Announcement PDF decidrindustries.ai FY 2 6 in Review and Outlook Decidr AI Industries · ASX: DAI · September 2026 Buil t the plat f orm. Now running the business. The forward-looking statements included in these materials involve subjective judgement and analysis and are subject to significant uncer t ainties, risks and contingencies, many of which are outside the control of, and are unknown to, Decidr AI Industries Limited ("Decidr" or "DAI" or "the Company"). In par t icular, they speak only as of the date of these materials, they are based on par t icular events, conditions or circumstances stated in the materials, they assume the success of the Company's business strategies, and they are subject to significant regulatory, business, competitive, currency and economic uncer t ainties and risks. Decidr AI Industries Limited disclaims any obligation or under t aking to disseminate any updates or revisions to any forward-looking statements in these materials to reflect any change in expectations in relation to any forward-looking statements or any change in events, conditions or circumstances on which any such statement is based.You should monitor any announcements by the Company lodged with the ASX. Nothing in these materials shall under any circumstances create an implication that there has been no change in the af f airs of the Company since the date of these materials. No representation, warranty or assurance (express or implied) is given or made in relation to any forward-looking statement by any person (including the Company). In par t icular, no representation, warranty or assurance (express or implied) is given in relation to any underlying assumption or that any forward-looking statement will be achieved. Actual future events and conditions may vary materially from the forward-looking statements and the assumptions on which the forward- looking statements are based. Given these uncer t ainties, readers are cautioned not to place undue reliance on such forward-looking statements.To the maximum extent permitted by law, Decidr disclaims all liability and responsibility (including without limitation any liability arising from faul t or negligence) for any direct or indirect loss or damage which may arise or be suf f ered through use or reliance on anything contained in, or omitted from, this presentation. All amounts are in AUD unless stated otherwise. Some amounts may not add due to rounding. Annualised exit-rate revenue (AER) is a non-IFRS measure that has not been audited or reviewed. FY 2 6 figures are preliminary and unaudited unless stated otherwise. This presentation has been authorised for release by the Board of DAI. Impor t ant notice 2 Decidr Ai Industries 3 What we do Every organisation will run on AI. Decidr’s aim is to be the operating layer that makes that work in practice. • Business knowledge captured once and reused. • Work executed as small, auditable tasks, every step recorded. • Cost per task falls as AI models get cheaper; price is set by the work done. • Customers keep control of their data, their models and their rules. DecidrOS captures how a business works and runs its workflows as small, recorded tasks inside the systems it already uses. Sugarwork captures what experienced people know, including in meetings via Rumi, so it's reused rather than lost. Sold direct to customer, via sof t ware par t ners, consul t ing and IT firms, or the AWS and Google Cloud marketplaces. Vision Listing ASX: DAI US subsidiary Decidr U.S. Inc., Delaware, 95.87% owned Bases Sydney and New York Products DecidrOS, Sugarwork, Rumi AER at 30 June 2026 A$13.6M FY 2 6 statutory revenue A$4.73M Par t nerships 14+, incl. AWS, Google Cloud, SCX.AI Key facts FY 2 6 was the year the plat f orm reached paying customers in production.What follows: what was delivered, what's changed since 30 June, and the outlook. An ASX-listed company building an operating system that runs business workflows on AI, under the customer's own rules. What Decidr is 4 AER is the prior month’s invoiced revenue mul t iplied by 12, excluding Edible Beauty. A lead indicator, not statutory or annual revenue. A$13.6M Annualised exit-rate revenue at 30 june 2026 7.5× the prior corresponding quar t er and 1.7× March 2026. Six consecutive