Your browser is not supported. Please update it.

9 June 2017

CollectAI pushes AI-based invoice chasing to enterprise

CollectAI ‘s CCO, Steve Emecz, presented the case for machine-learning as a debt collection tool, in a discussion Riot had at IoT Tech Expo in Berlin. The startup’s pitch is a faster way to collect the vast amount of money that is tied up in overdue invoices, using a self-learning algorithm that chases up customers and adapts to their behavior – as a service.

It’s a very interesting use of the AI-based technology, being put to use in an application that has a pretty clear use-case – and not in applications that are on the horizon. For many large businesses, such a system has huge potential improvements in operational costs and cash flow, but companies like CollectAI have an uphill struggle on their hands in order to make inroads into the giants of the industries’ market share.

Collect AI came into being via the Otto Group, a European online retailer based in Germany. Emecz explained that Otto has a history of building companies to solve its internal business problems, using these offshoots to service is group companies as well as external businesses too.

As such, CollectAI is expanding from its Otto-focused base, and hopes to move into enterprises looking to improve their cash-flow. Putting the market into context, Emecz said that around 40% of invoices in the EU are overdue, representing billions of Euros that are being chased with old systems and methods.

To counter the legacy approach, CollectAI is offering a digitized and automated system that uses machine-learning feedback to adapt its pursuit to match the customer – able to offer different communication channels (phone call, SMS, email, mail etc.), different payment methods, and the best time of day to reach out. The company pitches this as akin to having your best employee working all day, every day, and much more quickly.

Monitoring the user behavior, through things like click-rates in email, allows the AI-based system to determine the best course of action, and it can escalate appropriately if it feels like it is being ineffective. The first variable it uses is time of day, before moving onto changing the channel, and finally the tone of the message – shifting from a friendly reminder to a warning.

Emecz explained that the system helps to spot the hang-ups that prevent invoices being paid, such as users clicking through a reminder email but then running into difficulties making the payment, or users ignoring emails in the morning but being more likely to read them in the evenings. Adapting to the data is the key here, and that’s where CollectAI is selling its secret-sauce.

In our conversation, Emecz said that the system can achieve 50-80% reductions in the amount of manual work carried out by staff, and that the customer data generated also provides some pretty valuable customer insights. As for pricing, CollectAI estimates the volume of customer interactions and then prices accordingly – with Emecz saying that the cost is usually a fraction of the savings, and so very easy to justify.

Benchmarking the system using a variety of KPIs, such as time, cost, service, and returns disputes, CollectAI is confident about its value to these enterprises – especially to the likes of telcos, service operators, or utilities that need to maintain long-term positive relationships with their customers.

On this note, Emecz says that the time saved for these businesses allows them to take a more proactive approach to customer engagements – something that is vital for brand-oriented businesses. To this end, redeployed staff could be put to use in tasks that would reduce churn, such as better handling complaints and returns, or for promoting the business in new channels like social media.

Emecz also provided the example of a client with a small call-center, of around 200 staff, which had struggled to scale past that number of workers in the past. It found that the automation functions allowed it to use the same number of workers to take on new business opportunities – chasing leads, pursuing social media, and positive customer interactions.

However, Emecz noted that the integration between CollectAI’s system and the traditional business tools is a little clunky – saying that while many companies talk about API integrations (“You can’t handle the API!”), it often comes down to CSV file downloads and deliveries.

Merging new technologies like CollectAI with the old and often incredibly customized business software that runs an enterprise is not straightforward. However, the company does offer integrations with the likes of SAP, DateV and Oracle for ERP systems, as well as CRM systems like OTRS, Zendesk, and NetSuite.

Currently, CollectAI has some twenty clients, a mix of B2B and B2C business. Emecz said that the majority were B2B deals, but that the greater volume was in the B2C sector – and that the majority of the deals were based in Germany. With around 40 staff, growing quickly since its founding last year, the company was built by Liquid Labs – a company dedicated to launching Otto’s projects.