Hi, it’s Kerry here standing in for Damian, who will hopefully return from his travels energized and eager to host The Last Mile next week 😎
Meanwhile I hosted a LinkedIn live this week with my co-conspirators Michael (Fish) Fisher, and Dan Aslet where we revisited the idea of the AI first contact center that we first discussed in our ebook on the subject.
When we first shared this vision 18 months ago, we were pushing the boundaries of what was possible, but we were still able to demonstrate how Gen AI could transform contact centers through a virtuous cycle of automation, assistance, insight, and optimization.
Ultimately, we saw a future where the contact center becomes a self-improving system with AI agents identifying opportunities to improve automation, and assist human agents, with Gen AI coding agents making the updates. A cycle that we felt sure would lead to the ‘no-agent’ contact center – where all front-line contact is handled by AI, with senior agents training and supervising the AI, and handling escalations.
Since then AI has got better, faster, cheaper and easier to build and optimize – if you do it right. But most of the organizations we talk with are still building like it’s 2022!
In the LinkedIn live event, Dan highlighted the struggles he’d experienced with low-code platforms. While easy to start with, these tools quickly hit limitations.
Switching to Agents SDK and vibe-coding allowed rapid prototyping—provided organizations clearly understand their Standard Operating Procedures (SOPs).
Fish brought up another critical bottleneck: outdated business models.
Traditional seat-based licensing is increasingly incompatible with AI-first operations, where consumption, or outcome-based models are essential. Procurement teams grapple with understanding pricing based on tokens or minutes, leading to cautious hesitation instead of decisive action.
As we predicted, contact center vendors also remain behind, with CCaaS providers bolting on AI capability 12 to 18 months after its available on hyper scale cloud platforms like Microsoft Azure, Amazon AWS, and Google Cloud Platform.
This reinforces what we’ve been cautioning since the original ebook: waiting on vendor roadmaps is a competitive risk.
18 months ago we laid out three options for organizations considering their next step with contact center AI: Disrupt, Keep-Up, or Wait.
In the discussion we all agreed that waiting is no longer an option: with highly regulated fintechs like Klarna going all-in on Gen AI and Shopify bundling conversational agents with their platform, it’s time to move ahead, or get left behind.
Check out the 30-minute replay of for the full discussion:
The AI First Contact Center: From Hype to How
Kerry
PS: If you’re new here, this newsletter brings you the best from Waterfield Tech experts at the frontier of AI, CX, and IT. Kerry posts weekly at The Dualist, and Fish and Dan share their thoughts every other week at Outside Shot and Daichotome.
Here’s what went down this week.
Bleeding Edge
Early signals you should keep on your radar
AI wins gold medals in Math Olympiad. ExperimentalAI models from Open AI and Google both solved 5 of 6 International Math Olympiad problems – achieving gold medals in a world-first for AI, demonstrating that AI is turning abstract reasoning into a commodity. (Reuters)
Study Mode flips ChatGPT into a coach. OpenAI’s new toggle turns ChatGPT into a teacher – exactly the “active-learning” pattern that drove the 127 % learning boost we covered two weeks ago. (OpenAI)
Leading Edge
Proven moves you can copy today.
ChatGPT gets its own agent. Pro, Plus and Team users can now hand ChatGPT a goal and it’ll spin up a virtual computer with a full browser that lets it not just search out information, but take action across the web. I’ve yet to hand over my credit card details but it’s research is top-notch! (OpenAI)
Big Tech’s AI bets pay off: Microsoft, Alphabet and Meta all reported that AI is already boosting search, advertising and cloud revenue. Microsoft claimed 100 million Copilot users. Alphabet claims Gemini has 450 million monthly users, while ChatGPT attracts roughly 500 million weekly users. Notice how OpenAI report weekly active, Google monthly active, and Microsoft clearly don’t want to admit how often paying customers actually use copilot! (Reuters)
Off the Ledge
Hype and headaches we’re steering clear of.
Google grudgingly signs EU AI code: Alphabet will sign the EU’s voluntary AI code of practice, which explains how to comply with Europe’s landmark AI Act. But Google’s legal chief warned that the code could slow approvals, require disclosure of trade secrets and diverge from copyright law—risks he says could chill European model development (Reuters)
Mark Cuban says the AI arms-race is “getting ugly”: Meta is reportedly flashing $100 million offers—and a $15 billion Scale AI tie-up—to poach OpenAI and Google researchers. Maybe I should throw my hat in the ring 🤣 (Business Insider)