quar t ers of growth. A$1.8m Jun 25 A$2.5m Sep 25 A$4m Dec 25 A$8.06m Mar 26 A$13.6m Jun 26 With revenue lines maturing, the group will begin to include ARR and TCV repor t ing figures from the December repor t ing period FY 2 6 · Year in review Annualised exit-rate revenue (AER) of A$13.6M at 30 June 2026 AER · A$M 5 A$4.73M 12.1 A$23.4M 100% ~180 A$33.1M Funds were allocated into product, go-to-market and the US expansion. Funded by A$33.2M raised in FY 2 6 (A$19.0M in September 2025, A$14.2M in May 2026). FY 2 6 at a glance FY 2 6 financial and operating summary Preliminary and unaudited, from the Appendix 4E lodged 31 August 2026. Ownership of decidr.Ai Updates in june 2026 decidros release Net loss af t er tax Statutory revenue Par t ner and direct revenue growth Cash at 30 june 2026 Up 99% on FY 2 5 (A$2.4M) A$3.19M in FY 2 6 vs A$0.26M in FY 2 5 (Appendix 4E revenue lines) From A$7.8M a year earlier From 51%; Sugarwork owned via Decidr U.S. Inc The largest single release to date Includes A$9.3M D&A and A$2.6M share-based payments × 6 Cash position and cost program Cash A$23.4M at 30 June 2026. A$5M annualised cost reduction program initiated in Q 4 FY 2 6 . A$5M Annualised cost reductions targeted The A$5m is an annualised reduction target, not an achieved FY 2 6 saving. FY 2 6 cash movement · A$m · preliminary 4E cash flow 7.8 Jun 2025 +37.8 Net financing −19.5 Operating use −2.6 Investing use 23.4 Jun 2026 Cash and financial discipline 01 Market context 8 Enterprise AI adoption in FY 2 6 The market in FY 2 6 Adoption is broad; operational deployment at scale remains limited.The constraint is knowledge of how work is done, held by people rather than systems. 88% of businesses were using AI in 2025, up from 55% in 2023 Stanford HAI, AI Index 2026 2% of organisations have operationalised AI at scale, though 83% want it in core processes IBM (2025); Morningstar (May 2026) 64% of 2026 enterprise AI budgets are reserved for core business functions, not experiments MIT research, 2026 95% of generative AI pilots fail to reach measurable P&L impact MIT NANDA, State of AI, Aug 2025 10 annual fall in the cost of AI execution from infrastructure and algorithmic ef f iciency Q 4 FY 2 6 investor webinar 45% of Australian businesses face regular challenges because only a few employees know key workflows Australian AI Readiness 2026 Buyers now ask where the AI runs and who controls the data. Sovereignty has become a purchase requirement. × 9 Decidr compared with AI assistants Ai assistants Sovereign operational ai One chat, one answer Knows the internet, not your business Stops when the conversation ends Someone still has to check the work Every team, end to end Buil t on the business’s own knowledge Runs with nobody at the keyboard Every step logged and auditable Assistants are suppliers to Decidr, not competitors. Decidr routes tasks to them and switches as cheaper models arrive. Where DAI fits 9 Assistants suppor t an individual. Decidr runs business workflows end to end. 10 How the plat f orm works: capture, coordinate, compound Capture Sugarwork and Rumi record how people actually do the work and handle the exceptions.The customer owns what is captured. Subsea cable company $3.8M DUPLICATED EFFORT FOUND 480 to 738 hours a week 1% margin on $1.5B of work Delivered in 10 weeks A hunch became a business case Coordinate DecidrOS breaks workflows into small, auditable tasks and sends each one to the cheapest reliable option: code, a model or a person. eBev · venue ordering $0.55 PER VENUE WEEKLY,WAS $192 385 min to under 1 min 62% to 96% accuracy 94 to 95% task-level gross margin 10 to 20x ROI from year two Compound The ontology and Agentic Graph link organisations, people, products and workflows so each deployment makes the next one cheaper. The agentic graph · in build 143,325 BUSINESSES SEEDED ON THE GRAPH 87,911 Pre-identified links 7,605 taxonomy terms 3.6M attributed data points Capture and Coordinate are selling today. Compound is the par t still being buil t . How it works 01 02 03 Each step, with a customer resul t for the two steps in commercial use. Business & financial per f ormance 02 12 Revenue channels and stage of development 01 Seeding DecidrOS (BetaOS) Early adopters and foundation customers deploying the operating system. 