065 Garbage In, Gospel Out: Why AI Needs Governance, Contracts & Decision Intelligence with Thaija Dickerson

Everyone is talking about AI, but why are so many organisations still struggling to see real business value?

In this episode of Information Matters, Jacqueline Stockwell sits down with Thaija Dickerson, Co-Founder of Contract IQ, data scientist, and author of 10X Contract Value, to explore the missing link between data, governance, contracts, and business outcomes. 

Together they challenge the idea that more data automatically leads to better decisions and discuss why organisations need decision-ready intelligence, not just dashboards and reports. They explore how contracts can become critical business assets, why governance should be seen as the Department of How rather than the Department of No, and how leaders can bridge the gap between AI insights and real-world action. 

Jacqueline and Thaija discuss:

• Why contracts should be treated as Tier 1 business data 
• The biggest reason AI projects fail to deliver measurable value
• How organisations can move from information overload to decision intelligence
• The role of governance in accelerating AI adoption safely and effectively 
• Why ownership, accountability and authority matter more than technology alone 
• How information leaders can influence business outcomes and demonstrate value 

One of Jacqueline's favourite moments from the discussion centres around the phrase:

"Garbage in, Gospel out."

As organisations increasingly rely on AI-generated outputs, the quality, governance and trustworthiness of information has never been more important. 

If you're an information manager, records manager, governance professional, privacy specialist, data leader, or executive exploring how to turn AI ambition into measurable business results, this episode is packed with practical insights.

Connect with Thaija Dickerson

🌐 Kontrak IQ: https://www.kontrakiq.io

 💼 LinkedIn: https://www.linkedin.com/in/thaidickerson

 🎙️ 10X Contract Value Podcast: https://open.spotify.com/show/0ZWpqLxDQ6wueK0937JLzA

 📚 10X Contract Value Newsletter: https://www.linkedin.com/newsletters/10x-contract-value-731843510135938252

About Information Matters with Jacqueline Stockwell

Information Matters is the podcast for information leaders, governance professionals, records managers, privacy experts and data specialists who want to increase their influence, demonstrate value and lead confidently in the age of AI. Hosted by international bestselling author and CEO Jacqueline Stockwell.

Transcript

SPEAKER_01 0:00

Hello and welcome to today's show. Now we've all heard that data is the new oil, but the truth is most companies are drowning in data and have no idea how to actually use it to move momentum forward. Taider is co-founder of Contract IQ and the author of 10 times contract value. Today we're talking about why AI often fails to get the job done.

SPEAKER_00 0:24

I was going to say thank you so much for having me. I'm honestly really excited about this conversation because I think a lot of leaders are feeling the same tension right now, right? They have more data than ever, they have more technology and ever. And then they're being told that AI is going to transform their entire business. But many of them are still stuck in pilots and they're asking, okay, where is the measurable value? So I think that this is a space that a lot of leaders are looking for answers, and it's the area that I care about. So thank you so much for having me.

SPEAKER_01 0:53

You argue that contracts should be treated as tier one critical data. For a leader trying to connect AI to value, why is the legal nature of contract data more important than the statistical nature of it?

SPEAKER_00 1:07

The way that I think about it is pretty simple, right? Most companies are not short on data. They are really short on usable business leverage. They may know that costs are rising. They may see that their margins are under pressure. They may even know that they have a customer that doesn't meet commitments or a supplier miss certain deadlines. But then they have the question of, okay, what can the business actually do about it? And that is where contract data is different. So kind of putting on my data scientist hat, you know, descriptive analytics will tell us what happened, right? Predictive analytics will help us understand what to anticipate will happen next. But that contract data is really will help us move forward towards an actual action because it'll tell us what we agreed to, what rights the company actually has, what obligations exist, and what leverage the business may have to act right now. And that is why I believe that enrich contract data should be treated as tier one critical data. And honestly, in this market, leaders do not just need more information. They need decision-ready intelligence. And that really shows them where they have the leverage in the business, right? Where they are exposed and where they can actually recover value. And that's when, and I will say when contracts are treated as just files in a repository, companies leave so much value on the table, right? When it comes to their revenue, their profitability, compliance, performance, and even risk. But when they treat their contracts as data-rich business assets, they become enforceable rules across the business. And then they're able to understand what triggers certain actions, what conditions can or must be met, and then what outcomes the business has a right to pursue. You're all based in America, right? Yes, in Houston, Texas, in the US.

SPEAKER_01 2:55

Many leaders view governance as the department of no. But how do you rebrand governance as the department of how when it comes to deploying artificial intelligence? Yes.

