Five UK startups pioneer AI across the Consumer Fintech Spectrum

ai-finance2040-620x354

Image Source

AI is no longer a differentiator amongst startups, it has become a default feature that most firms will need to have as one of their core capabilities. UK has never been short of AI success stories, with one of the first being Deepmind, that was acquired by Google in 2014. The Fintech wave was just getting started then, and there have been some good tales in the UK-AI-Consumer Fintech space, across various sub-clusters.

Personal Finance Management: Cleo

Cleo, an AI assistant that helps customers manage personal finance was founded in 2015, and after two years of work was launched commercially this year. The AI assistant taps into consumers’ bank accounts and helps them save money. Since launch, Cleo have close to £400 Million worth of assets under management.

They have integration with Facebook messenger to manage payments, and have an interface that has managed to retain 70% of their users even after 3 months of signing up – thats pretty impressive. VCs have been pouring their money into PFMs in anticipation of PSD2, and Cleo has recently closed a £2 Million round in which Skype’s Niklas Zennstrom participated.

RegTech: Onfido 

London based Onfido, was recently announced by the World Economic Forum as one of the “Technology Pioneers”. They help verify people’s identities digitally. Founded by three entrepreneurs from Oxford University, Onfido uses AI to perform background checks and spot frauds. They validate a user’s identity by comparing biometrics with identity documents. Identities can then be cross-referenced against international credit and watchlist databases

They operate across the globe in about 195 countries with a team of 150 employees, and support close to 1500 businesses. They have recently managed $30 Million in funding from Salesforce ventures and IdInvest Partners.

InsurTech: Tractable

AI within Insurance has seen some really good use cases in the last couple of years. Tractable is a deep learning startup specializing in computer vision to solve specific high impact problems. Their imagery algorithm can be used to quickly perform visual inspection of an accident and provide data digitally to support insurance claims.

In the past where even minor accidents took a few days if not weeks to settle, AI programmes can quickly scan the damage and digitally assess the cost to fix them. The deep learning capability that they have built in collaboration with Cambridge Machine Learning Group, the Visual Geometry Group (Oxford University) and the Neuroscience unit at UCL, can now assess damage to a vehicle more accurately than a human expert. Tractable closed their Series A round of $8 Million last month.

Mortgages: Habito

I wrote about Habito in detail in one of my previous posts. They are an AI capable online Mortgage broker based out of London. Their AI algorithm helps identify the best deal in the market, and also get real time approvals for mortgages. Customers have access to over 60 mortgage lenders through their platform.

Earlier this year, they closed a £5.5 Million funding round which they would be using to build an end to end AI enabled real time mortgage experience.

Credit Scoring: Aire

Aire offers AI enabled alternative credit scoring capability to lenders, where they get insights into borrowers with thin credit files. Aire has an interactive virtual questionnaire that provides insights on top of traditional credit data. Through this new capability, lenders on average have managed to increase credit approvals by upto 14% without increasing risk exposure.

Aire has recently signed partnerships with Zopa and Toyota Financial Services, and closed a funding round of $5 Million last month.

AI for decades has been a fairy tale as it lacked the data volumes and data quality to provide the right insights. However, with Social Media, Open data, and PSD2 providing a firm footing this time, the API era should see some Consumer Fintech AI success stories over the next few years.


Arunkumar Krishnakumar is a Fintech thought leader and an investor. 

Get fresh daily insights from an amazing team of Fintech thought leaders around the world. Ride the Fintech wave by reading us daily in your email.


 

Key Trends in Q2 Fintech M&A activity – Payments lead the way

Fintech deal activity hit a peak in Q4 2015, and as discussed in a previous post, steadily went down through most of last year. However, this year after a good start in Q1, there was a strong rebound in Q2 2017, and recent news have been pointing to some big ticket deals happening within the payments space.

As per KPMG’s quarterly report, globally Fintech investments hit a healthy $8.4 Billion across 293 deals. Rebounds were particularly noticeable in both Europe and UK. Fintechs in Europe managed to attract $2 Billion (in investments) in Q2, which is more than double the Q1 number ($880 Million).

VCQ2-1

Some of the key trends from the report,

  • Corporate Venture Capital continue to increase their involvement in Fintech deals
  • Asia sees a dip in Q2 investments due to low China deal activity
  • Regtech deals could create a record year 2017. At the current pace its likely to surpass 2015 and 2016 activity (in size and count)
  • Focus moves from B2C (customer experience) to B2B (mid and back office efficiencies)

VCQ2-2

Apart from the VC activity, Private Equity firms have turned their attention to Payments, as the deal sizes within payments start to increase. In the last eight weeks we have had some M&As and private equity deals announced within payments.

