HIVEMIND

HIVEMIND EFTA00803391 Software to build, clean and enrich data sets. HIVEMIND We specialise in generating structured data from an unstructured world. Datasets as diverse as the content of documents or videos, group opinions, or the likelihood of future events. We believe humans and computers complement each other powerfully, and that together they can tackle problems neither can complete alone. EFTA00803392 The problem Data-driven business decisions require structured data. 80% of data is unstructured', much more potential data is unrecorded. Transformative insights remain inaccessible to businesses within this raw material. Refining it into valuable data assets is hard to do flexibly, accurately and at scale. 1 Gartner: EFTA00803393 The solution Hivemind co-ordinates cognitive and automated processes to allow you to unlock the insights hidden in unstructured or unrecorded data. Break down big problems into bite-sized questions of parsing and judgement Distribute them either to humans or machines as appropriate Intelligently aggregate the responses to build data sets and ensure data quality Chain tasks together into sophisticated workflows to tackle hard data problems EFTA00803394 The impact Create bespoke datasets pertinent to your problems Transform your internal documents intq valuable assets I Clean and enrich your existing structured data Systematise manual data workflows and reduce senior staff time on basic data tasks Predict future outcomes by aggregating the knowledge of global experts or your own staff EFTA00803395 The product 0 A cloud-based platform with two interfaces: one for task creation, one for task completion. Hivemind Studio Create, monitor, download Hivemind Workbench Collect EFTA00803396 Financial research groups and the data vendors feeding that research 250 hedge funds > .$51)AUM -200 asset managers >$5Ob AUM -250 "alternative" & market data vendors Target markets Data management operations Cross sector applications: legal, across all financial firms medical, retail, tech - and many others Annual spend on financial market data and news > $27b By the end of 2O19, 9O% of large organisations will have a Chief Data Officer ( ) EFTA00803397 Business model Low barrier to entry. Cost scales with time, users, features and support. Annual License Fee (£000s) A 240+ - 120 24 O - Lite Core Pro features Dedicated support Pro Enterprise features Dedicated support Custom deployment Unlimited use Enterprise 250 5,000 25,000 Unlimited 1000 hr increase £5000 p/a Usage (hours) EFTA00803398 At Winton... 4 YEARS IN OPERATION 30+ BESPOKE DATASETS 8.5 million+ COMPLETED TASKS 175+ \ MAN-YEARS OF WORK HIVEMIND so far Since spin out 4 WEEKS IN OPERATION 14 QUALIFIED PROSPECT + 6 TRIALS UNDERWAY (0) In the press... The Economist FT Live... Credit Suisse Prime Services Conference 2017 Morgan Stanley European Quantamental Conference 2018 SBAI s Artificial Intelligence Roundtable 2018, New York Alpha FMC & Illuminate Financial Fintech Showcase, 2018 The Technical Analyst's Alternative Data 2018 Event (forthcoming) SBAI s Artificial Intelligence Roundtable 2018, London (forthcoming) Manager Magazin Awards Event 2018, Frankfurt (forthcoming) „..) EFTA00803399 Challenges Competition Human as parser • Figure Eight • Amazon Mechanical Turk Robotic process automation • Automation Anywhere • WorkFusion Human as expert • Lumenogic • Consensus Point Operational Challenges Distribution channels • Currently no distribution partners Burden of consultancy • Learning curve of using Hivemind Finance and the cloud • How do we convince a naturally conservative industry to embrace a cloud-based solution? Workforce provision • Develop partnerships with work- outsourcers (e.g. Cloud Factory) and improve integration with crowd source environments (e.g. Mechanical Turk) EFTA00803400 What makes H IVEM IND different? CORE CONCEPT • Focus on using computers to help humans solve hard problems more easily, rather than on trying to replicate or replace human intelligence. FLEXIBILITY • Not tied to a specific workforce • Deals with wide-range of use cases across a data scientist's workload • Simple integration with a client's internal workflow SOPHISTICATION • Built for complicated workflows and dense, heterogeneous data sources • Sophisticated aggregation of human judgement and expertise • Validated through four years of use by demanding data organisation FUNDAMENTAL DATA VALUES • A practical solution for both big projects and daily workload • No reliance on buzzwords • Data quality is more important than data size EFTA00803401 Team CEO Dan Mitchell • Director, Head of Research Data @ Winton • Prev. Oxford Uni DATA SCIENCE Christian Gilson • VP, Data Scientist @ Winton ENGINEERING Alex Dawes • VP, Technology @ Winton • Prev. Oxford Uni Mark Roulston Alex Taroghion CHIEF REVENUE OFFICER • MD, Research @ Winton • VP, Engineering @ Winton Henrik Grunditz • Prev. Met. Office, Penn State Uni, Caltech, Cambridge Uni • Prev. RBS, Barclays • SVP, Bus. Dev. @ Winton • Prev. MSCI, Accenture, Imperial College Riaz Karim • VP, Engineering @ Winton • Prev. JP Morgan, Goldman Sachs The team has worked together at Winton and on Hivemind over the last 4 years EFTA00803402 Projected financials (£000s) (O) EOY 20181 EOY 2019 EOY 2020 EOY 2021 EOY 2022 Enterprise Clients (- 240k) 1 2 3 5 12 Professional Clients (- 120k) 2 4 10 25 65 Core Clients (- 24k) 3 6 16 38 100 Revenue 240 906 1830 5112 13080 Revenue Growth (%) 0 139 102 179 175 Total Costs -562 -2416 -3738 -6015 -9323 EBIT -322 -1510 -1908 -903 3757 Capital Raised 394.5 Secured Loan 250.0 -279.4 Cash Position 322.3 -1187.5 -3375.2 -4277.8 -827.2 Headcount 7 18 30 40 57 Avg Rev / Employee 34.3 50.3 61.0 127.8 229.5 1 EOY 2018 numbers based on 6 months to Dec EFTA00803403 Funding Seeking a £5m funding round to implement initial strategy PEOPLE Management / advisor to supplement business experience Sales team core to driving business expansion, including in the US. Data scientists to address "consultancy burden" and build / support value-add services. Developers to expand the platform PARTNERSHIPS Distribution partners to help scale the sales effort. Workforce provider to answer the initial "who's going to do it- question and allow us to provide an end-to-end service if required Data aggregation platform to provide distributional services to new "alternative" data vendors PRODUCT A more feature-rich platform including continued development of our prediction market service Focus on ease of use to avoid initial "consultancy burden" on our data scientists Cloud security requirements for large sell-side organisations EFTA00803404