SPEAKER_00 3:08

You know, unfortunately, like you mentioned, um governance has gotten the bad rap as the place of no, right? Instead of the place of how. And they've built a reputation for just slowing things down. And governance does really need a rebrand because good governance should help the business actually move faster because there's clear rules in place. That's going to be especially important as we move more and more into AI because people need clarity, right? And I would say there's a bucket, a few buckets of areas that the governance should focus on. Um, first, the data. What data can be used and who should have access to it, right? Human in the loop. When does a human need to review an output or approve an action from the AI? When it comes to recommendations for AI, you know, because everything is progressing from insight to now, what you should actually do, what recommendations should be allowed, right? So it's recommendations with guardrails, not just whatever the AI comes up with. It's these, these are the guardrails that we're gonna have in place for recommendations. Then in ownership, when there is a recommendation, who owns that decision, right? Many times you have an insight and a recommendation, but who's gonna carry that out to make sure that it's executed and own it? Errors and hallucinations, right? When there is discrepancies with the AI, who's gonna be accountable for the AI when it gets it wrong? And then I would say the last part is controls for scale. So when you're ready to take the training wheels off of a pilot, right? What needs to, what controls need to be put in place as you're scaling that? So those would be the the primary areas I would recommend that companies focus on for governance because teams shouldn't be guessing it when it comes to governance, right? They should be not running into friction, but instead becoming governance becoming the operating system and a business accelerator. So I believe that that's how the governance department transitions from being the department of no to the department of.

SPEAKER_01 5:05

Yeah, and it's really interesting because I took a lot about branding. So I think an information management service should be branded. It's easier to sell something. And I use the word sell, you know, people in our industry don't like selling, but we are effectively selling our service to business to get influenced to buy in to get the stuff done within the organization. And I think it's really important where we look at information um management services department of know. It's a kind of a rebrand. There's lots of organisations out there that rebrand, create a slogan, think of something you know useful and effective that you can actually say, Hi, I'm here. Yes, and this is what I do. And it's easier for you to create that band to drive forward. I did a presentation with um my general manager uh at the RMS in May, and that was around branding. So they had an MM project, and they were moving the M drive. So all the organization talked about was MMs and the MM project when the drive closes down and they were dressed in t-shirts, they had signs, reports, and it was really interesting how you can get an organization enthused with some of the initiatives for information management just by branding it something different and giving out MMs and getting them to do taxonomies in MMs, peanuts, you know, all those types of things. So it's very interesting based on what you're saying as well, is to rebrand and get that sort of fresh the department of no. Agreed. Agreed. If artificial intent artificial intelligence identifies as one million dollars value leakage in a contract, but no one is assigned to claw it back, the date that data is worthless. How do you define decision ownership in a way that AI generated actually result in a bank deposit?

SPEAKER_00 6:53

Yes, that is a very good question. And it's honestly one of the biggest gaps that I see today. Companies are very excited when AI finds something, right? It finds leakage, it finds risk, it finds a pricing issue. Then the question, like I mentioned, is then what's supposed to happen, right? You know, who is gonna call the third party or who's gonna actually take on the action? Who's gonna validate contract information? Can you actually take a certain action? You know, who talks to finance? There's so many different things that need to actually happen, but there's no clear decisioning trace that is there. And so an insight can just sit there, right? It's nice to have, but who actually carries that out to execution? And that's a critical point where I see AI value often die, right? It's not in the model, it's not in the dashboard, and it's not even in the handoff between, I mean, it is in the handoff between the actual insight and the action. And so that's where you see the bottom line impact. And when insight doesn't actually have a clear owner, it's just a prettier version of the same old problem. So to me, decision ownership really means every insight needs a named owner, right? At the minimum, it should have a role, right? A named role. And it needs to have a clear next step and the authority that uh that needs to be in place or an escalation decision. Then you also have the measured outcome because you just don't want to have insights. You want to actually see how it's impacting the business.

SPEAKER_01 8:21

Yeah, brilliant. Now you talk about execution versus intelligence. So you've said AI fails execution. Is that a software limitation or is it because we haven't mapped the right decision rights, or is it specifically an AI output?

