Category: Uncategorized

Automation needs structure to function.

This is because all automation is rule based. 

It has to function on an “if this, then that” basis, and it doesn’t work if the “this” isn’t where it should be.

Most AI at the moment is being used to try and structure information after it’s created, rather than during. 

“We didn’t classify that, so we’re using AI to do it” or “We didn’t structure that document, so we’re using natural language processing to populate some meta-data about it”.

Every time I see a new pitch for an AI tool, I wonder how much money an organisation will have to spend on the promise of a magical result, that could have been done far more simply (and cheaply) with a better structure.

I think structure to support automation is one of the great opportunities for RIM professionals. So many organisations with a great records managers are going to benefit from their approach to structure, and their deep understanding of the power of meta-data.

There’s a really simple relationship in customer service.

Investment in service is determined by switching cost.

The harder it is to switch to another service provider, the worse the customer service generally is. 

So service just has to be good enough to prevent “customers” switching governments, or finding a new job.

That’s not very good.

Recurrent conversation:

IT – “do we need all of this information”

Records – “they’re records”

IT – “yes but surely some of them are high value and some are low value”

Records – “they’re records”

IT – “do we really need to keep all of them”

Records – “they’re records”

I’ve also had hundreds of conversations with practitioners who tell me that it’s not important to define records.

So we keep having the recurrent conversation, because IT don’t understand records, because we can’t tell them.

Yesterday I wrote about the de-professionalisation of records.

Ultimately, it comes down to trading off accuracy for perceived efficiency.

Efficiency is only perceived as efficiency because what we really do is shift the cost of records from a professional team, to other teams in a way that we can’t measure.

That’s bad, but accuracy is far more troubling.

Accuracy is what you will build the entire future of your organisation on. Any process that relies on the records relies on their accuracy.

Unfortunately, what we end up trying to measure is the cost of work that won’t get done because it simply isn’t feasible, because our records aren’t accurate enough to produce reliable results, and the costs of getting them there are so large that it’s not feasible.

There are going to be winners and losers in this.

Winners will have high quality records, and will be able to take advantage of machine learning and many, many more automation technologies as they become available. They’ll also pass audits – which is nice.

Losers will have to do one of two things –

1. Start a records program to produce high quality records.
2. Wait for strong AI that can do the work anyway (in 70 years time).

The way to win is to automate and architect. When information enters the organisation in pre-defined structures (or semi-structures), and gets handled automatically you can have both efficiency and accuracy.

Is really a tradeoff between the cost and quality of records.

There’s an interesting trend in records.

We’re de-professionalising record keeping.

Record keeping duties have been given to non-professionals – “business users.”

What hasn’t moved though, is accountability.

Records managers still have accountability for the quality of organisational records.

Their ability to ensure quality though generally boils down to the ability to “ask nicely again”.

So the quality of records we’re producing is deteriorating, while at the same time the compliance statements are being signed, and the bill for the records team is going down.

What’s really happening is that we’re – 

1. Moving the cost of records from a records team to business users (non-professionals).
2. Accepting a deterioration in the quality of our organisational records.
3. Making large portions of the cost of records invisible.

The reasons are understandable, if record keeping is always done by professionals, the number of record keepers required scales linearly with the size of the organisation.

I have to wonder though, if the long term costs will be more expensive than the perceived gains.

Are we making a trade off? Or just making it harder to understand the costs we’re incurring, and making the value that we’re providing look like it came at no cost?

Records provide evidence of how we’ve worked in the past.

AI is about how we should exercise it into the future.

AI of all forms has to learn.

It needs to learn from how we’ve done things in the past.

It needs to learn from our records.

Records are a compensating control for power.

Regulated industries only get power because they agree to keep records.

That’s how it works.

And these days it’s particularly important to understand because every organisation is regulated – by everything from accounting standards to occupational health and safety.

Regulation (or legislation) says that “in order to have this power, you have to exercise it under these conditions”

Every now and then, auditors will show up and demand to see your records – to make sure you are using the power you’ve been given under the conditions required.

If you can’t produce records, they’ll put you out of business.

Records are the compensating control for power. The only way to trust is to be able to verify, and the only really feasible way to verify efficiently is to check records. 

This applies to government agencies as well. 

The only way the public trust, is through the ability to verify.

The only reason Government and regulated organisations get power is because they agree to keep records, when they don’t keep records well, one of two things generally happens – 

1. Power gets taken away by a regulator.
2. Corruption and failure forces closure of the organisation.

Is that culture forms around whoever is leading records.

Culture seems to form around one of two world views – 

1. “Records is important, it’s how we do things here”
2. “We hate records” and it’s “a waste of time”.

In organisations where records is lead well, a culture that believes in the importance of records forms.

Culture is important because it’s self enforcing. Culture dictates behaviour when no one is around to manage.

So there are only two questions to ask – which world view is leading records in your organisation? Is it yours?

Automation provides two possibilities – 

1. Get things right at scale.
2. Get things wrong at scale.

When things go wrong, they’re going to go wrong at scale.

This means that audit needs to happen more often.

Like this:

1. Review your legislation/regulation/standard and identify everything you need to prove your organisation did.
2. Identify information you already capture that can prove it.
3. Develop processes to capture what’s missing.
4. Implement your processes.
5. Audit the process regularly to make sure it’s working.

Develop procedures to collate the information you capture into a complete record for your auditor.

That’s it. Everything else is organisational willpower and good change management.

This is part of what records managers do.

Great ones will help you get it right.

Then:

1. Your audits will be quick and cheap.
2. Your auditors will love you (and flag you as low risk, and leave you alone).
3. You’ll have a minimum of organisational angst at audit time.
4. You can focus on making your customers feel great about what you do.

Great records managers are the route to a more sane world for any regulated industry, if only they knew it.