5 No-Brainer Ideas to help you get started with AI

Dawid Naude, Director, Pathfindr

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?

The same can be said for a working group without any $, resources, focus or help. Or you might be ‘encouraging AI’ but telling people they need to use it without putting anything they’re working on in it (come on, you know better).

But what if you had to get a plan together tomorrow, what if all of a sudden you had a bucket of money and needed to demonstrate AI that is felt across your whole organisation within the next 3 months? What are the 5 things you should do?

A note here - this is for broad AI in your organisation. It won’t have a massive direct impact on any function, but it will lay the foundation on which everything else is built and start having your entire organisation using AI every day.

In short - these 5 ideas are necessary but not sufficient - to make a massive measurable impact (50%+ productivity savings). That will come in the immediate projects after the first 5.

Five ideas to get going

Give your team a secure ChatGPT, and show them how to use it

Build an enterprise knowledge assistant (and show them how to use it)

Make your data conversational (….. and show them how to use it)

Bump up your website search

Pick 3 artefacts to automate end-to-end

1. Give your team a secure ChatGPT and show them how to use it

Chat Copilot tools like Gemini, ChatGPT, Copilot for MS365 can make a huge difference to productivity, focus, and job satisfaction. Think of a grad who’s getting a ton of new terms thrown at them they don’t understand, or a new manager not sure how to structure a status report, or someone who has English as a second language (common particularly in IT) and needs to read some complex documentation or write an important email.

These tools can make a huge impact, but your team needs to be able to put their data into it. They need to be able to put a requirements specification in it and ask if they’ve missed anything obvious (and it comes back with - “you should also include a table of security and permissions”), or a full-length document from a client and ask it to summarise the key points.

You already know that your employees (and likely even you) have put things into ChatGPT free personal version that you/they probably shouldn’t have. It’s happening every day across your whole workforce. Walk the halls and you’ll see ChatGPT open on several monitors.

You need to give them a secure way to do this.

There are 3 obvious ways you can do this. 1 - You can buy a premium version of ChatGPT (team/enterprise), MS Copilot, Gemini, etc. You’ll be opted out of model training, and you’ll be trusting the provider with your data based on how they say they’ll use it and protect it. (just like you do every day with Salesforce, ServiceNow, Hubspot, Adobe, Workday, etc).

2 - You can host it in your cloud of choice with their recommended tool, where your data and compute already sits. Eg. Azure OpenAI, Amazon Q, etc. A bit more complex but being made easier every day. Probably the best middle ground for most organisations.

3 - You host it where you want, how you want - From your own server, to your own device. No data ever passes outside your company walls. Great for highly regulated industries and defence departments. Complexity varies, but you’re taking on a lot of the work that providers take off your hands.

Now you have a spot where people can go all-out and use these tools.

But you must show them how to use it - or else only about 10% will use it regularly and get benefit.

2. Build an enterprise knowledge assistant (and show them how to use it)

Your company has a lot of information, both structured and unstructured. Centrally organised or individually contributed. Some items are used frequently, others once or twice a year (and often results in manual effort for someone else because it’s hard to find, like contacting HR about a leave policy).

But what if you had a simple chatbot that you could ask “Can I take sick leave during annual leave” or “Can you show me all our research into the utilities industry”. Or how about “what training courses are available for someone managing a team for the first time” (and the result is “Leading for performance course” - where there was no mention of leading or performance in the search term, but found the best result.

Give your company an easy way to find this knowledge, it’ll cite the underlying source and they can go into it for more information.

This needs to be deliberate or else it’ll be garbage. If you just point it to Sharepoint folders all over the place it’s going to be messy and confused. Make it curated and specific, and give them a feedback mechanism.

Think of the employee experience for a new joiner - “just ask this tool”.

Again - you need to show your team how to use these tools and evangelise the heck out of them, set expectations and also make your team feel like they’re part of the AI journey, and that it’s not happening to them.

3. Make your data conversational (….. and show them how to use it)

You have a lot of valuable data (if you don’t think you do, you just haven’t thought about it properly). That data might be in a good state or a bad state, either way, you can use this data with AI to get insights. A large language model is the equivalent of having a highly performing graduate-level data analyst working for you 24/7. Commands like “Can you pull together a chart that shows how many customers we acquired in the past 6 weeks in these 3 areas” is a data analyst task that can now be given to AI if you put the conversational layer over the top of it.

If your graduate data analyst can point out that there seems to be a lot of duplicates, then AI can too, and if you're able to explain to that data analyst why those duplicates exist and how to handle them (‘yeah, just treat those as the same customer, they repeat 3x for each record), then you can tell AI that too.

That’s the starting point, but it gets much more advanced. The next step is to ask “What are our financials for the past 6 months telling us?”, “What products do you think we should stop selling?”. Your definition of each of these can also be built into it.

Business Intelligence teams would feel a lot of frustration here. Typically they’ve gone to tons of effort (and $) to build self-service dashboards with filters that business users still struggle to use, even though it’s simple. But now their effort will be rewarded through access through a chat interface in MS Teams, slack, or wherever their work is simply ‘asked’, as they would typically get asked, and they can spend their time building the next data product.

4. Bump up your website search

Website searches vary greatly. Most haven’t had an upgrade in over 5 years, yet the technology has completely changed. It’s gone from keyword searches to ‘semantic searches’ (meaning it understands the meaning of the search). It’ll interpret “I’m looking for leadership training courses” the same as “promoted to manager, looking for guides”.

The next step is having it search across and through documents and producing it as a briefing. “promoted to manager, looking for guides” could be answered as “There are 3 courses available for this, 2 are on demand and there’s 1 in person, as well as 5 playbooks listed below. There’s also a community group that meets once a month for this”. How powerful would this be internally and externally? You already have great content, but it can’t be found.

5. Pick 3 artefacts to automate end to end

What if you never had to create a status report, memo, pitch, or requirements specification from scratch ever again?

What if all your team leads were messaged automatically on MS Teams or on Email and asked to reply with their updates for the week? And what if those updates were automatically collated, summarised, and populated a PPT/Word/Confluence document automatically.

If you’re in this world, you know that you’ve just saved at least several meetings, hundreds of messages and emails, and probably about 3-4 hours of work for the project manager. In total, this process alone could collectively save over 10 hours each week. But what if it also looked at the amount of user stories completed and populated that summary too? What if it also updated a table of who is taking leave when?

Now picture this same process for creating a client proposal where you’re first just asked a few questions by a chatbot “Who’s the customer, how long is the presentation, what insight are you looking to show, how many team members, what’s the product” - and it automatically created an entire PPT or PDF.

Wrapup - Necessary but not sufficient, what’s next?

These are great no-brainers and can get you a massive win over the next 3 months, and have your whole organisation feeling AI.

But. It won’t make a material difference across the whole organisation, you won’t see gains of 15%+ across the whole workforce, yet. For some people with the artefact creation in particular, it will help them massively.

What have is a foundation, momentum and a culture that is loudly embracing AI instead of having it done to them.

What you need to do next is continually add skills to your knowledge assistant, and chatbots - make it be possible to produce a briefing report just through a question, log leave, raise an IT support ticket without ever leaving your collaboration tool.

You need to also then pick processes you want to infuse AI with, not artefacts or people. Look at your entire order fulfilment, onboarding, IT support or marketing process and ambitiously improve it with AI.

 

 

Other Blogs from Dawid


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.

AI Value not in Chat

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.

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.