SPEAKER_00 8:37

Another great question. And I'm gonna tell you it's it's both, right? So some of the issue is software design, and some of it's honestly just the company's overall operating model. So AI is gonna identify the patterns, it's gonna identify risks, it's gonna summarize the contracts, it's gonna make recommendations, but the execution context really still sits with the business, right? The accountability, the authority, what's the actual workflow. And this is where I find that leaders are often underestimating, right? Uh a lot of companies wanna connect the AI to the data, but they do not know what decision rights are gonna be there that the AI actually needs, right? So the AI produces an insight, but no one knows who owns the approval of the action, like I mentioned before. Who owns that risk? Is it gonna be legal? Is it gonna be operations? Is it procurement or sales? Like where what is gonna happen after that actual insight produces itself? And that's where AI will often stall, as I mentioned. Now, every company is nuanced, right? But most companies are not as unique as they would like to think that they are when it comes to decision flows. They still need um approval paths and escalation points and controls and measurable outcomes. But this is the moat where I think that solution providers can really help with this opportunity. So the best solution providers should not just surface insights, they should be actually helping their customers step into that execution model. So if you want AI to drive execution, you have to map out the output from the authority, owner, the process, and the business results. So that's why I say I think it's a it's a hand on hand, and I'm thinking I'm seeing more AI solution providers step up to that, and then some leaders realize that that's a gap that they have to start looking at before trying to throw AI at the problem.

SPEAKER_01 10:31

Yeah, sort your information out before you put AI over the top of it. Yes. There's a new saying I've been talking about quite uh quite a bit over the last couple of podcasts. It's sort of garbaging gospel out because obviously everyone's taking AI as gospel, which I just think is an amazing expression, if I'm totally honest. Agree. Yeah. So let's talk about the authority as outcomes. You said you were a data scientist, the data scientists have the tools but no power. And executives with the power, but actually have no data to kind of make those decisions. How does a leader bridge in the gap so that that person with the authority to act is actually seeing the AI decisions intelligence based on good data?

SPEAKER_00 11:12

Yes, so I the one principle that I use across the board in many aspects is start with the end in mind, right? And that's how you're really going to truly connect AI with the bottom line. You really have to bridge the gap from what decision or outcome you're trying to actually achieve. And honestly, too many AI programs start with what data do we have? But the better question is what decision are we actually trying to improve? And when you're trying to protect margin or cash flow, reduce leakage, avoid compliance exposures, you really need to have that identified and then work backward. So once that decision is actually identified, you can identify what data is actually needed. Because you don't need all data for all decisions, right? You need to identify exactly what data you need, what data needs to be cleaned as a result, but you understand what data is needed. And sometimes when you identify the data, it may not require that as much manipulation for that particular data set as you may anticipate, given the whole grand scheme of things. But you know, what data is needed, who's gonna own that decision and who's gonna have the authority to act, and then what is the actual workflow? And then when you get an outcome, how are you gonna actually measure that outcome? And that is what I've seen really change those executive conversations. You're not just bringing executives more data, right? You're bringing them decision-ready intelligence that's actually tied to outcomes that they are accountable for. The data scientists that they understand the model, right? And the executives, they understand and own the outcomes, but the bridge is really going to be that decision intelligence.

SPEAKER_01 12:53

So, talking about those measurable outcomes, can you walk me through and listeners through a before and after scenario? So, for example, what does a company look like when these five elements, so AI, data, governance, ownership, and authority are finally all tied up together?

SPEAKER_00 13:11

The before state is you have searchable documents and searchable information, right? The after state is now you have everything interconnected for decision-ready intelligence that actually goes to or strengthens the bottom line. So, what this looks like before, the company may be able, for example, to have a repository with their contracts or even a solution with their data in there. But the business is still acting reactively, right? A situation comes up and then they're sprounging around trying to figure out what should happen. I'll use a supplier increase because we're seeing a lot of inflation and tariffs coming into the market, or we have been. That's our new norm. Um, but when it happens, you know, who is going to approve it, right? And has does a contract, what does a contract allow? There's so many details in all of these disparate systems that if you don't have that clearly flushed out, there's a lot of holes and inconsistencies and a lot of leakage that actually happens. So, you know, everyone's busy, right? Finance is chasing numbers, legal is reviewing documents, you know, procurement is managing suppliers, and operations is just trying to keep the boat from sinking, right? But the problem is, is everyone's working from partial information. So what happens in this afterstate is that the company is has a connected intelligence layer. AI is gonna surface that a contract needs certain attention, and the data is gonna be structured around a business decision, and governance then defines what can be trusted. Ownership then makes someone that's gonna be responsible and accountable for what needs to happen. And then who has a final authority to make a call, right? And that's where the company is moving from just where are our contracts? Where what do they allow to where are we actually leaking value? Where are we exposed? Where can we recover margin? And where do we have an opportunity to act before a window closes? And that's really the shift of going from, you know, just searching to actually having decision intelligence.

