Here at Intercom, our mission is to make internet business personal. But in order for an internet business to be personal, it must also be possible for everyone to access.
More than one billion people worldwide live with a disability – that’s more than 15% of the global population. Without assistive technologies like screen readers, the web is inaccessible or hard for them to use. Think of it this way: it’s like entering your neighborhood coffee shop and if you’re in a wheelchair, discovering that the counter is too tall for you to reach.
“We believe internet businesses should be able to communicate with everyone, regardless of how their visitors interact with the web”
We believe businesses should be able to communicate with everyone on their website, regardless of how their visitors interact with the web. This isn’t just a company philosophy; it’s also an engineering commitment. To prioritize accessibility in our Messenger, we took a hard look at the technical improvements we needed to make and turned what were often fuzzy requirements into real, meaningful solutions.
What we achieved is making our web Messenger accessible and compliant with the Web Content Accessibility Guidelines 2.0 Level AA.
A shared framework for web accessibility
The Web Content Accessibility Guidelines (WCAG) are a shared set of technical standards that explain how to make web content accessible to people with disabilities. Its 12 guidelines are organized around four main principles, which provide the foundation for web accessibility:
- Perceivable: Users must be able to perceive the content in some way, using one or more of their senses. For instance, images that convey meaningful information should have alternative text provided.
- Operable: Users must be able to control UI elements. For example, all functionality like buttons and form elements should be accessible using keyboard controls.
- Understandable: The content must be understandable to its users. That means things like the language of the page should be detectable in the code.
- Robust: The content must be developed using well-known and adopted web standards. In other words, your code should be easily parsed and interpreted by different browsers and user agents like screen readers.
Our engineering work started by exploring the WCAG guidelines and then identifying all the areas in our web Messenger that needed improvement. As we quickly learned, turning these four principles into real solutions was simpler on paper than in practice.
Turning fuzzy requirements into real solutions
The WCAG guidelines are extensive – across the four principles, there are nearly 100 sections – and some areas are quite fuzzy. Requirements like “meaningful sequence” and “focus order” are very broad in scope, especially for applications like ours that get embedded in many different environments.
While not entirely autonomous, ML-based AIs have the potential to take over repetitive yet nuanced finance functions, as well as provide useful insights to support higher-level business development in the office of finance.
“There wasn’t always a direct or obvious correlation between the accessibility guidelines and what we needed to build”
These fuzzy requirements meant there wasn’t always a direct or obvious correlation between the WCAG guidelines and what we needed to build. We encountered issues that didn’t have clear answers online, leaving it up to us to come up with the right technical solutions.
Supervised-ML AIs – The Helpful Accountant
Basic accounting functions that are nuanced but predictably consistent are a great match for this type of AI. Basic Accounts Payable, Receivable, Payroll tasks can be completed accurately and efficiently, tailored to the needs of team members. Report-building will also become easier as this AI can gather data and pre-fill templates on a regular basis. Finance departments will have a very reliable 24/7 accountant that can perform simple tasks and find data on their behalf.
Unsupervised-ML AIs – The Predictive Data Analyst
As an AI that predicts rather than decides, Unsupervised-ML AIs can be a powerful ally in building realistic budgets and forecasts. Drawing past data out of an ERP, it can analyze existing numbers and provide finance teams with possibilities on what the future can look like. Moreover, it can offer the probability of whether specific fiscal scenarios will play out and how likely their numbers will change due to industry trends. This will give finance teams a stronger foundation on which to budget and forecast, allowing them to make better business decisions for the future.
Reinforcement-ML AIs – Advisor to the CFO
This is the AI that can hyper-focus on big picture objectives such as what steps the CFO would need to take to maximize revenue. Working in tandem with multiple data sources including the finance team and the CFO themselves, the Reinforcement-ML AI can analyze a company’s financial footsteps to determine which decisions made the biggest differences in revenue. These insights in turn can be sorted into negative and positive changes and mapped out to recommend the order and timing for these changes. All-in-all, the CFO will gain a very useful advisor who can provide timely and relevant information on important long-term decisions.
While AI technology is still in its infancy, it has the potential to greatly enhance and benefit finance & accounting teams. Here at Limelight, we have already begun exploring the potential possibilities of embedding AI in our own platform. Click here to book your demo and see Limelight in action.
