“The Innovation Lab is inriver’s speedboat for ingenuity and progress.”
May 27, 2024Viktor Bergqvist, a manager at inriver Innovation Lab, looks at the technologies that are rapidly changing the game in PIM.
“Everyone at inriver is an innovator. Our job is to accelerate that innovation.”
Advancements in technology are transforming the world of Product Information Management (PIM) at breakneck speed.
One developer at the vanguard is Viktor Bergqvist, a manager at inriver’s Innovation Lab.
He sat down with us to share some of the exciting new PIM capabilities everyone’s talking about and the impact they are having on manufacturers, brands, and retailers.
Hi Viktor. You manage a team of developers at the inriver Innovation Lab. What is that?
Innovation Lab acts as a kind of speedboat for ingenuity and technological progress at inriver.
We are a global network of teams dedicated to exploring new ideas and capabilities in the PIM market. Working closely with product managers and customers, we track trends and evaluate new types of technology to see how we can use them to enhance the PIM experience.
Everyone at inriver is an innovator in some way. But our job is to accelerate innovation, transforming fresh thinking into prototypes, putting them to the test, garnering customer feedback, and iterating fast. It’s also about knowing when to sunset an idea that isn’t solving a customer pain point.
What are the latest tech innovations that are making a big difference to PIM customers right now?
Several exciting advancements are shaking up the PIM landscape in 2024, from bulk content creation to streamlined workflows and large language model fine-tuning.
One of the new tools I’m most excited about is Batch Editing, a function within the inriver AI-powered content generation module, Inspire. It allows you to generate text that enriches more than 100 products in one go in seconds, using the same promptings. You can also translate products in bulk into multiple languages.
At the same time, we have maintained all the detailed functionality of Inspire, which leverages ChatGPT AI to generate SEO-optimized product descriptions and social posts of your desired length that are hyper-localized to target consumer groups, from fitness enthusiasts to farmers, Amazon Marketplace visitors to businesses.
To maintain human oversight, there’s a preview and validation feature that allows you to check everything is in order before you syndicate content to all your channels.
For overstretched product marketing teams responsible for creating unique and compelling content on vast numbers of products on multiple communication platforms, this is a real gamechanger. They can work much more efficiently while still maintaining visibility and quality control over what they push to market.
That sounds fantastic. You also mentioned streamlined workflows. Tell us about that.
Sure. Workflows is a new inriver feature unveiled at PIMpoint Americas 2024 that makes it easier for PIM users to focus on specific assignments and collaborate across teams.
It allows workflow creators to define process steps for specific tasks in the product lifecycle, such as enrichment or evaluation, and translate those steps into workflows that match their specific needs. Instead of seeing thousands of irrelevant fields that could distract or confuse, users see just a subset of the platform’s features tailored to their task at hand, allowing them to work more confidently and fluently. Meanwhile, workflow creators can more effectively track the status of each task.
In addition to this are the new extended capabilities related to workspaces. These allow inriver partners to develop specific workflow workspaces that can target specific customer needs. These can then be solved more effectively and efficiently in batch mode.
And large language model tuning?
When we talk about AI-powered product information management, we’re really talking about PIM platforms that have integrated large language models (LLMs). This is the type of AI that is trained on enormous datasets and can process information at superhuman speed and then generate human-quality content. Although the big LLMs like ChatGPT and Google Gemini are getting better all the time at generating multilingual content in nuanced and bespoke ways, they are not necessarily tuned in to your company’s unique tone of voice.
At inriver, we’ve recently introduced a Fine Tuning feature that allows you to input data specific to your business so that you can instruct the model to your requirements. It means you essentially create your own AI model, fine-tuned to use the on-brand turns-of-phrase, terminology, and preferred spelling that make your products stand out in the marketplace. This is an area that inriver is continuing to investigate and develop.
Standing out in the marketplace is certainly important. What makes the inriver PIM stand out?
The inriver PIM platform is the most complete on the market, able to manage data across the entire product journey, from data onboarding to robust data-driven evaluation on how products are performing on the digital shelf. Here too, AI is helping. Our AI-powered Evaluate module keeps a smart eye on how products are displayed and reports back on whether customers are inspired and converting.
Another hallmark of inriver is our fully extensible data model that means our platform can scale to our customers’ needs as their product lines, suppliers, and channels constantly shift and grow. And we have an ever-growing network of technology partners and alliances providing complementary products and platforms. So, when we’re working on a new technological capability at inriver Innovation Lab, we’re not just thinking of how we can use it in our own software but how our customers and partners can utilize it through the extension capabilities we have.
Tell us a bit about your career to date and what prompted your move to inriver.
While studying computer engineering in the early 2000s, I took part in a research program about how humans and computers interact in the IT space. Having graduated in the wake of the dot-com crash, I started off working as a marketing manager, before moving into developing B2B and B2C e-commerce sites. This taught me the vital importance of having structured data in place and, in fact, I even built my own PIM system to house all the information required to market and sell our products. That realization and experience inevitably brought me to inriver’s door.
Finally Viktor, what are you working on in the Innovation Lab right now? It must be an exciting time to be working at the cutting edge of PIM technology.
It certainly is. A key focus area for our team right now is developing the Inspire module of inriver Enrich to make sure the prompts that generate these dynamic product stories are as fine-tuned as we can make them. I think the next big leaps forward will be in multimodal LLMs that can process vast amounts of inputs from multiple formats, including text, images, audio, PDFs, and code. The result will be a level of hyper-personalized and engaging product storytelling that surpasses anything we’ve seen so far.
LLMs have supercharged PIM development, giving us powerful tools that were unimaginable before. Where once we spent days writing complex code by hand, we can now focus on providing the right training data and context for the LLM. And where we used to bring a new piece of software to the market say once a year, now we are rapidly iterating all the time, resulting in smaller releases that incrementally improve the customer experience on a regular basis. The biggest challenge of my job is keeping up with the pace of change, but that’s also why I love it.
Thanks, Viktor. Keep innovating!
want to see the inriver PIM in action?
Schedule a personalized, guided demo with an inriver expert today to see how the inriver PIM can get more value from your product information.