SPEAKER_01 15:10

I want to talk about red flags. So what is the one sign a company's data strategy is all talk and no execution?

SPEAKER_00 15:18

Such a good question. That is, that is, that is. And I would say that the red flag is when you're seeing that the data is not actually influencing and changing decisions. It's just honestly it's just decoration. If a company can just, they can show a dashboard, but they cannot show what actual decisions were changed as a result from the insights from the dashboards, who acted on it and what outcome was measured as a result. To me, that's just a red flag. At that point, it's not really a data strategy, it's just reporting. That we've we've been in that for the last several decades, right? The the line of sight has to be clear. Data has to go to a decision and decision to an action and an action to value. Otherwise, it's just it's like I said, it's decoration.

SPEAKER_01 16:02

Yeah, and just making sure if that data's come from AI, that you check the data that is correct from AI before it even gets to that point.

SPEAKER_00 16:10

Exactly. And that brings back the governance loop that you mentioned, having those having those clear rules in place. Yeah, it does indeed.

SPEAKER_01 16:18

Let's talk about mindset. So for listeners, if they could fix one thing in their information governance um tomorrow to get closer to the 10 times value, where should they start?

SPEAKER_00 16:30

I think many times teams feel like they have to boil the ocean. So I would say don't do that, right? Pick one business's decision that actually matters. If it's pricing validation, is margin leakage, is cash flow, or any other type of measurable outcome that's that that the company is actually prioritizing. Then ask what information is needed. You know, who's gonna own the decision, who has authority. I can you see I have this same, you know, this same mantra, because it's it's very um is a clear workflow, but that's how you really move from information governance to value governance, right? This once you have that foundation is clear, AI becomes a multiplier rather than it's just continuing to produce uh more information.

SPEAKER_01 17:16

Yeah, so what did I say? Garbage in, gospel out.

SPEAKER_00 17:20

Exactly, exactly.

SPEAKER_01 17:22

So you're a fellow MBA and an MS like myself in 2006, which is more valuable for an information entrepreneur, and I'll say that because there's a lot of things around organizations where entrepreneurs are very much out there taking risks, financial risks. And then there's entrepreneurs that are within organizations that are doing exactly the same thing but don't have that financial risk. So entrepreneurs, yeah. That's why I like to call them informational entrepreneurs.

SPEAKER_00 17:52

I agree.

SPEAKER_01 17:52

Knowing how to read a balance sheet or knowing how to audit and AI's decision log, what's the one thing that they should do?

SPEAKER_00 18:00

To be honest with you, I think leaders need to have both, right? I think that I'm a foundation principles person. Like you have to have a foundational layer there, and AI is just a tool that you use to accelerate things. So technology doesn't really replace those fundamentals, right? Reading a balance sheet helps you understand where value is created, where it is lost, you know, where the business has exposure. But then on the other side, being able to audit AI decision logic helps you understand whether the intelligence is actually influencing the outcomes that actually can be trusted. So the future leader really has to be um the new bilingual, right? And and technology and then the fundamentals. They need to understand the business model, but they also need to understand how the data and AI are shaping those decisions inside the model.

SPEAKER_01 18:48

Yeah, I agree. And I'd add to that as well. There's a lot of fear around AI is gonna take our jobs. And I think it's really important that we remember we are people. Yes, and people utilize AI. Yes. We feed it with things that our creativity and expands. And and I really think that there's a massive future for information leaders that they influence organization, yeah, utilising technology, utilising AI, but there's still gonna be a value and a need for governance, for compliance, for you know, influencing within organizations. Yeah, and it goes back to right at the start, information is gold. Yes, agree. And yeah, we need oil to look after. And the new oil, absolutely. It's been absolutely insanely amazing to have you on the show. You have your own podcast as well. Do you want to tell listeners what podcast that is and how they can reach that podcast?

SPEAKER_00 19:41

Definitely. So I have the 10x contract value podcast, um, where you can actually find it on YouTube and other podcast platforms. I also have a book, 10x Contract Value. It's an Amazon best-selling book. So it really helps leaders think about the unlock between their contracts and creating business value.

SPEAKER_01 20:02

If today's conversation has really got you thinking about the kind of leader you want to be or the culture you're helping create, I would absolutely love to hear your thoughts. Please connect with me on LinkedIn. Let me know what your biggest takeaway is, and don't forget to subscribe to Information Matters for more conversations that help information leaders and change makers creating different workplaces where people can thrive. Until next time, be bold, be brave, and be beautiful.