14% 02 Commercialising Embedded par t ners Sof t ware companies that sell Decidr-buil t AI apps to their own customers. Includes Global System Integrators (“GSI”) - also known as consul t ing companies 54% 03 Emerging Commercial resellers Consul t ing and IT services firms that resell Sugarwork and Decidr to large companies. Contracts in US, AU, SG 04 Growth Sugarwork + Rumi Writing down what exper t s know, then training a large language model the customer owns (“sovereign model”) 32% The strategy OF AER OF AER OF AER (DIRECT ENTERPRISE) Share of AER at 30 June 2026 by channel. + others + others + others 13 Every deal adds a node. Some deals add a network Par t ners attach thousands of businesses at once. Enterprises attach one at a time. Foundation customers prove the schema that both rely on. Foundation customers Targeted, scaled ver t icals ECONOMICS: A$100,000 ARR per customer Enterprise customers One enterprise at a time ECONOMICS: A$50k pilot to ~A$1M+ lifetime Embedded par t ners One build, whole network ECONOMICS: 50% of workflow revenue via the par t ner, no acquisition cost Every customer and par t ner is buil t on the same Agentic Graph and can transact with the others. One shared schema, reused on every deployment. Go-to-market Decidr Decidr Sugarwork Graph Rumi Decidr Graph Graph Sovereign AI Sugarwork 14 Customer deployment timeline: an aged care provider Pilot to a two-year agreement in six months. An Australian aged care and retirement living provider; customer not named. Jan 2026 First conversation 30 Mar 2026 A$50k pilot signed Apr to Jun 2026 Knowledge capture: 42 sessions Within 12 weeks Findings and dashboard delivered 30 Jun 2026 A$300k two-year agreement Sep 2026 Second project star t s The trigger Aged care reforms raised invoice volumes while margins tightened. Knowledge of how the finance process actually ran sat with a small number of people.The engagement later found 24 processes that depend on one person. What we buil t 22 people across 8 functions took par t in 42 knowledge-share sessions, about 69 hours in total. 65 processes were catalogued and scored. Outputs: process map, dependency map, RACI, a register of 62 pain points, a resource heat map and 27 sequenced recommendations. What happens next A two-year agreement covering 4 to 5 projects a year.The second project, workforce management, star t s in September. A $60k invoice-automation proposal from our SAP investment is in demonstration with Decidr, buil t on the assessment. 12 weeks Pilot signed to findings and workflows delivered A$350k Contracted to date: A$50k pilot and A$300k two-year agreement Sovereign Model Discussions underway using captured knowledge from earlier projects Customer case study Sugarwork 15 Distribution channels AWS, Google Cloud, sovereign compute in Australia, and the US commercial base. US landing and network Oct 2025 Decidr U.S. Inc established (Delaware) Jan 2026 SOSA International Landing Pad, backed by NYCEDC Jul 2026 US Board: Jenny Fielding, Vanessa Liu,Vinny Lingham FY 2 7 New York as commercial base for Nor t h American growth Amazon Web Services Mar 2025 Strategic par t nership announced Q 3 FY 2 6 Marketplace Fast Track listing live Jun 2026 DecidrOS live on Marketplace; AWS and Stripe billing Jul 2026 Accepted into ISV Accelerate; co-sell with AWS sales Google Cloud and sovereign compute Q 4 FY 2 6 Google Cloud relationship initiated Mar 2026 MOU with an APJ data centre provider for NVIDIA inventory May 2026 SCX.ai: first Australian node for sovereign inference Roadmap Additional US, EU and APAC nodes scheduled We reach large companies through the buying channels they already use, without a direct sales force in every market. Distribution Decidr Sugarwork Decidr Sugarwork Sovereign AI Graph 16 Product delivery, Q 4 FY 2 6 Released in the June 2026 release cycle, and the current roadmap. Cheaper AI models lower our cost per task. Our prices do not fall with them. Technology progress ~180 updates in the June 2026 release cycle 13 new integrations added 4 messaging sur f aces via Agent Relay BYOK run on your own model and infrastructure Shipped in FY 2 6 • DecidrOS Beta and CRM/RevOps agents. First version of DecidrOS, with AI that per f orms real sales