99% of AI value won't be from chat interfaces
Dawid Naude, Director, Pathfindr
Chat is only incremental
The biggest mistake you’re making right now is thinking about AI in terms of ChatGPT, or chat interfaces. These assistant tools are great generalist support applications and a demonstration of the incredible power of LLM’s.
But. They are a single step in a process, typically involving a lot of Ctrl+c, Ctrl+v, a lot of prompting, trial and error, to get a consistent outcome.
This won’t ever have a consistently repeatable measurable impact. Two marketers will not only get different results because of how they both use the platforms, but even if they both used them the same way, they will get different results.
To be clear, I’m not discounting that ChatGPT, Copilot, Gemini, etc won’t play a significant role, they absolutely will, but don’t fool yourself into thinking that by rolling out ChatGPT, even with us facilitating a masterclass, will have any ground shattering impact across the organisation. Absolutely people will be more productive and they’ll report it how they use it, and how much time they’ve saved, but I don’t expect the CFO to be forecasting a 20% reduction in payroll cost off of 100 $25 licences.
End-To-End is transformational
Your biggest impact with Generative AI will come in the end-to-end processes you automate without needing the user to string together a dozen prompts and a ton of copy-paste.
Take the below example, here’s how you’re able to automate 50+ hours worth of research and artefact creation with a single command. It involved mapping the entire process of how a research analyst at a private equity client went about assessing a possible investment. Tons of googling, financial analysis of annual reports, searching competitors, Linkedin, media. Then creating the investment memo itself, in this case a PPT/Google Slide.
In this particular example it’s taken over 200 sources in under half an hour about about $7.23 in GPT4 tokens.
There are so many examples of being able to string together processes in your day to day.
Think of a project status report process where you need to get inputs from several people and sign off from someone else. Right now, should you be willing to dive into it, you could automate this whole process.
Even human input and review
You could initiate the process with a command or have it scheduled, this first thing the bot/agent would do is create a new version of the status report in confluence/ppt/wherever and populate it with the templated information, then it would automatically message each stream lead on MS Teams for their update for this week, showing them their updates from last week. They would reply directly in the chat message (and be reminded until they do). All the responses would be consolidated by the agent, the first draft of the status report created, burndown information and charts sourced from Jira, sent to the project manager for review (again, a MS Teams ping ‘your status report is ready for review’). It would then also make suggestions of things that may have been missed. Once changes have been made by the PM, then a simple “submit to client” will create the first draft of the email and attachment for review, a quick updates and hit send.
All of that is possible right now
Take the example below from one of the platforms we work with called RelevanceAI (great aussie startup), where it does cold sales outreach by researching the customer, sending them an email, creating a task in a CRM, then even facilitating the back and forth between a prospect and booking a meeting. The first time the manager gets involved is when the meeting is booked.
Earlier this week Microsoft announced their roadmap and features for their Agent platform, an extension of MS Copilot. Things like long term memory, asynchronous tasks, event monitoring, waiting for responses, etc. Have a look at the promotional video on their website.
If you’re going to take AI seriously, and you absolutely should, make sure you immediately look beyond ChatGPT (but still use ChatGPT), and start figuring out whole processes you can automate end to end. If you can describe the process to someone, you can absolutely describe it to AI and kick it off. The worst case is it creates a sloppy first draft to build on.
5 No Brainers
For all the hype around AI, it’s often not clear to business where they should get started. Some convince themselves they have a plan, you’re in this category if you say “We’ve given some people access to MS Copilot”. Let me guess - they seem to like the MS Teams meeting summarisation right?
AI Poor Data
I don’t know if it’s simply a convenient answer whilst there are hundreds of other competing priorities, or if it’s a misinformed opinion, but “our data isn’t ready yet to take advantage of AI” is something we hear regularly. Or similar variants like needing to spend 2024 getting data ready and then they’ll reassess AI in 2025.
How to think about copilots
Bots, like people, have two types of usefulness - knowledge and skills. Things it knows, and work it can do. In this Blog, Dawid Naude of Pathfindr provides practical advice to help you roll out or review your deployment of co-pilots and chatbots within your organisation, to generate both a short time impact and longer term strategic value.
Use whats in front of you
The first step your company should take is learning how to use ChatGPT properly. Very few do. There’s a lot of talk of AI, and businesses need to start tinkering right away. There is a huge opportunity accessible right now, right in front of you. You can completely automate processes with autonomous agents, have all your content created automatically for your learning management platform, and even enable the equivalent of a data science team with a few clicks.
10 Boring Problems
Whilst we talk about ethics, avatars, deep fakes and other magical technology, the reality is that for most businesses, the value in AI will come from solving very boring problems. The average desk worker has a lot of boring tasks. There’s a substantial amount of download, copy, paste, save-as, etc. There’s a lot of responding to messages that could be found on the website. There’s a lot of vlookups and pivottables, meeting minutes and deal updates. We love solving boring problems, and get less excited about deep fakes and avatars.
How accurate is AI
Is it accurate? Along with “Is my data safe”, this is the most common question about AI. We’ve all experienced mild embraces with inaccuracy, possibly even hallucination of AI models. So how accurate is AI, let’s get a bit more specific.
Early AI pioneers
Our clients are trailblazers of AI, and it’s clear what the leaders are doing.
AI Assessments
Checking, marking, critiquing, assessing, scoring We recently reflected on a common theme we’re noticing in the work we’re delivering. The projects had different names like “accreditation certifier”, “clause validator” or “skilled migrant assessment”, but they were all doing the same thing, assessing something. What we also realised is that AI is often great at a first draft, or a last pass. In this case, an initial assessment.