synthetic data, and implement the AI-First engineering operating model for the platform team.

Focus

Shared foundation build

Unified data integration across Core, WatchWire, AgileQuest, Locatee
Customer data, business rules, security, compliance, reporting bridge
ADP pipeline & deployment infrastructure

Focus

First agent MVP + closed beta

Rebuild first agent on ADP vertical slice
Deploy to design-partner environments
Closed beta on real customer data, human oversight
Measure outcomes against success criteria, iterate

Focus

Second agent prototype

Build second agent prototype on ADP vertical slice
Generate synthetic data for second-agent domain
Design-partner feedback & concept validation

Focus

Design partners & operating model

Onboard 2–3 design partners; connect data through shared foundation
Incremental platform team transformation: sprint cadence, DoD, WoW
Product management uplift; change management for GTM shift

Outcomes · Segment 2

A functional, end-to-end agentic platform slice, first agent validated in closed beta against real customer data.

Functional shared foundation

First agent MVP in closed beta

Second agent prototype

Control plane operational

2–3 design partners onboarded

AI-First operating model v2

Seg 3 scale-up plan

Gate · M2

Shared foundation + first-agent MVP demonstrated to Tango ELT + design partners.

Approval of scale-up plan, with elaborated Segment 3 outcome criteria, gates Segment 3 investment.

Objective

Scale the shared foundation to production grade, advance the first agent to production, advance the second agent through MVP, customer validation, and into production, develop the remaining three agents (Workplace Intelligence, Space Optimization/Maintenance, Site Selection) through development and validation with production readiness, broaden the design-partner program to 3–5 customers, and embed organizational capability so Tango's team operates independently.

Focus

Platform hardening & first-agent production

Harden shared foundation for production: performance, security audit, compliance
Extend first agent to production based on beta feedback
Fleet-wide observability, cost management, evaluation at scale

Focus

Second agent to production & agents 3–5

Second agent: prototype → MVP → customer validation → production
Agents 3–5: workflow logic, synthetic evaluation, control plane
Validate agents 3–5 to production readiness

Focus

Design-partner expansion & GTM

Expand design-partner program to 3–5 customers
Outcomes-led sales playbook, competitive positioning, pricing
Thought leadership and analyst briefings
Value proof via won new bookings within FY'26

Focus

Capability transfer & org enablement

PM practices, engineering restructuring, cross-functional patterns
Progressive ownership transfer; Tango leads sprints by end
Phase 2 recommendations

Outcomes · Segment 3

A production-grade agentic platform, 2 agents in production, 3 at production readiness, Tango-led, design-partner-validated.

Production-grade shared foundation

First + second agents in production

Agents 3–5 at production readiness

Fleet-wide control plane at scale

3–5 design partners active

GTM enablement playbook

Won bookings within FY'26

Tango platform team operating independently

Gate · M3

Tango ELT evaluates production platform, full agent portfolio, design-partner results.

Decision: transition to independent ops, or extend into Phase 2 for portfolio hardening and expansion beyond initial 5 agents.

Status

Phase 2 is not part of this engagement.

It is an optional path forward Tango ELT can choose at the M3 gate. Scope, duration, and commercial structure are defined jointly at that point, grounded in Segment 3 outcomes.

If elected

What it could cover

Harden initial 5-agent portfolio in production
Develop additional agents beyond the initial portfolio
Scale customer migration toward GA

How it's shaped

Defined at the M3 gate

Scope & duration set based on Segment 3 evidence
Design-partner outcomes inform agent priorities
Tango's hiring trajectory determines co-build balance
Commercial structure negotiated separately

Commercial Vehicle

A commercial model built for the work itself, not the work as imagined.

Fixed-fee for outcomes. Gated segment commitments. Each milestone gate is both an acceptance checkpoint and a scope elaboration point, so commercial specificity scales with what we actually learn.

Gated

Three segments. Each segment requires the prior gate's outcomes to be approved before investment commits.

Fixed-fee for outcomes

Not time and materials. Each segment is priced against defined outcomes and measurable success criteria.

Progressive elaboration

Segment 1 outcomes are fully defined at start. Segment 2 and 3 outcomes are baselined as a deliverable of the preceding segment.

M1 in 30 days

An ELT-ready implementation plan within 30 days of engagement start. Architecture, plan, first-agent prototype results.

A Working Demo

This is how Segment 1 actually works.