and revenue tasks. • Mul t i-plat f orm query engine. Ask questions in normal language across systems and data; every AI call is costed. • DocFlow. Reads and understands documents, then uses them to trigger work and AI workflows. • Entity ranking, data layer, graph objects. Maps how people, customers, data and processes connect, and acts on those relationships. • AWS Marketplace and billing. Find, buy and pay for Decidr through AWS. • The data model and engine took longer than planned.What we are building has not changed. Roadmap • DecidrOS sel f -service release and Agentic Graph developer layer, March quar t er 2027 (working date) • Rumi plat f ormed into Sugarwork, December quar t er • A model the customer owns, 2027 17 How the Decidr plat f orm is buil t Seven layers. Each one is a capability the layer above relies on. The three highlighted layers are the ones currently in focused production. Products Decidr, Sugarwork and Rumi. Screens are views over the same model, kept simple for each role. Workflows and agents Work is broken into small tasks. Each task goes to the cheapest reliable option: code, a model, a person or a marketplace vendor. Every step is recorded. Wallets A budget is set before any task runs, for people and agents alike. Cheaper AI models lower the cost per task; price is set by the work done. Events engine Every action and decision is written once, permanently, with who did it and why. Current state and history are read from that record. Entities (the ontology) One shared model of the business: organisations, people, products, workflows, decisions and goals. Extended as customers need it. Accounts One account per organisation. I t s data, its models and its rules stay its own. Auth Identity and permissions.Who can see and do what, by role. Runs On Sold four ways Direct, embedded par t ners, resellers, AWS and Google Cloud marketplaces. The three foundations Finishing these moved the sel f -service release by a quar t er. Every later deployment reuses them. Customer control Own data, own models (bring your own key), own rules. The sel f -service release lets customers set up Decidr themselves. I t sits on the same six layers beneath the products. The plat f orm 18 Harnesses vs Schema: what they are and why they matter The model is rarely what decides whether AI does real work.The sof t ware around it is. AI model Answers a question. Remembers nothing, uses no tools, spends no budget, keeps no record. Harness The sof t ware around the model. Memory, tools, a record of every action, a budget, and oversight. Working AI Work that gets finished, can be checked, and can be trusted with money and customers. What a harness is An AI model answers questions.To do a job it needs more: memory, tools, a record of what it did, a budget, and someone watching.That layer is the harness. Coding assistants were the first products to need one, so theirs are the most advanced today. Why it matters Most AI projects fail between the demo and real use. 95% of generative AI pilots show no measurable profit impact.The model is rarely the problem.The harness is. Why being ahead matters The leading coding harnesses are adding these features now, one at a time, for single developers. Decidr buil t them into the plat f orm for whole organisations. That is what took longer to build, and what every customer reuses.The next page shows the comparison. 95% figure: MIT NANDA, State of AI in Business, August 2025 A harness turns a model into something a business can rely on. Decidr buil t its harness into the plat f orm. The plat f orm 19 Ahead of the harnesses What Decidr already has, what the planned agent layer adds, and how that stacks against the five coding-agent harnesses assessed: Claude Code, OpenCode, DeepSeek, OpenChamber and Oto-Dock.The harness ecosystem is converging on ideas Decidr made structural decisions about years earlier. Capability comparison The harnesses are adding these capabilities one at a time. Decidr buil t them into the plat f orm foundation. The plat f orm dsh Claude Code OpenCode OpenChamber Oto-Dock Decidr Immutable event log Durable workflow engine Flexible domain data model Process supervision Provenance on every fact Budget gates execution KEY Shipped in that system Par t ial or emergent Absent Decidr column: buil t today, except supervision, which is plat f orm-native plus a planned oversight process. Sources: Decidr engineering assessment of the five harnesses from published source code,10 September 2026, and Decidr architecture documentation. 