Bridging the AI Gap: How GenAI Enhances Non-AI Native SaaS Products
Three use cases for GenAI in traditional SaaS products
The past couple of weeks I tested around 20 non-“AI native” SaaS products that are now implementing AI powered features (most are in beta). I was interested to see how traditional SaaS could benefit from adopting AI. And from my tests, I see three main areas for which GenAI seems interesting to adopt for traditional SaaS products:
To reduce product friction.
To expand user bases by democratizing product access.
To generate leads thanks to AI powered marketing tools.
Use Case #1 - GenAI to lower product friction
Product friction in software can take many forms, from installation and setup frictions to unintuitive user interfaces or data import challenges. No software product is frictionless. In this perspective, GenAI can help to ease many of these friction points.
Here are five types of product friction that I saw GenAI features address:
“Non Sequential” Tasks Automation: Traditional software automation focuses on well-defined task sequences. E.g: “If a user registers on my website, then send her an email and add her info on Hubspot”. The advent of GenAI has introduced a new paradigm where AI can understand the context and generate outputs for the user (even based on unstructured data). This opens up the automation of an increasing number of tasks where the context is more fluid and less sequential such as reporting, data audits, or cost estimations. For example, several SaaS for the restaurant industry are testing AI powered features that automatically answer customers’ reviews or analyze their third-party platform orders to provide recommendations for reducing losses coming from disputed orders.
The “I Need a Co-worker” Problem: Often, users in a SaaS product can initiate a task but lack specific skills to complete it. For example anyone can create a presentation on Powerpoint but not everyone can make it professional-looking (most people need a designer). Same with spreadsheet software that anyone from marketing to sales or product can use, but conducting in-depth data analysis can require the help of a data scientist. GenAI powered features are now embedded in software like presentation and spreadsheet tools, acting as virtual co-workers with specialized skills, thus reducing the “I need a co-worker” problem.
The Blank Slate Problem: Many SaaS products require initial data or configuration to be useful, which can be a significant barrier to adoption. GenAI can assist users in populating interfaces, drafting articles, or creating brand imagery, effectively overcoming the blank slate problem. For instance, AI can be used to help product managers generate mind maps much faster or to fill up your CRM with data.
The Search Problem: Professionals like lawyers or doctors often face the daunting task of searching through vast amounts of text and records. GenAI can streamline this process, quickly surfacing pertinent information and even uncovering insights that might have been overlooked.
Action Autocomplete: Deciding on subsequent actions in a SaaS product can be a point of friction. AI can accelerate this process by suggesting or even completing these actions. Microsoft’s Co-pilot approach is a great example, where AI assists in decision-making and task execution. Before reliable autonomous AI assistants can automate tasks from start-to-end (which will probably take time), I think that the autocomplete model is the one that will spread first.
Use Case #2 - GenAI features to expand your user base
If the first use case is about lowering frictions for existing users, the second use case is about broadening the user base and enabling individuals who previously lacked the required skills to use these tools. This approach basically democratizes access to SaaS products, broadening their market reach.
An obvious example of this democratization can be seen in the creative software category (like the Adobe Suite). These creative software suites, historically the domain of professional designers, now incorporate AI-powered features that enable anyone without formal design training to produce high-quality images and videos. But I believe that this trend will probably impact an increasing number of B2B software categories.
For instance, GenAI is opening up access to tools typically reserved for data analysts, such as those requiring to know how to to write SQL query. Similarly, several analytics products are experimenting with features that help non-specialists in creating and interpreting data dashboards. This market broadening isn't limited to data-centric tools; it's also happening in areas dominated by developers (see Vercel assistant) and marketers, now becoming accessible to a larger audience, such as small business owners for example.
It's important to note that this "user base expansion" facilitated by GenAI is still at an earlier stage compared to the initiatives I listed in the previous section (lowering product friction). I also believe that not all software categories will be revolutionized by this trend. In many cases, new "AI native" SaaS products will be better positioned to capture and expand existing markets.
Use Case #3: GenAI features for lead generation
The final use case is much more speculative but could be an interesting one for traditional SaaS founders: Building ChatGPTs instances to solve specific pain points for users and provide these solutions for free as lead generation tools. This approach is not about adding AI features to existing products but about using the recently launched ChatGPTs marketplace as a marketing distribution channel.
I’m currently writing my thoughts about this topic as I already tested 10-15 GPTs (both consumer and B2B) and will share them in a following post.
But are we there yet?
From testing these 20 non-“AI native” products that are implementing GenAI powered features, it's clear that GenAI has the potential to significantly enhance existing products without the need for complete overhauls.
However we are still far from getting amazing results. IMO seamlessly integrating AI-powered features into existing products presents a real UX/UI challenge as well as a quality of result one. At the moment the output quality varies, and in many cases, it's not yet at a point where it significantly reduces the targeted friction (I was more often disappointed than amazed during my tests).
However I believe that there's a reason for optimism. In my opinion the current state mainly reflects the fact that we haven’t nailed the UI/UX part. For example I’m not sure that every internal AI assistant should be chat/prompt based. And the data part is also not nailed yet. Getting great results with GenAI often depends on how well you feed it with relevant data or on how well you have adapted a prompt to your own needs. I think that most of the AI powered features I tested don’t yet take into account the specificities of each user (lack of personalization).
A telling sign is that many AI assistants currently stand as separate components within SaaS products, not fully integrated into the core experience and are still “in beta”. I think we will be there once these AI functionalities become invisible to the user.
But these challenges are also opportunities. I believe that for many non-”AI native” SaaS founders it’s worth spending months on polishing the AI implementation in their product.