This demo was built on AI/works to demonstrate the art of the possible. It is a working multi-agent system for lease renewals. Not a mockup. Not a wireframe. The first agent we'll prototype with you in Segment 1 is selected by your team, not ours, against your synthetic data, and put in front of design-partner customers before any production commitment.

tango-leasing-prototype.pages.dev

Open in new tab ↗

Treat this as a window into how we work, not a preview of the deliverable. AI/works lets us go from concept to working evaluation against synthetic data in days, so design-partner feedback can shape the portfolio before commercial commitments get made.

Outcomes that matter to Tango

This is not a services proposal. It is a reinvention bet.

Every segment is measured against the outcomes Tango's own ELT proposal laid out: reposition the market narrative, unlock the acquisition moat, and get ahead of the AI-native upstarts before they expand beyond their initial niches.

Reposition the market narrative

From "legacy IWMS" to "AI-first workplace platform." Re-open deals Tango is currently losing to AI-native entrants like VergeSense, Density, and Facilio. Win the category-shift buyers like Apple and NVIDIA.

Protect the $70M base

The shared foundation layer lets existing customers benefit from AI capabilities without a forced platform migration. Continuity for today's customers while the new platform matures.

Unlock the acquisition moat

WatchWire, AgileQuest, and Locatee finally compound instead of fragmenting. Lease + occupancy + energy data, in one agent, is what no point-solution can match.

Hedge the investment

Shared foundation delivers value even if agentic adoption takes longer than planned. The data work is load-bearing either way, agents are the upside, not the whole bet.

The Portfolio Tango Identified

Five candidate agents.

The agent concepts Tango identified, the portfolio we'd prioritize, sequence, and prototype together in Segment 1. Click any to read its description.

Outcome Arc

Credibility in 6. Validation in 12. Scale in 24.

Tango's own outcome arc, broader than the 48-week engagement window. The TW engagement closes around month 12 with a production-ready platform; month 24 is Tango executing on what we built.

6 mo

Establish credibility and speed

Shared foundation under active development. First agent working prototype. 3–5 design-partner customers committed. Market signal that Tango is no longer a legacy player.

12 mo

Validate and demonstrate

First agents in closed beta on real customer data. Project Hope and Project Empire substantially complete. Market narrative shifted. Pipeline and employee base re-energized.

24 mo

Scale and differentiate

New AI platform generally available with initial agents. Customers migrating through the shared foundation. New acquisition driven by AI value proposition. Competitive losses declining.

~12 months post-engagement · Tango-led

What You'll Own

Every artifact compounds. Every artifact is Tango's.

Each segment leaves Tango with a concrete asset, not a deliverables binder. Thoughtworks retains rights to its own platforms; everything produced for Tango belongs to Tango.

Segment 1 · 6 weeks

An evidence-based foundation

Shared foundation architecture blueprint
AI platform target-state architecture
First-agent prototype + synthetic data evaluation
First-agent requirements with implementation plan
Agent prioritization scorecard & portfolio roadmap
Data estate inventory & readiness assessment
ELT implementation plan (30-day deliverable)
Design-partner selection criteria & shortlist
AI-First engineering operating model (v1)
Segment 2 execution plan with measurable outcomes

Segment 2 · 14 weeks

A functional, end-to-end platform slice

Functional shared foundation layer
First agent MVP validated in closed beta on real data
Second agent prototype evaluated on synthetic data
Agent control plane operational (governance, observability, cost)
2–3 design partners onboarded and in closed beta
AI-First operating model (v2) validated in practice
Refined change management plan & org readiness
Segment 3 scale-up plan grounded in design-partner evidence

Segment 3 · 28 weeks

A production-grade agentic platform

Production-grade shared foundation (performance, security, compliance)
First and second agents in production with design partners
Agents 3–5 developed, validated, at production readiness
Fleet-wide observability, cost controls, evaluation at scale
3–5 active design partners driving GTM signal
GTM enablement playbook & market positioning assets
Value proof via won new bookings within FY'26
Tango platform team operating independently
Phase 2 recommendations

Team Archetypes

A senior-weighted team, organized by capability, not headcount.

Because this is a fixed-fee-for-outcomes engagement, the composition flexes to the work. Six capability archetypes scale across 48 weeks alongside Tango's founding AI platform team.

Strategic Leadership

Segments 1–3 · Onshore

Engagement vision, ELT alignment, implementation plan ownership, milestone readouts. Owns design-partner strategy and GTM enablement. Includes delivery transformation leadership for the AI-First operating model and waterfall-to-agile transition.