  • Igenico acquires Bambora for $1.5 Billion. This happened after Ingenico tried a hostile takeover of WorldPay assets
  • Worldpay merged with Vantiv with a £9.1 Billion deal. The new firm will be jointly led by Vantiv’s Charles Drucker and Worldpay’s Philip Jansen.
  • Worldline acquires Digital River World Payments and First Data Baltics, giving them operational positions in the Nordics and in the Baltics.
  • Visa invested in Klarna – how much they invested and at what Valuation is not disclosed.
  • Blackstone and CVC announce acquisition of Paysafe for £2.9 Billion

These are some of the top stories, but the key takeaway is that money is flowing the Fintech way, again!! Both in the VC and the PE space!!


Arunkumar Krishnakumar is a Fintech thought leader and an investor. 

Get fresh daily insights from an amazing team of Fintech thought leaders around the world. Ride the Fintech wave by reading us daily in your email.


 

GDPR vs PSD2 – Banks may abandon PSD2 due to conflicting policies

KEEP-CALM-AND-PREPARE-FOR-GDPR

Image Source

About a year ago, Bernard had written a post on PSD2, and discussed different levels of maturity in regulations. He highlighted that PSD2 was a regulation meant to open up the market for innovative consumer banking use cases and solutions. However, the same regulator (EBA) have set a timeline for General Data Protection Regulation (GDPR) in 2018 alongside PSD2.

We have discussed PSD2 and its implications for banks, fintech firms and consumers at length in the past. So, let me focus on GDPR and what it means to firms and consumers. The purpose of GDPR is to ensure consumers give informed consent before companies can share their personal data with third parties. Pre-ticked check boxes and inactivity from consumers can no longer be assumed as their consent to data sharing post GDPR.

Unlike PSD2, GDPR applies to businesses in the EU processing consumer data, not just Financial services firms. Also, for non-EU businesses GDPR applies, if an EU resident’s personal data is processed in connection with goods/services offered.

The Data Protection Act (DPA) provided consumers with right of subject access – which meant consumers can request a company for data that the firm had collected about them. Currently many businesses charge a fee to provide this data to consumers, but post GDPR, firms can’t charge this fee.

As consumers, we can instruct firms when to collect our data and stay on top of it using the right of subject access. Now what does this have to do with PSD2? PSD2’s purpose is to enable consumer data sharing, where as GDPR’s purpose seems to be to try and cut down on data sharing.

GDPR

PSD2 is about financial services firms sharing customer data with third parties who they may not necessarily have a contractual agreement with. These third parties may then come up with innovative use cases by processing consumer data.

So, to be compliant with PSD2, banks should ask for customer’s consent to share their data with third parties. But to be compliant with GDPR, data processing by third parties will also need explicit customer consent. How is a bank supposed to be responsible for the processing of consumer data performed by a third party, it has no contractual agreement with?

While this hasn’t been explicitly mentioned as a process required to be GDPR compliant, my guess is, it would be upon the Banks to ensure third parties (that they share consumer data with) have consumers’ consent to process their data.

Unlike PSD2, that doesn’t have any punitive charges, violation of GDPR might result in a fine of upto €20 Million or 4% of Global turnover. And knowing the way banks deal with regulatory compliance, nothing motivates them more than a fine hanging over their heads.

This means, where there are conflicting regulations, and lack of clarity on a standard approach to data sharing, banks will focus completely on implementing the punitive GDPR. In someways, GDPR may also become an excuse for banks for not implementing PSD2 and avoid sharing what they feel is their asset – consumer data. Watch this space!!


Arunkumar Krishnakumar is a Fintech thought leader and an investor. 

Get fresh daily insights from an amazing team of Fintech thought leaders around the world. Ride the Fintech wave by reading us daily in your email.


 

Fraud Detection using AI and Mastercard’s acquisition spree

“Progress is made by the improvement of people, not the improvement of machines.”

fraud-detection

Image Source

As more consumers turn to digital banking for their everyday transactions they will generate huge amounts of data that banks can use to identify trends and highlight suspicious behavior.

As digital transaction volumes increase, and real-time payments become the norm, banking solutions to identify frauds are often inadequate. In most cases these systems will need to determine if a transaction is genuine or not in a fraction of a second. Thanks to the AI wave, as fraudsters get better, machines spotting them get better too.

Cybercrime is estimated to cost the global economy 400 billion dollars. Credit card fraud accounts for a large proportion of this cost. Artificial Intelligence (AI) can provide faster, cheaper and more accurate fraud detection.

Some of the key considerations of a payments infrastructure (using AI) while solving the fraud detection problem are,

  1. Initiate the payments safely
  2. Handle billions of transactions
  3. Identify relationships through graph maps
  4. Social media integration and Sentiment analysis
  5. Behavioural analysis
  6. Adapt quickly as fraudsters evolve their modus operandi

An AI system can use thousands of data points in every transaction and do a fuzzy lookup to billions of other transactions to identify patterns, coincidences and anomalies.

