SCG’s 90-Day Transformation Journey

At Gaffex [music] SCG, we embrace AI. We

experiment with different AI tools, even paying out of pocket, hoping to spend less [music] time on admin tium. The

average SCG officer spends 57% [music] of his time on repetitive tasks. That's

about 2.85 days lost per week. More than

80% of SG officers have been using AI.

We do it in small ways like trying what pair can do and building trial apps with AI bots. when we thought, what if we can

AI bots. when we thought, what if we can turn this trial and error into something bigger? Meet AI first. This isn't just

bigger? Meet AI first. This isn't just another hackathon. It is our chance to

another hackathon. It is our chance to [music] build AI powered solutions even though we aren't AI experts.

AI first works on two fronts. We hold

monthly screens where teams tackle real problems. Teams are supported with AI clinics, workshops, and learning journeys. We also offer access [music]

journeys. We also offer access [music] to a growing collection of guides, tutorials, and tools to enhance our AI skills. We've run three sprints [music]

skills. We've run three sprints [music] so far with more than 200 gap techies working on 73 AI powered solutions to make our work smarter.

>> We are cross divisional team for finance [music] and digital governance. Our

challenge month financial reporting that's a 12-hour marathon against the clock. Our team manually consolidates

clock. Our team manually consolidates data from over a dozen sources. is a

high stakes process with constant error risk. So we built an end-to-end AI

risk. So we built an end-to-end AI powered treasury [music] management system that automatically integrates and analyzes data with offices controlling every critical decision. The system

requires human confirmation each step data extraction validation and report approval. We tested it with internal

approval. We tested it with internal audit and finance colleagues for valuable user inputs. We transformed

that 12-hour audio into just one hour, 92% faster, saving 17 mandates annually.

Officers remain in complete control while AI handles the heavy lifting. We

are GFK's policy performance and rewards team trying to solve every officer's struggle with writing code performance evaluations. There's no existing [music]

evaluations. There's no existing [music] evaluation templates based on GFK schemas and panel expectations. We built

[music] app praise AI an AI evaluation buddy trained on all gartk role schemas tough evaluators questions and best practices from expert writers. Our

solution guides you through seven steps.

It helps us create impactful evaluations [music] that showcase the actual contributions of the officers. 90% of testers confirm

more accurate impactful evaluations.

users cut evaluation time from a full day to just a few hours while significantly improving quality of their writeups. We're from the uh people and

writeups. We're from the uh people and organization division. Uh our team was

organization division. Uh our team was focused on tapping the resistance from individuals to even attempt adopting AI into [music] their day-to-day activities. Taking into consideration

activities. Taking into consideration that individuals might feel unfamiliar or unsecure about using external tools like chat gi [music] claude gemini we looked at what was available inhouse and

eventually identify AI boards which is a gaffech tool as a solution to this by using AI we can now generate smart summaries detect action items with owners and deadlines and instantly share

follow-ups all from the meeting transcribe instead of having to previously manually [music] review it with AI boards even those who once found AI daunting now see how easy it can [music] be. The result, greater speed,

[music] be. The result, greater speed, clearer outcomes, and higher productivity. Our solution has made AI

productivity. Our solution has made AI adoption simpler.

>> We're from finance handling huge [music] data sets for cost and revenue analysis.

We wanted to see how AI could help. So,

when we tested government approved tools like pair and AI bot, it actually couldn't really handle the complex math that we needed. We found Julius AI which has the financial capabilities that we

were looking for but using external tools would mean that we had to anonymize our sensitive data manually.

We actually converted our information required into food terms. So while this concept proved that it worked but translating insights back [music] into the real business context was a little

tricky. What is our key takeaway? So

tricky. What is our key takeaway? So

using AI for financial analysis actually needed three things. A very clear scope, a right AI tool and a very clean data set. We are from the digital governance

set. We are from the digital governance and we tackle formal meeting notes on a monthly basis while juggling with providing quality assessment. Preparing

minutes used to drain our time. So we

experimented with AI tools. MS teams

co-pilot turn our minutes quickly but couldn't match government formatting standards. Pair had potential but still

standards. Pair had potential but still trapped us in tedious copy. [music]

Then transcribe transform our workflow.

It delivers formal meeting minutes in the format we needed with high accuracy.

This lashes our 3-day marathon down to just one day for the first draft. The

result, we have unlocked a time for what truly matters. Diving [music] deep into

truly matters. Diving [music] deep into project team assessments and delivering real value.

>> We are from the digital governance [music] capability enablement team. We

need to decide whether to renew our million-doll contract. But we face

million-doll contract. But we face mountains of disconnected data [music] and considerations. Yet somehow we need

and considerations. Yet somehow we need to take all this discontined chaos plus a dozen of conflicting departmental opinion and forge it into a single confident yes [music] or no decision. So

we leverage AI to build deal breakaker.

First it pull together the data to offer a complete big picture. Second it gave you instant insight through an AI bot.

Third a decision optimizer help you to consider different scenario instead of waiting for someone [music] else analysis. Decision maker now can

analysis. Decision maker now can explore, question and model scenario themsel.

>> And that's our AI first story. Whether

you're from another part of Gtech or another public agency, [music] join us in turning small AI experiments into scalable success. Get in touch with us.

scalable success. Get in touch with us.