AI Platform Engineering

Segments 1–3 · Onshore + Offshore

Shared foundation architecture, AI platform target-state design, agent architecture, agentic workflows, LLM integration, control plane configuration. Builds alongside Tango engineers with progressive ownership transfer.

Data Foundation

Segments 1–3 · Onshore + Offshore

Cross-product data model, integration patterns, data readiness, governance, scalable foundations for agent context and reporting. Schema reconciliation across Core, WatchWire, AgileQuest, Locatee. Tapers as the foundation stabilizes.

Quality & Evaluation

Segments 2–3 · Offshore

Agent evaluation rigor: synthetic data validation, agent outcome measurement, evaluation harness ownership, design-partner beta measurement. Secondary coverage of traditional QA alongside Tango's QA function.

AI/works + Agent/works Specialists

Segments 1–3 · Onshore

Technical oversight of AI/works (reverse engineering, prototyping, forward engineering acceleration) and Agent/works (control plane reference architecture, observability, governance). Embedded across teams and workstreams.

Tango Co-Build Counterparts

Segments 1–3 · Tango-supplied

Product Owner, CRE Domain SME, Product Designer / UX Lead, QA Engineer, and 5–6 Platform Engineers. Co-development from day one ensures the platform is Tango's to operate, not just Tango's to receive.

On-Ramp Profile

A senior-heavy start. A delivery-heavy peak.

Seg 1 · Wks 1–6

→

~14

Seg 2 · Wks 7–20

→

~17

Seg 3 · Wks 21–48

Indicative FTE counts only. Composition flexes to the work as scope is elaborated at each gate. Tango co-development resources are sized separately and confirmed at engagement start.

Thoughtworks Tooling

Two proprietary platforms. Embedded in how we deliver.

AI/works accelerates how we deliver. Agent/works runs in your stack as the control plane for your agents, with favorable terms if Tango participates as a design partner.

Reverse engineering (CodeConcise) accelerates Segment 1 codebase and domain understanding, extracting structured specifications and knowledge graphs from Tango's existing estate to compress immersion timeline.
Rapid prototyping powers the Segment 1 first-agent prototype, enabling working agent evaluation against synthetic data, independent of ADP infrastructure.
Forward engineering (TWist) accelerates microservice development for the shared foundation: data integration services, business rules API layer, reporting bridge components.

AI/works is embedded in our delivery methodology. All outputs are Tango-owned.

A/w

Agent/works

Composable agent governance & control plane

Licensed as the control plane for Tango's ADP and agent portfolio. Production governance, observability, cost controls, evaluation, and audit trails. All running in Tango's stack from Day 1.
Favorable licensing terms if Tango elects to act as a design partner. The design-partner relationship gives Tango influence over the Agent/works roadmap and preferential commercial terms.
Integration architecture, configuration scope, and rollout plan are defined during Segment 1 and delivered at M1 as part of the implementation plan.

ASC 842

IFRS 16

SOC 2

Thoughtworks retains IP rights to both platforms. All engagement-produced artifacts and code are Tango-owned.

Governance & Risk

Lightweight cadence. Transparent escalation. Risks managed from Day 1.

Four governance tiers, calibrated by altitude. The top risks are named and mitigated in the contract, not buried in an appendix.

Level

Cadence

Participants

Purpose

Operational

Daily

Platform squad + TW team

Co-sourced collaboration, progress, blockers

Track 1 Coordination

Weekly

TW Team Leads + Hope/Empire leads

Shared dependency alignment, extraction progress, risk escalation

Steering

Bi-Weekly

Tango + TW leadership

Performance, RAIDs, ways-of-working, escalations

Executive

Monthly

Tango ELT + TW leadership

Milestone readouts, gate approvals, go-forward decisions

High Impact

Oracle ADF complexity / Project Empire dependency

Mitigation

Extraction is Tango-led (Project Empire). TW designs the shared foundation to consume extracted services, not raw ADF artifacts. Dependency tracked explicitly from Segment 1 via weekly Track 1 coordination.

High Impact

Data estate fragmentation exceeds estimate

Mitigation

Segment 1 data estate assessment provides early visibility. Foundation designed to accommodate heterogeneous schemas. Segment 2 scope bounded to highest-readiness data domains; later domains re-sequenced as clarity develops.