Most payment giants are increasingly turning to AI and Mastercard is no exception. They have been acquiring firms focusing on fraud detection as AI deal activity hit all time highs in Q1 2017. In March 2017 Mastercard announced the acquisition of NuData Security to deliver online and mobile anti-fraud solutions using session and biometric indicators.

AI_MnA_Q1-17_2


“Our unprecedented use of artificial intelligence on our network is already proving successful. With the acquisition of Brighterion, we will further extend our capabilities to support the consumer experience.”

– Ajay Bhalla, President of Enterprise risk and security for Mastercard


 

Earlier this month, Mastercard announced the acquisition of Brighterion. Brighterion’s portfolio of AI and machine learning technologies provide real-time intelligence from all data sources regardless of type, complexity and volume. Its smart agent technology will be added to Mastercard’s suite of security products already using AI.

“Progress is made by the improvement of people through the improvement of machines.”


Arunkumar Krishnakumar is a Fintech thought leader and an investor. 

Get fresh daily insights from an amazing team of Fintech thought leaders around the world. Ride the Fintech wave by reading us daily in your email.


 

China and India – bright spots in Corporate Venture Capital recovery

cvc

Image Source

Corporate Venture Capital (CVC) has seen a steady rise for the last few years. Between 2011 and 2016 the number of large corporates establishing their own Venture Capital capability nearly tripled. Last year CVCs participated in about 40% of VC deals that happened in Asia. With major global events such as Brexit and Trump dampening investor appetite in VC funds, CVC deal volumes saw a global dip of 2% last year. However, there are data points indicating its likely to take off in a big way this year.

The CVC world has had challenges due to the compensation structure for the Partners and lack of nimble decision making capabilities. Partners at CVCs have traditionally not been compensated as well as a Partner in an independent VC fund. Also, if the VC arm of a Corporate cannot make independent decisions in quick time, it affects both the startups and the corporates. Sometimes CVCs lose good deals due to their lack of agility, but often they hurt startups by making them wait for investment decisions, and turn them down after a few months of due diligence.

Global KPMG

The brighter side of the deal is that, once the deal has gone through, the Corporate can be a massive launchpad for the startup. And for this very reason, Startups seem to be more forgiving of the red tape and the bureaucracy that they have to go through to get the deal closed.

That said, there is no denying that CVCs have evolved their models over the last few years, and are here to stay. In Asia, Softbank announced the launch of their $100 Billion fund, with $25 Billion of their skin in the game and Apple contributing $1 Billion, the fund has seen good traction since launch. The other big announcements were Baidu’s $3 Billion fund and Samsung’s $1 Billion fund.

Chart_01c

CVCs in India had a good year 2016, until Q4, where only 4 deals were closed. However the general trend last year was that the deal count went up, but the size of the deals came down compared to 2015. This is in contrast to most other top VC ecosystems in US, China and Europe where capital was moving towards more matured (growth stage) firms. VCs in India preferred smaller sized deals in early stage firms, resulting in a higher deal count. The only other region that had had CVC investments go up in the first half of last year was UK, but there was a slow down in H2 2016 post Brexit.

China on the other hand, had a slowdown in CVC investments last year after a strong 2015. This trend is expected to change because, between China and the US there were about 53 new CVCs that were launched last year. They are expected to get more active this year. Also, many VCs and CVCs are eyeing China for Fintech deals in a big way since the start of 2017.

China

Image Source

If CVCs in India perform anywhere close to what they did in Q1 and Q2 2016 and if the new CVCs in China start deploying, we are likely to see a CVC recovery. With VC investments shrinking globally over the last eight years, and CVC’s slice of the VC pie at an all time high (17%), the recovery of CVC might just be what the VC industry needs.


Arunkumar Krishnakumar is a Fintech thought leader and an investor. 

Get fresh daily insights from an amazing team of Fintech thought leaders around the world. Ride the Fintech wave by reading us daily in your email.


 

 

 

 

From a Blockchain based to a Blockchain inspired world, SWIFT could deliver verdict at Sibos

test-hand-chains

Image Source

This time last year, the dust hadn’t settled on the Blockchain hype, and several key players within Fintech and Financial Services were quite upbeat about the possibilities. However, as results of PoCs from various consortiums, central banks and payment providers emerged, the results were mixed. Daily Fintech covered an article on R3’s miseries towards the end of last year when Goldman Sachs left the consortium.

Since then, R3 publicly moved away from Blockchain, into a Blockchain inspired world using an open source distributed ledger named Corda. The R3 consortium lost three major banks towards the end of last year. This is vastly attributed to the fact that they chose to move away from a pure Blockchain implementation to a Distributed Ledger implementation for Corda.