20 How we get there is the whole question The problem Knowledge work today 21 Work runs down a production line, not through a meeting room The breakthrough The production line is deterministic. Build the rails once, then route each task to the right executor 22 Decidr turns knowledge into structured workflows that can be executed, improved and reused The breakthrough Via Decidr, work is harnessed in manufacturing-like, deterministic production lines that can be reused, are auditable, quality controlled and inputs can be tested and swapped for optimal economic outcomes. 23 The same graph of work customized for similar functions across industries The breakthrough The vocabulary changes between industries.The structure doesn’t.Task standards buil t once hold everywhere 24 Graphs make markets The breakthrough Search, social and payments each became a market when someone named the objects. Work has no such graph yet. 1,2,3,4 : eMarketer, Statista, McKinsey Global Payments Report, MIT Project Iceberg 03 Outlook 26 Commitments and outcomes, FY 2 6 to CY 2 7 Commitments made in ASX announcements and webinars, their due dates and status. Open items run to CY 2 7 . Category Commitments Accountability Commercialisation & Revenue Go1, ELMO and eBev to contribute revenue; Cash receipts to catch up with AER; DecidrOS sel f -service release Conver t commercial traction into recurring revenue and cash, while creating a more scalable route to customer acquisition Product & Technology DecidrOS Beta launch; Agentic Graph plat f orm extension live; Sovereign model fine-tuning product Build dif f erentiated, scalable technology that expands the addressable oppor t unity and strengthens DecidrOS's strategic position US Market & Strategic Positioning Enter the US via Sugarwork; Appoint a Sugarwork CEO; AWS case studies shared Establish credibility and commercial capability in the world's largest AI market, suppor t ed by customers, leadership and ecosystem validation Capital & Operating Discipline A$5m of annualised cost reductions Extend runway and improve operating leverage, giving the business more time to conver t product and commercial investment into revenue Benefit 27 Update since 30 June 2026 September quar t er Q 3 Cash Receipts up vs Q 2 The quar t erly trajectory is consistent with the Group's focus on conver t ing pipeline to collected revenue in a timely manner. Product dates DecidrOS sel f -service to include business and consumer organisations Schema, entities and events have progressed and will be included in the H 2 DecidrOS release. DecidrOS will have two main components BusinessOS and LifeOS. Both will operate on an open-sourced Decidr schema allowing Businesses and consumers to build workflows. Agentic Graph production containing marketplace dynamics continues inside the FY 2 7 window already stated. The Group's improving cash receipt trajectory, suppor t ed by a deliberate reduction in its operating cost base, is extending runway and positioning Decidr for its next phase of commercial growth What has changed September quar t er Cost base reducing The Group has under t aken a deliberate re-indexing of its cost base, with strict capital allocation disciplines and governance frameworks embedded across all business units to extend runway and suppor t the Group's path to commercial scale. 28 Outlook for the December and March quar t ers Expected milestones and dates. Dates are calendar quar t ers; working dates are marked. 