Greg Coleshill, Chief Commercial and Operating Officer
If you want AI to move the numbers that matter, stop treating it like a science project.
The quickest path from promise to profit is to place AI in the hands of the people where value is created. Not a solution trying to “boil the ocean” or solve everything in one fell swoop. Teammates with clear roles, sat beside your people, improving outcomes in front of them whilst delivering demonstrable quick wins and engagement. Building the road ahead to greater adoption and scale.
Many pilots look great on a slide but fail to land in the real world.
You know the pattern.
A team trials a tool. A few people get curious. Early feedback is hopeful. Then momentum fades.
It is not because the tech is weak or the work has changed. The tool sits next to the task rather than inside it. No workflow change means no lasting behaviour change. No behaviour change means no measurable gain.
So what to do? With AI adoption in mind - pick a single job to be done, put AI inside it, and judge success on a metric leaders already care about. Time to first sale? First contact resolution? Claim cycle time? Pick one, but make sure it matters. If you cannot point to a number that moves, you do not have a use case. You have a demo.
I’ve already said recently, language matters.
When people hear “platform” they think optional. When they hear “teammate” they expect help.
A teammate has a name, a job description and a measure of success. A platform has a login.
Good teammates do three things well:
People adopt what helps them finish with confidence.
Speed and safety can live together if you design them together. The right rails live where the work lives.
Treat oversight as a product feature, not a committee meeting. Red team before launch. Monitor after launch like any other operational risk. The point is not to slow delivery. It is to catch errors early and learn from them.
You do not need ten use cases to change the culture. You need one per function that pays for itself. Here is a pattern that works:
Once you have one win per function, pattern match. Reuse the parts that made it work. You will scale faster by cloning a proven design than by inventing new toys.
Sales. New reps spend their first month finding the basics and guessing the tone. An onboarding teammate pulls the right materials, teaches the pitch in steps and drafts the first five outreach messages in the brand voice. The human edits and sends. The metric is time to first qualified meeting. When that number drops, managers feel the lift and adoption sticks.
Customer service. A quality checker reads replies before they leave the agent’s desk, flags accuracy issues, suggests a clearer sentence and attaches the right policy link. The agent stays in control. The metric is first contact resolution. When reopens fall, the question is not “is this worth it” but “when does it roll out to the next queue”.
Everywhere else. If there is a repeatable decision with defined inputs and a clear standard of quality, you have a candidate for a teammate. Finance approvals. Claims triage. Promotions compliance. The difference between a showcase and a result is whether the teammate lives in the place where the decision gets made.
The fast movers do not hide AI in a technical lane. They put it in the hands of the teams who own outcomes.
Product, operations, risk and compliance shape the role, the rails and the metric together. The language stays human. The feedback loop is short. Release weekly, not annually. Most value comes from fit and iteration, not from model size.
Get specific.
Specific about the job the AI will do.
Specific about the rail that keeps it safe.
Specific about the number it will move.
Put AI where people actually work and call it a teammate. Do that and pilots turn into profit, not just presentations.
We deliver AI teammates for regulated businesses. We enable productivity, safely, with real-time guardrails.
We believe the future of work is AI teammates collaborating with humans to lift outcomes. Others share that belief. Where we differ is how it comes to life:
We encourage leaders to see AI differently. Stop treating it like software. Treat it like a teammate. Like any new hire, it needs onboarding and coaching, and people need time and evidence to trust it before it reaches peak productivity.
November 5, 2025
Greg Coleshill, Chief Commercial and Operating Officer
November 5, 2025
If you want AI to move the numbers that matter, stop treating it like a science project.
The quickest path from promise to profit is to place AI in the hands of the people where value is created. Not a solution trying to “boil the ocean” or solve everything in one fell swoop. Teammates with clear roles, sat beside your people, improving outcomes in front of them whilst delivering demonstrable quick wins and engagement. Building the road ahead to greater adoption and scale.
Many pilots look great on a slide but fail to land in the real world.
You know the pattern.
A team trials a tool. A few people get curious. Early feedback is hopeful. Then momentum fades.
It is not because the tech is weak or the work has changed. The tool sits next to the task rather than inside it. No workflow change means no lasting behaviour change. No behaviour change means no measurable gain.
So what to do? With AI adoption in mind - pick a single job to be done, put AI inside it, and judge success on a metric leaders already care about. Time to first sale? First contact resolution? Claim cycle time? Pick one, but make sure it matters. If you cannot point to a number that moves, you do not have a use case. You have a demo.
I’ve already said recently, language matters.
When people hear “platform” they think optional. When they hear “teammate” they expect help.
A teammate has a name, a job description and a measure of success. A platform has a login.
Good teammates do three things well:
People adopt what helps them finish with confidence.
Speed and safety can live together if you design them together. The right rails live where the work lives.
Treat oversight as a product feature, not a committee meeting. Red team before launch. Monitor after launch like any other operational risk. The point is not to slow delivery. It is to catch errors early and learn from them.
You do not need ten use cases to change the culture. You need one per function that pays for itself. Here is a pattern that works:
Once you have one win per function, pattern match. Reuse the parts that made it work. You will scale faster by cloning a proven design than by inventing new toys.
Sales. New reps spend their first month finding the basics and guessing the tone. An onboarding teammate pulls the right materials, teaches the pitch in steps and drafts the first five outreach messages in the brand voice. The human edits and sends. The metric is time to first qualified meeting. When that number drops, managers feel the lift and adoption sticks.
Customer service. A quality checker reads replies before they leave the agent’s desk, flags accuracy issues, suggests a clearer sentence and attaches the right policy link. The agent stays in control. The metric is first contact resolution. When reopens fall, the question is not “is this worth it” but “when does it roll out to the next queue”.
Everywhere else. If there is a repeatable decision with defined inputs and a clear standard of quality, you have a candidate for a teammate. Finance approvals. Claims triage. Promotions compliance. The difference between a showcase and a result is whether the teammate lives in the place where the decision gets made.
The fast movers do not hide AI in a technical lane. They put it in the hands of the teams who own outcomes.
Product, operations, risk and compliance shape the role, the rails and the metric together. The language stays human. The feedback loop is short. Release weekly, not annually. Most value comes from fit and iteration, not from model size.
Get specific.
Specific about the job the AI will do.
Specific about the rail that keeps it safe.
Specific about the number it will move.
Put AI where people actually work and call it a teammate. Do that and pilots turn into profit, not just presentations.
We deliver AI teammates for regulated businesses. We enable productivity, safely, with real-time guardrails.
We believe the future of work is AI teammates collaborating with humans to lift outcomes. Others share that belief. Where we differ is how it comes to life:
We encourage leaders to see AI differently. Stop treating it like software. Treat it like a teammate. Like any new hire, it needs onboarding and coaching, and people need time and evidence to trust it before it reaches peak productivity.
