“AI won’t replace humans but humans who can effectively integrate AI into their workflow will replace those who do not”
As VP of R&D Customer Engineering, my team’s remit is simple: mine market signals, engage with customers and partners and turn that into insights that fuel our innovation and go-to-market. Success in this function shows up first in sales productivity: better customer conversations, higher win‑rates, cleaner forecasts and second in being an “outside-in” voice for our executives.
Traditionally, compiling and processing reams of information has been foundational to this function. To this end, we have been inspired to look to A.I as an enabler of growth, scale and depth of insights. After almost two years of us shepherding AI from prototype to production, here are 10 lessons that we have learned. I would love if you add some of your own in the comments below!
1. Relevance Has a 24‑Hour Half‑Life — Use it or lose it.
Market launches, new compliance regulations, surprise M&As come at us at a staggering pace — but the signal can easily get lost in the noise. We have fine tuned our Agentic A.I platform for quality over quantity allowing us to find, filter and summarize market intelligence and internal communications into our training set nightly.
2. Get to the Best Answer Fast — Scale Expertise
While customers and the market in general is more informed and prepared than ever before they still expect direct answers to technical questions. The skills and experience of our internal human experts are critical to scaling but capacity is always limited. We have found that direct transcription of “expert conversations” has been a great way to feed AI with high protein context allowing it to to better respond to technical inquiries.
3. Capture the Customer Whisper— Improve your Messaging
When we listen carefully we find that customers are often very clear about their requirements and what they expect from vendors that they work with. AI possesses an incredible power to “listen without prejudice” helping us to better cater our messaging and proposition. The learnings from these conversations are also critical for consolidating key themes for our marketing organization and bridging the power of our platform to the market.
4. Deeply Understand “Why you win and lose”
Traditional spreadsheets and data visualization tools are great at showing trends but have lacked in outlining the true anatomy of a deal and the critical factors that drove it’s outcome. We have spent significant effort fine‑tuning AI agents to look deeper allowing it to build it’s own opinion of our wins/losses and a wider understanding the full deal lifecycle.
5. Architecture Matters – “Beyond the Widget”
While the quick win may look like a chat widget bundled inside of one of your applications longer term the magic of A.I is tied to the architecture of your data library, agents/experts and inference. Where is your data being fed from? What experts exist within your model and how are they segregated? What LLM models are you leveraging in the interpretation of questions and delivery of answers? What visibility do you have to the questions being asked and the interactions with your consumers? These are all evolving areas but customization is a critical differentiator here.
6. Behavior Change and Trust Take Time
A brilliant model alone does not drive change. Every major release needs a clear beta launch with committed early adopters, formalized success criteria and objective evaluation. In addition, tools need to be integrated into the workflow rather than sitting as a separate app.
7. Start with Pull but evolve to Push
Every new innovation starts as a new app or a new tool that is made available to users but over time the AI system needs to integrated into collaboration, CRM and productivity tools in order to bring the right insights to the right people at the right time. Said another way the aspirational outcome is delivering answers before the question is asked!
8. Human + Machine — Not Human vs. Machine
AI provides an incredible engine to complement human expertise and has tremendous potential to accelerate workflow, improve productivity and make all of our jobs more interesting. But remember, AI does not replace human domain expertise, judgment and decision making and when used in the right way, it enhances it.
The Road Ahead
Generative AI today feels like early cloud—promising, chaotic, inevitable. Tooling will mature, budgets will stabilise, and regulations will crystallise. The leaders five years out will be the teams logging the unglamorous reps today: scrubbing data, wiring guardrails, and embedding sales‑centric feedback loops before sprint planning.
Got a key A.I learning to share in your own world? Share it below!
Leave a Reply