At Assist Scout, our mission is to empower companies to thrill extra clients. We place super worth on the “delight” a part of that equation. We’ve by no means included ticket numbers in emails to your clients; as an alternative, we name them conversations. We by no means say somebody’s out there to speak until they’re. We didn’t rush to construct chatbots that strike out too typically and frustrate clients. We make merchandise optimized on your buyer first, which is why a lot of you select Assist Scout.
On account of current progress within the subject of AI — notably the adoption of enormous language fashions (LLMs) — we’re getting into the golden age of delighting clients. This expertise allows us to construct issues we wouldn’t have dreamed of after we based the corporate in 2011, however we’re having quite a lot of enjoyable constructing them now. It’s modified our perspective on AI and what it might probably do for rising companies.
What made this AI cycle totally different
We’ve adopted all the “AI for buyer help” hype cycles very carefully. We don’t consider ourselves as specialists — we’re builders, all the time in quest of new instruments to assist us (and also you) delight extra clients. We tried all the instruments. We constructed issues with them, we used them internally, however all of them fell quick. We spent quite a lot of time in two totally different areas:
We used quite a lot of chatbot instruments and received to know the expertise powering them. Generally, the chatbots have been arguably not utilizing AI, and their capabilities have been restricted. In the end, we discovered that search was a greater UX than a chatbot.
We tried a number of various things with machine studying — predictive AI for tagging, draft responses, workflows, and Docs article suggestions. The issue was that for the mannequin to be semi-accurate, it required extra knowledge than 95% of our clients had.
The brand new instruments (LLMs) in a short time blew us away and solved a number of of the challenges we had skilled beforehand. LLMs have sensible purposes that may profit each enterprise, notably buyer help groups. In actual fact, “improved buyer expertise” is among the high optimistic outcomes enterprise are experiencing on account of their adoption of generative AI instruments:
In February 2023, we put collectively a small group of parents to spend 30 days studying, prototyping concepts, and constructing issues with LLMs (OpenAI, particularly). After these 30 days, unbelievably, the crew had constructed seven issues we have been enthusiastic about. Proper then and there, we began re-working the product roadmap.
For sure, our ideas about whether or not AI is “the fitting instrument for the job” have developed considerably. When thoughtfully utilized, this expertise will allow you to thrill extra clients with out having to undertake clunky enterprise-level instruments your crew doesn’t need to use. You may love your help platform once more.
What about buyer help jobs?
LLMs are highly effective as a result of they’ll learn and write English extremely effectively. In some circumstances, it makes them look a lot smarter than they really are. None of us can predict the longer term, however I’ve an optimistic view on this subject (like many others). I consider these instruments are going to assist buyer help professionals do even higher work, which has a basic influence on the client and the enterprise.
For us, and for the companies we speak to, that is about greater than expertise disrupting the labor market (which has occurred for hundreds of years). It’s about giving everybody magical instruments they didn’t have entry to earlier than. Some customer-facing roles might evolve or disappear, however with that change will come a chance to adapt, study new expertise, and do extra fulfilling work.
No matter how good AI expertise will get, it is inconceivable to rely solely on AI with out inadvertently (or worse but, purposefully) creating countless loops or useless ends for purchasers in search of help. Whereas decrease ticket quantity might look good on paper, we construct AI for happier clients, not deflected tickets. Human assist ought to all the time be straightforward to search out when AI isn’t the fitting answer. So for us, we simply don’t see a future with out individuals being a part of the help journey.
The web modified the world as we all know it — in good methods and unhealthy, however we consider far more so for good. This expertise may very well be simply as impactful. These instruments can increase the inventive course of, allow us to study sooner, and pace up complicated duties. They’ll do quite a lot of good, and we’re eager to discover methods of making use of that in our product. We actually are getting into the golden age of delighting clients.
The street forward
We haven’t been this excited a couple of new expertise since founding the corporate, however there’s nonetheless quite a lot of work forward. We’re wanting to focus extra of our consideration on crafting pleasant person and buyer experiences, backed by AI.
