Subscribe to Impira's blog
Stay up to date with all things Impira, automation, document processing, and industry best practices.
At Impira, we are on a mission to free people from monotonous tasks to create space for more meaningful work. I’m proud of everything we did in 2021 to further our mission – both internally for our team and externally for our customers.
This year was transformational for us. We kicked it off by launching our self-service product and spent the year learning together with our customers. Since then, we've seen over 10,000 organizations sign up for Impira with the goal of making their work easier through intelligent data processing. Along the way, we hired several key team members in multiple geographies and countries and had several opportunities to spend some quality time together in person.
A shift to self-service
When we started Impira in 2017, the machine learning (ML) space was nascent and neither vendors nor customers knew how to create or consume ML-oriented products that could enable business automation (plenty has been written about this). Since then, we’ve had the opportunity to work with several thought leaders, from growing startups like Dashy Dash to large enterprises like Colgate-Palmolive and learned a few key lessons:
- Companies tend to have many “small data” problems that benefit from automation. For example, a process might involve reading thousands of financial statements a month and compiling the data into a database. Each process requires a slightly different ML model.
- Data operations professionals have the budget and skillset to drive automation within savvy organizations. They’re experts at bringing together vendors, developers, and subject matter experts to solve automation problems.
- Automation involves more than just data extraction. Yes, if you’re working with an invoice, you need to pull the line items out of it to automate the processes surrounding it. However, you might also need to join them with another dataset, integrate with accounting software, or both.
The challenges of last year opened our eyes, and we realized that the only way to serve organizations who need intelligent document processing in a scalable and sustainable way would be to deliver a self-service product. Why? Because otherwise we’d be training custom machine learning models for each customer, and therefore not building a true software company. Because of our shift to self-service, we were able to put our heads down and develop some truly innovative technology: a one-shot learning model to extract data from documents, incremental materialization to support real-time training, and countless user experience improvements to make the product easy to use for any type of user. With these features, customers can train their own machine learning models to extract data from documents in less than five minutes without writing a single line of code. In February of this year, we launched these technologies behind a simple, self-service signup on impira.com. Since then, 10,000+ organizations have used Impira to process data from over a quarter million documents. That’s over 29,000 hours saved for our users!
Throughout 2021, we’ve iterated on our core product and its feature set alongside our customers. Along the way, we’ve shipped some incredible features like table extraction, review workflow, webhooks, Snowflake integration, NLP-based extraction, in-app file modification, lots of new IQL syntax, and infrastructure that can keep up with our constantly growing scale.
Product-led growth in our DNA
Technology is only one part of delivering a self-service product. To drive a go-to-market successfully, you need the right people and processes in place. This year, we made two key hires: Lorilyn McCue (Head of Product) and Megan Dorcey (Head of Marketing). Lorilyn joined us from Slack and Megan from Stack Overflow, two companies that have mastered product-led growth. Through their leadership, we’ve made a number of investments to drive growth, including:
- A heavy emphasis on in-product, self-service onboarding
- A new consumption-based pricing model that enables customers to start small and scale with their use of the product
- A sales process that emphasizes customer success with the product before engaging in broader business discussions
- A large repository of publicly available content including blog posts and documentation designed to make customers more successful
- A metrics framework that allows us to measure the impact of our experiments
Embracing a hybrid team
Thanks to public progress in tackling the COVID-19 pandemic and widespread availability of the vaccine, we were able to partially reopen our office in May. In doing so, we officially became a hybrid company.
Several of our key employees live outside of the San Francisco Bay area where we are headquartered, and we have no intention of asking them to relocate. At the same time, many folks in San Francisco, including myself, missed in-person collaboration and the office environment. We felt the best way to operate going forward was with a fully-optional hybrid office environment.
As you can imagine, creating a well-functioning hybrid office environment is a challenge to implement so we approached it from multiple angles. First, we decided that office attendance was completely optional for even San Francisco-based team members. Everyone has different working styles and personal circumstances, so it didn’t make sense to wash over those unique needs simply because a team member lived in close proximity to the office. If we wouldn’t be asking remote team members to move, we wouldn’t force local team members to come in if they didn’t want to. In practice, we found that some local folks prefer to primarily work from home, some like to come into the office two to three days a week, and some come in every day.
Second, for meetings with remote attendees, we ask each individual attendee (in-office and remote) to dial-in to create an even playing field amongst participants. We’ll continue to experiment with different ideas that promote a positive and productive meeting experience for all our team members. In terms of live collaboration, we still haven’t cracked the code on how to create a great remote whiteboarding experience, so please let us know if you have any ideas!
Finally, we reinvested the budget that was originally spent on in-office lunches (and other office expenses) into a flexible travel budget that teams could use for frequent offsite gatherings. This allowed us to try out all kinds of activities including a get together for the entire engineering team, small team roadmap offsite meetings, and even a virtual wine-tasting for the marketing team.
Remote hiring has been a boon for diversity and inclusion at Impira. A fully remote interviewing process has created opportunities to be more objective. For example, many of our interviews now contain anonymous take-home assignments that are reviewed asynchronously rather than with in-person conversations. As a result, we've become a truly hybrid team. We've recruited top-tier talent from across the country, which has supported our diversity goals overall. We benefit from a wider range of perspectives and insights than we could by simply hiring within the San Francisco market.
This past year was tough in so many ways, but we worked hard and leveraged difficult circumstances as opportunities to pivot to something better. This makes us really excited about the upcoming year. At the end of the day, the Impira team feels more productive, energized, and dynamic than ever.
Looking to 2022
As we look to the year ahead, we’re going after some ambitious goals by allowing ourselves to dream big and aim high. Our marketing and sales team will collaborate closely with our customers because listening and responding to their needs is an amazing way to grow and reach new heights. Our product team will be laser-focused on improving our core capability. This means quickly and accurately extracting data from documents with as few as 1-10 example labels. They’ll also make Impira easier to implement into customer environments through more integrations, query capabilities, and API improvements.
On the business side, we’ll be investing in more use-case specific content and resources to help customers get started faster, especially when working with financial documents, purchase orders, or invoices. Finally, we’ll be investing in our user community so that customers can share best practices, ask questions, and collaborate on their automation journeys.
With this year came great challenges, but I’m proud of our team for navigating them well. We’ve worked hard (and so have you), so we’re all taking the final week of the year to spend with family and friends and recharge for the year ahead. See you all in 2022.