01 Sugarwork and par t ners • Sugarwork revenue will continue to grow at a double-digit rate in the September quar t er • New CEO appointed and 5 sales resources onboarded in May-Jul ‘26 • ~90 deals in pipeline with customers Decidr and the Agentic Graph • Foundation customers repor t ed each quar t er by stage: assessment, orchestration, live • Agentic Graph developer layer, March quar t er 2027 (working date) • Decidr sel f -service release, H 2 2026 • Rumi plat f ormed into Sugarwork, December quar t er 2026 • Sovereign model development, 2027 Cash and people • Extended runway (against Q 4 Appendix C) as strict cost discipline underpins all capital allocation • Reshape of org design to better optimize product and tech resources based on exper t ise • Commercial momentum building with pipeline conversion accelerating into Q 4 , suppor t ed by new customer activations across mul t iple product lines • Board and governance infrastructure strengthened during the period, including the establishment of the US advisory board Strong pipeline of deals in both the US and Australia for Sugarwork and par t ners. Outlook 02 03 decidrindustries.ai Questions Decidr AI Industries Limited · ASX: DAI Appendix. Customer case studies Mercha, Scentia, a subsea cable company, Appen, Lcubed, Supergoop. 31 Customer case study: Mercha, proposal generation Decidr buil t a proposal-generation system for Mercha, Australia’s leading promotional merchandise plat f orm. Cost to Mercha $70K Initial set-up cost Resul t s for Mercha 2.5hrs to ~10sec Proposal turnaround time 200 to thousands Monthly proposal capacity Success story Before Decidr, proposals required hours of manual work across pricing, sourcing, formatting, and mock-up creation. Every quote depended on operational knowledge spread across inboxes, spreadsheets, and the sales team’s heads. Today, Decidr automatically qualifies leads, assembles proposals, generates branded mock-ups, and prepares pricing in seconds instead of hours, allowing Mercha to scale proposal volume without scaling headcount. 32 Data prep Analysis Execution Task Read Research Find Identify Select Calc Calc Generate Draf t Assemble Send Track Time Required (Human Sales Mgr) 5m 20m 10m 10m 20m 20m 15m 25m 15m 10m 5m 5m Time Required (Agent: Mercha AI) 0m 0m 0m 0m 0m 0m 0m 0m 0m 0m 0m 0m TASK TIME 160m (2.5hrs) ~10 sec LEADS FULFILLED 25% 100% ANNUAL CAPACITY 50/mo Thousands / mo Human Agent Per-step times reconciled to sourced range of 2–3 hrs per proposal pre-Decidr, ~10 sec post-Decidr Mercha proposal workflow Mercha proposal workflow: task time before and af t er Decidr 33 Mercha proposal workflow: product screens Leads are enriched and qualified automatically. Pricing, mock-ups, and proposal assets are generated and assembled. Teams review, adjust, and send proposals in minutes. Mercha proposal workflow 34 Mercha: outcomes beyond the initial workflow The Proposal Generator gave Mercha something even more valuable than speed: structure. Memory, context, and repeatable workflows were wired into the business for the first time and can extend across the rest of the company. Proposal per f ormance improves with every workflow Trackable links feed open, click, and conversion data back into DecidrOS. Every proposal improves recommendations, pricing logic, and copy choices behind the next one. Operational coordination extends across the business The same operating layer that runs proposals now extends to lead qualification, tiered pricing, supplier coordination, resource allocation, and forecasting. Internal operational knowledge becomes reusable infrastructure The team’s product, pricing, and customer judgement are mapped into Mercha’s instance, a golden-record view across HubSpot and Shopify. Operational systems remain stable as models improve Workflows and infrastructure stay intact as foundation models improve. Mercha owns their operating layer; Decidr keeps the substrate current underneath it. Success story 35 Mercha: customer statement Mercha didn’t wait for the per f ect plan.They chose their highest value star t ing point, the proposal generator. This operating layer can now extend across other decision-heavy processes without impacting headcount. “I f you believe AI is going to drastically transform every par t of every business in every way — which it will — then you need to par t ner with somebody who’s building a solution for that future.That’s why we’ve par t nered with Decidr.” BEN READ, CEO, MERCHA Q 36 Customer case study: Scentia, AI learning tools Decidr buil t three AI learning tools inside the learning plat f orm of Scentia, an Australian education and training group. Scope of the deployment 3 AI tools buil t and embedded 200+ future skills courses in scope 3 specialist education brands Tens of thousands