The three banks Goldman Sachs, Santander and JP Morgan left the consortium and invested in Axoni that was a pure Blockchain firm. It got worse when R3 blogged that they were not a Blockchain firm, and had always been a distributed ledger company and got trolled on social media for that.

This was shortly followed by the news that SWIFT had launched its inter-bank payments platform that it believed would be the future of its cross border payments platform. The platform was called GPI (Global Payments Innovation), and had a founding consortium of 12 global banks. The GPI, at that time was based out of traditional technologies and not Blockchain. However, earlier this month, SWIFT announced that GPI was being beta tested on Blockchain with 22 new banks validating the system. Verdict on this PoC is going to be at Sibos later this year.

Swift GPI

Image Source

Apart from this, the Bank of England (BoE) haven’t delivered a conclusive verdict on the PoC with Ripple for Cross border payments. The detailed report on the PoC was released earlier this month. The key message was:

” Cross-border payments when applied to wholesale markets present different challenges than when compared with retail and corporate transactions, which the Ripple product is designed to handle. The availability of liquidity is one such challenge, and the PoC allowed the Bank and Ripple to begin exploring these questions. “

In other words “Ripple’s solution wasn’t fit for purpose”, although Ripple chose to see it differently.  A few days later, Ripple announced that a pure Blockchain based approach was not scalable for banks and advocated a “Hybrid approach”.

Wearing my technology hat on, I see some fundamental lessons here, and I may be repeating what has been so often mentioned.

  • Find technologies that can solve your problems – it may not have to be Blockchain.
  • Do not interchangeably use Decentralised Ledgers and Blockchains. You can photocopy on a Canon machine too (not just on Xerox).
  • Innovation doesn’t always have to be on sexy technology. SQL Server and Oracle can do the job too.
  • Simplicity is often overlooked and massively underrated.

I believe that SWIFT’s announcement of the results of their Blockchain PoC at Sibos could provide a decisive direction for Blockchain in Financial Services/Payments. And it might well be “Let’s Move On”.


Arunkumar Krishnakumar is a Fintech thought leader and an investor. 

Get fresh daily insights from an amazing team of Fintech thought leaders around the world. Ride the Fintech wave by reading us daily in your email.


 

India GST – One Nation One Tax Model could be a boost for Start ups

image1

Image Source

The Goods and Service Tax (GST) was implemented by India on 01 July, in line with a One Nation One Tax Model. GST is a single tax on the supply of goods and services, right from the manufacturer to the consumer. Historically the taxation of goods and services in India followed its federal structure, with various state regimes co-existing with the central tax rules. One of the key advantages of this unified tax model is to ensure businesses could have tax neutrality wherever they do business in the country. Such a big revamp doesn’t happen without its challenges, however the benefits are immense.

GST is a mechanism of seamless tax-credits throughout the value-chain, and across boundaries of States. This doesn’t mean that there wouldn’t be a state tax. There will be two components of GST – Central GST (CGST) and State GST (SGST). Centre would levy and collect Central Goods and Services Tax (CGST), and States would levy and collect the State Goods and Services Tax (SGST) on all transactions within a State. The input tax credit of CGST would be available for discharging the CGST liability on the output at each stage. Similarly, the credit of SGST paid on inputs would be allowed for paying the SGST on output.

 

Image result for GST India

Image Source

At the heart of the GST  implementation is a state of the art IT infrastructure. The tax-payer gets a single GST portal and can manage all their GST related reporting. The system supports connecting through mobile, tablets or even through API integration. The system also provides end to end automation for tax authorities, with capabilities such as reporting and analytics.

In the past, businesses could tell their auditors the amount of tax they wanted to pay for the year, and the books were prepared accordingly. That is no longer possible. For every transaction reported by a business, there would be need to be a matching transaction from a supplier or a buyer. However, this also removes the bureaucracy and the complexities for businesses that constantly have to do inter-state transactions.

For an SME that is looking to VAT register across multiple states, GST allows a simple central process of registering itself. They don’t have to navigate through the complexities of different state tax regimes anymore.

GST also extends the basic exemption limit to pay VAT from a previous Rs.5 Lakh per annum to Rs.20 Lakhs per annum. These rules are slightly different in some states where the exemption is at Rs.10 Lakhs.

GST views goods and services components as one and the same, which means there is no confusion around what is a good vs service. This not only simplifies the process massively for the tax payers, but will also stop instances where accountants used the ambiguity between the two (Goods and Services) to evade taxes.

However, more importantly in my view, start-ups/SMEs can expand operations across multiple states without the burden of tax on interstate sales. With GST, tax credits would be transferred irrespective of where the transaction happened in the country. This is a whole new landscape that businesses would exploit.


Arunkumar Krishnakumar is a Fintech thought-leader and an investor. 

Get fresh daily insights from an amazing team of Fintech thought leaders around the world. Ride the Fintech wave by reading us daily in your email.