Medium Impact

Agent portfolio velocity, 5 agents in 48 weeks

Mitigation

Portfolio sequencing validated at M1 against actual data and design-partner feedback. Tango co-development capacity sized for sustained delivery. Agents 3–5 may re-sequence to Phase 2 if capacity falls short.

Medium Impact

Design-partner recruitment stalls

Mitigation

Selection criteria defined in Segment 1 with Tango sales leadership. Minimum viable design-partner profile established early. Fallback: agents proven on synthetic data with real-data validation deferred.

Medium Impact

Platform team capacity, legacy incident burden

Mitigation

Platform team fenced off from legacy operations per the reinvention proposal. TW co-sourced model provides capacity buffer. Weekly steering surfaces capacity risks; scope bounded at milestone gates.

Medium Impact

Organizational change resistance during transition

Mitigation

Delivery transformation lead dedicated across all three segments. Change management plan developed in parallel with technical delivery. GTM framed around outcomes, not automation.

Critical Dependencies

What Tango supplies. What we track from Day 1.

Six items sit on the critical path. Each is named, owned, and surfaced at the right governance tier, so a dependency slip becomes a steering conversation, not a Segment 3 surprise.

Project Empire extraction progress

Stored-procedure extraction directly gates shared foundation integration timeline. Weekly Track 1 cadence established in Segment 1 to maintain critical-path visibility.

Design-partner legal agreements

Customer availability and legal/data-sharing agreements must be in place before the closed beta in Segment 2.

Data infrastructure readiness

Data warehouse maturity (Redshift consolidation) and API availability assessed in Segment 1. Material gaps may require scope adjustment at M1.

Tango co-development capacity

Delivering the full agent portfolio in Segment 3 requires sustained Tango co-development capacity (5–6 engineers) across multiple concurrent workstreams. Shortfall may re-sequence agents 3–5 into Phase 2.

Shared foundation domain coverage

Agents 3–5 need data domains the foundation hasn't covered yet: occupancy, energy, maintenance, facilities. These get extended by mid-Segment 3. Coverage gaps may re-sequence agents into Phase 2.

ELT gate decision cadence

Milestone gate decisions must land within one week of readout to maintain engagement momentum and team continuity. Built into the executive cadence.

Prerequisites

Three things we assume in place by engagement start.

Tango authorizes formation of the dedicated AI platform team (founding group + product lead) concurrent with or prior to Segment 1 kickoff
Project Hope and Project Empire continue in parallel under Tango ownership

Timely access to all product codebases, data stores, infrastructure, and documentation across Core, WatchWire, AgileQuest, Locatee, and E&S
Required AI/ML tooling, cloud infrastructure, and sandbox environments procured and available by Segment 2 build initiation

Design-partner customers are identified and willing to participate within Segment 2 timeline; Tango sales leadership drives recruitment with TW advisory support
All engagement artifacts owned by Tango; TW retains rights to pre-existing IP (AI/works reference architecture, Agent/works source code, methodologies)

Next Steps

Three steps from this site to a signed engagement.

A simple sequence. No surprises. Commercial and legal start in parallel only after we are aligned on the work itself.

Tango reviews this site (today)

Walk the narrative and the prototype with the team you'd want in tomorrow's alignment conversation. Capture questions and pushback so we can address them live.

Alignment conversation, Wednesday April 29, 2:00–2:30 PM ET

Already on the calendar. A working session to mutually align on the engagement approach: principles, segment shape, gate criteria, co-build model, and the agent portfolio shortlist. The goal is a shared point of view, not a sales meeting.

Kickoff commercial & legal in parallel

Once aligned on approach, we run pricing and contracting tracks in parallel, so M0 kickoff is days away, not months.

Once we kickoff, you have an ELT-ready implementation plan within 30 days, before any Segment 2 commitment is made. That is the smallest, fastest unit of evidence we can offer.

This site is for Tango leadership review.

Agentic Reinvention.

Every company today needs an agentic system strategy.

The Agentic Organization.

Three principles govern this engagement.

Desirability is proven before commitment scales.

From kickoff to production-grade in 48 weeks.

A commercial model built for the work itself, not the work as imagined.

This is how Segment 1 actually works.

This is not a services proposal. It is a reinvention bet.

Every artifact compounds. Every artifact is Tango's.

A senior-weighted team, organized by capability, not headcount.

Two proprietary platforms. Embedded in how we deliver.

Lightweight cadence. Transparent escalation. Risks managed from Day 1.

What Tango supplies. What we track from Day 1.

Three steps from this site to a signed engagement.