.avif)
Hype, no hype, overblown or overrated, ethical or evil. A whirlwind of thoughts have been swirling around since Open AI dropped the bot at the end of 2022.
This piece is not about that, (though our opinions are wide, rich and deep).
This piece is simply about the practical and common sense adoption of helpful tools. For over a year, we have explored, tested and applied a range of apps, platforms and use cases for AI and automation. Here’s what we learned.
Learning 1: You can either be busy trying to find the best AI tech or you can be busy trying trying to solve the right problem with AI tech.
💡 Focusing solely on discovering the "best" AI technology often leads to a counterproductive cycle of perpetual tool-hopping, while overlooking the fundamental purpose of tech: solving real business problems and eliminating friction while trying to get things done. Reframing the use of tech as problem-solving rather than tool selection leads to more impact and scalability.
Successful teams first identify specific pain points in their workflows—whether it's customer service bottlenecks, data analysis inefficiencies, or repetitive administrative tasks—and then select AI solutions that directly address these challenges in an ongoing and sustainable way.
A problem-first approach ensures more tangible business value.
Learning 2: You need to step back from the technology, take a breath and see it for what it really is. Multiple apps with unique capabilities, working together and doing what needs to be done in real time.
💡 Technology is like an orchestra - individual instruments may be impressive on their own, but true magic happens when they play together in harmony, conducted by a clear purpose. Connecting specialised bots (some especially capable of scraping data, some at analysis, some at synthesis, some at visualisation etc…) is analogous to putting together multidisciplinary teams to take on completing processes day in and day out with literally a click. All business activities are important but not all of them especially exciting to perform.
When the output is critical for decision making and action (comprehensive competitive analyses and learning, administrative tasks, synthesising employee surveys for insights…) but the process is repetitive and tedious, these are the interesting ‘problems’ to address with well-designed intelligent automation.
Learning 3: Don’t look at the tech as only costs reduction. Look at it as something that can increase the speed to insights and info to help make smart business decisions.
💡 A long time ago, we were shaping a strategy for CX for a leading European bank. With data from CRM, call centres, branches, app and web analytics we were able to crunch the info and pin point the make or break moments when customers were angry to leave or to simply transact with another bank. Mind you, the outcome of this is loss of revenue. We designed CX initiatives around these customer pain points. It took a long time to extract valuable learning due to the sheer amount of data, from multiple sources, in multiple formats. Today, AI and automation systems designed for pattern recognition can do the same at a tiny fraction of the time.
Today, we would use the power of the tech to gain insight and design action and not just replace the call centres with bots.
Consider business driving outcomes as a goal and not go for a knee jerk reaction to cut, cut, cut.
Learning 4: Don’t throw out the baby with the bathwater. New tech with tested and trusted strategic frameworks are a pretty good combinations.
💡 We have always been fans of the JOBS TO BE DONE framework and journey mapping. Old and tested, they are powerful tools we use in order to kick off building AI & automation systems for clients in simple steps:
- understand the critical JTBD in an organisation, a business unit or team
- prioritise them based on IMPACT and SCALABILITY potential
- select experiments for pilots or evolving systems
- map the current task journey and optimise it with AI & automation
- apply agile methodologies of experimentation, iteration and testing
Learning 5: Talk is cheap and so are experiments. In a good way.
💡 Too often we see teams jumping in and eagerly, enthusiastically wanting to get sh&*t done. It pays off to step back and think, discuss and plan. Talking and planning for a couple of days with the above thinking literally costs nothing.
Aim before you fire.
To recap:
- Don’t look for use cases in the tech. Look for use cases in the business.
- Apps, API platforms and interoperability make things possible, not standalone bots.
- Don’t delay too long planning and experimenting. There are opportunity costs.
- Use cases are nearly infinite. Find yours. Look at our examples.
PS: no bots were used nor hurt in the writing of this article