learners suppor t ed each year Success story Scentia delivers career-focused learning across 200+ future skills courses and a por t folio of specialist education brands, including the Australian Institute of Management, Australian Business School and the Australasian College of Heal t h and Wellness. Course questions went to academic staf f , and assessment feedback arrived only af t er final submission. Today, three Decidr tools run inside Scentia's own learning environment: a course tutor that answers questions in context, an assessment tutor that returns rubric-aligned feedback before submission, and a generator that turns course material into interactive micro-courses. 37 Scentia: what changes for students and staf f The same operating layer changes how students ask, how staf f spend their time, and how courses are produced. Students ask more, and earlier Public Q&A and passive waiting for marks becomes private, confidence- building questions and active iteration before submission. Academic staf f move to higher-value work Repetitive Q&A and low-value marking becomes higher- value teaching, feedback and coaching. Learning designers publish in one click Manually rebuilding modules in authoring tools becomes authoring in Word and publishing with one click. The institution scales its content One-size-fits-all content, with AI seen as a replacement, becomes scalable micro-courses with AI as an augmentation layer. Success story 38 Scentia: customer statement Scentia wanted AI inside the learning journey rather than bol t ed on beside it, across mul t iple education brands and a single learning plat f orm. “We wanted a par t ner we could build with for the long term. Decidr stayed focused on what mattered: practical AI features that suppor t our students.” MATT STEJER, HEAD OF PRODUCT TECHNOLOGIES & AI, SCENTIA Success story WHY SCENTIA PARTNERED WITH DECIDR Scentia needed a par t ner able to build inside its existing learning plat f orm, suppor t mul t iple education brands at once, and stay focused on what students actually need. CASE STUDY 39 “ “I f in my head I could make a proposal, it would look like that. I just never could - it required [Sugarwork] sof t ware and […] brain to put it all together.” Company CFO Subsea cable company: $3.8M of duplicated ef f or t identified The challenge 39 interviews across 26 people in three overlapping teams, turned into evidence of exactly where work was being repeated. $3.8M annual cost of duplicated ef f or t identified 480–738 duplicated hours every week across the por t folio 1% margin improvement on $1.5B contracts What Sugarwork unlocked $1.25M–$2.65M in savings from quick wins alone — standardised repor t ing, integrated risk register, combined meeting prep A CEO-ready operating model and phased plan delivered in 10 weeks, not months Captured knowledge turned anecdote into a board-grade ef f iciency case. Ef f iciency ✓ ✓ SUBSEA CABLE COMPANY SUBSEA FIBRE OPTIC CABLE SYSTEMS ~900 employees, 8 cable ships, contracts of $300M–$1.5B. Af t er a restructure, role boundaries blurred and too many people were doing par t s of the same work, but leadership couldn’t act on what they couldn’t see. Open headcount requisitions stalled. CASE STUDY 40 What Sugarwork unlocked Knowledge transfer across four technical domains during a live restructure Auto-generated onboarding wayfinding plus a conversational inter f ace to find what matters “ “I t reduced onboarding time by 70%… enabling us to bridge geographical gaps and get to deep, tribal knowledge in areas we’d had no visibility over.” Mo Chahdi, Global VP, Employee Experience Continuity Continuity protected: exper t ise retained even as people moved on. $20M client delivery value de- risked and retained $750K annualised idle-labour savings 70% faster onboarding for 100+ engineers ✓ ✓ Appen: knowledge transfer during a restructure The challenge Exper t –learner pairs ran 250+ structured sessions over 6 weeks, captured and synthesised into onboarding assets new hires could query. APPEN TECHNOLOGY · ASX:APX · 1,000+ EMPLOYEES Appen restructured, moving engineering functions across geographies. Over 100 engineers across machine learning, DevOps, data science and QA had to transfer deep, tribal knowledge in just 6 weeks — while