The shift towards short-form video has revolutionised content consumption, and while creating unique, personalised video experiences manually remains impractical, recent breakthroughs in generative AI now deliver commercially viable avatars, making person-to-person connection scalable like never before.
The opportunities around personalised video have become a common theme in the client workshops Sputnik Digital has held with its high-profile clients in the insurance, healthcare and utilities industries in recent times.
Their prevalence prompted MD Andy Nicol to task its innovations team with developing a tool capable of delivering exactly this. Enter Hollr.io.
鈥淭he idea grew over a year or so as we held workshops with clients to explore their plans and hesitations around AI adoption ,鈥� he tells 老九品茶Cloud.
鈥淭he obstacles to proceeding across the board was having confidence in the technical capability, quality, scalability and cost.
鈥淥ur innovations team decided to build Hollr.io to allow us to better answer these questions. They have done a huge amount of work in understanding the AI avatar space, as well as all of the other elements that are needed to successfully implement a personalised video solution.
鈥淭he process iterated between speaking with clients and trying things out until we pulled together a use case, working tech and a splash of UX (user experience) into an MVP (minimum viable product).
鈥�Hollr.io is a consultancy service – and it鈥檚 ready to go.鈥�
I tried out the demo tool – you can do so too by clicking – and within minutes was sent a slick video message of a human-like AI rep delivering my personalised insurance quote renewal.
In a live environment, much of the data I entered – such as name and renewal price – would be merged from the client鈥檚 CRM, automating a great deal of the process while generating that crucial connection with the customer.
鈥淐lient implementations will vary depending on their digital readiness, and there is lots to do including scripting videos, understanding CRM data, compliance approval and video distribution channels,鈥� adds Nicol. 鈥淏ut we could have a client solution implemented in a month or two.鈥�
My video referenced a theft claim in the previous 12 months. 鈥淓very implementation is bespoke,鈥� says Nicol.
鈥淚ncreases in insurance premiums are clearly a concern for customers, and personalised video is an opportunity to explain the factors involved to soften the blow.
鈥淲e work with clients to understand their industry, their script, their brand, their data – and for which of their customers a personalised video is most appropriate.
鈥淲ith current AI costs, the full journey may cost 拢3-5 per message. At those levels it might be able to replace some direct mail where a first class stamp alone is 拢1.70, but with costs likely聽 to come down over time, we expect adoption of personalised video to spread. 鈥�
As well as the consultancy element, AI CPU time is not cheap – so Hollr.io will be a relatively premium service when it is taken to market in earnest.
鈥淚t would cost in the region of 拢75k to put together an end-to-end proof of concept to send to a small but statistically significant number of customers – but otherwise it鈥檚 ready to go,鈥� says Nicol.
鈥淚t鈥檚 live, but our plans for it are fairly early stage. We might keep it as an agency service within Sputnik, or we might spin it out.
鈥淎I is a changing landscape, and keeping on top of it will take some effort, which we鈥檙e geared up to do.鈥